VOLUME 63 No October-December2015oml.org.lb/Temp/Files/10ac7191-cd0e-465b-8a4d-cb4fae8118cd.pdf ·...

VOLUME 63 No 4 Octobre-DécembreOctober-December2015

JM

LJO

UR

NA

LM

ED

ICA

LL

IBA

NA

IS

Vo

l. 6

3 (

4)

- p

p 1

71

-23

2O

cto

bre

-Dé

ce

mb

re

20

15

Octo

be

r-D

ece

mb

er

LE

BA

NE

SE

ME

DIC

AL

JO

UR

NA

LL

MJ

ISSN 0023-9852 Indexed by • Index Medicus • Embase (Excerpta Medica)

• Cab Abstracts and Global Health - CABI Publishing

Lebanese Medical Journal 2015 • Volume 63 (4) I

JOURNAL MEDICAL LIBANAISLEBANESE MEDICAL JOURNAL

Publication du Comité ScientifiqueOrdre des Médecins du Liban

Publication of the Scientific CommitteeLebanese Order of Physicians

AARRTTIICCLLEESS OORRIIGGIINNAAUUXX // OORRIIGGIINNAALL AARRTTIICCLLEESS

Predictors of inflammatory local recurrenceafter breast-conserving therapy for breast cancerMatched case-control studyhttp://www.lebanesemedicaljournal.org/articles/63-4/original1.pdfR. AKOUM, E. K. ABDALLA, M. SAADE, A. AWDEH, F. ABI-AADN. BEJJANI, A GHOSSAIN, E. BRIHI, A. AUDI ................................................... 171

Parotid tumors: Is surgery always the only way?http://www.lebanesemedicaljournal.org/articles/63-4/original2.pdfC. FARAH, C. GHORRA, C. NASSIF, B. TABCHY, A. HADDAD ......................... 179

Effects of preoperative enoxaparin on bleeding aftercoronary artery bypass surgeryhttp://www.lebanesemedicaljournal.org/articles/63-4/original3.pdfM. ABI GHANEM, E. SALIBA, J. IRANI, R. DARWISH, N. KAWKABANIT. YARED, C. NAIM, B. ABU KHALIL ................................................................... 185

Profil gériatrique selon l’instrumentIdentification systématique des aînés à risque (ISAR)au service des urgences d’un centre hospitalier universitairehttp://www.lebanesemedicaljournal.org/articles/63-4/original4.pdfR. EL-HAYECK, R. BADDOURA, P. FADEL, A. WEHBÉ A. ZOGHBY, M. BERTHEL ..................................................................................... 191

False and true acute rheumatic feverA Lebanese experiencehttp://www.lebanesemedicaljournal.org/articles/63-4/original5.pdfN. JAWHAR, I. CHAMI, Z. SALIBA, G. CHEHAB .............................................. 198

Nasalance scores in Lebanese English-speaking adults using nasometric analysishttp://www.lebanesemedicaljournal.org/articles/63-4/original6.pdfA.L. HAMDAN, G. ZIADE, J. JABBOUR, G. KHNEIZER I. KUTKUT ................................................................................................................ 203

Formant characteristics of English-speaking Lebanese menhttp://www.lebanesemedicaljournal.org/articles/63-4/original7.pdfA.L. HAMDAN, G. ZIADE, Z. AL-ZAGHAL, D. TABRI, S. SINNOR. SAADE, J. JABBOUR, J. NASSAR ..................................................................... 209

Rédacteur en chefEditor in Chief

Adel BERBARI

Rédacteurs en chef adjointsAssociate Editors

Michel DAHER

Ghassan K. ABU-ALFA (USA)Elie BERBARI (USA)

Eileen O’REILLY (USA)

Directeur de la rédaction Managing Editor

Charles JAZRA

Secrétaire général General Secretary

Salim KABALAN

Comité de rédaction Editorial Board

George ARAJ

Elias ARBID

David ATALLAH

Nizar BITAR

Jamal BLEIK

Zaki GHORAYEB

Zuheir EL IMAD

Abdo JURJUS

Kamal KALLAB

Roland KASSAB

Adib MOUKARZEL

Mariam RAJAB

Fouad EL SAYED

Georges YARED

Majed YAZBECK

Fouad ZAAROUR

Coordinateur du site WebWebsite coordinator

Adib MOUKARZEL

Consultant en statistiquesStatistics AdvisorRobert KASPARIAN

Secrétaire de rédactionExecutive SecretaryJacqueline CHAHINE

II Lebanese Medical Journal 2015 • Volume 63 (4)

JMLJOURNAL MEDICAL

LIBANAIS

LMJLEBANESE

MEDICAL JOURNAL

Publication du Comité Scientifique

Ordre des Médecins du Liban

Publication ofthe Scientific Committee

Lebanese Order of Physicians

Adressertoute correspondance au

Rédacteur en ChefJournal Médical Libanais

Ordre des Médecins du LibanAutostrade Tawita

Furn el-Chebbak - BeyrouthLiban

Mailing address

Editor in ChiefLebanese Medical Journal

Lebanese Order of PhysiciansAutostrade Tawita

Furn el-Shebbak - BeirutLebanon

e-mail : [email protected]

www.lebanesemedicaljournal.org

Tel./Fax : +961 1 610710 ext 306

Production emoBeyrouth

Tel. : +961 1 339870ISSN 0023-9852

Indexed by • Index Medicus • Embase (Excerpta Medica)• Cab Abstracts and Global Health - CABI Publishing

MMIISSEESS AAUU PPOOIINNTT// IINN--DDEEPPTTHH RREEVVIIEEWWSS

Towards developing a sustainable Faculty Development Program An initiative of an American medical school in Lebanonhttp://www.lebanesemedicaljournal.org/articles/63-4/review1.pdfB. RAHAL, N. MANSOUR, G. ZAATARI ................................................................. 213

Management of metastatic colorectal cancerCurrent treatments and new therapieshttp://www.lebanesemedicaljournal.org/articles/63-4/review2.pdfA-M. HENAINE, G. CHAHINE, P. SALAMEH, E. ELIASM. MASSOUD, D. HARTMANN, G. AULAGNER, X. ARMOIRY ......................... 218

CCAASS CCLLIINNIIQQUUEESS// CCAASSEE RREEPPOORRTTSS

Cerebral aneurismal rupture with subarachnoid hemorrhage during pregnancy: a case reporthttp://www.lebanesemedicaljournal.org/articles/63-4/case1.pdfD. ATALLAH, F. MANSOUR, E. SAMAHAN. EL KASSIS, J. NASSIF ........................................................................................ 228

IINN MMEEMMOORRIIAAMM

Joseph Chucri Moubarak • 1931-2015http://www.lebanesemedicaljournal.org/articles/63-4/inmemoriam1.pdfG. SALIBA .................................................................................................................. 232

INTRODUCTION

Ipsilateral breast tumor recurrence (IBTR) after breast-conserving surgery (BCS) and whole-breast radiationoccurs in 4% to 10% of patients [1-3]. Late local recur-rence (> 2 years) in a conserved breast can usually betreated by further surgery with curative intent and re-mains associated with a favorable prognosis [4-5]. Earlylocal recurrence (< 2 years) may indicate a more aggres-

sive clinicobiological behavior that heralds the appear-ance of distant metastases in many cases [1,6-7]. A smallbut important subset of patients with local recurrencedevelops inflammatory local recurrence (ILR), which isassociated with a distinctly poorer prognosis. An uncom-mon finding representing 7% to 20% of recurrences [8-12], sometimes referred to as “secondary” inflammatorybreast cancer, ILR typically occurs within two years ofBCS, exhibits clinical and pathological characteristics of

Lebanese Medical Journal 2015 • Volume 63 (4) 171

AARRTTIICCLLEE OORRIIGGIINNAALL// OORRIIGGIINNAALL AARRTTIICCLLEEPREDICTORS OF INFLAMMATORY LOCAL RECURRENCE AFTER BREAST-CONSERVINGTHERAPY FOR BREAST CANCER: MATCHED CASE-CONTROL STUDYhttp://www.lebanesemedicaljournal.org/articles/63-4/original1.pdf

Riad AKOUM1, Eddie K. ABDALLA2, Michel SAADE1, Adnan AWDEH2, Fouad ABI-AAD2, Noha BEJJANI4

Antoine GHOSSAIN2, Emile BRIHI3, Akram AUDI5

Akoum R, Abdalla EK, Saade M, Awdeh A, Abi-Aad F, Bejjani N,Ghossain A, Brihi E, Audi A. Predictors of inflammatory localrecurrence after breast-conserving therapy for breast cancer:Matched case-control study. J Med Liban 2015 ; 63 (4) : 171-178.

Akoum R, Abdalla EK, Saade M, Awdeh A, Abi-Aad F, Bejjani N,Ghossain A, Brihi E, Audi A. Facteurs prédictifs de récidive localeinflammatoire après chirurgie conservatrice du cancer du sein : Étudecomparative cas-témoins. J Med Liban 2015 ; 63 (4) : 171-178.

Divisions of Medical Oncology1, Surgical Oncology2, Radiation Oncology3, Department of Pathology4, Internal Medicine Resident5,Lebanese American University Medical Center-Rizk Hospital, Beirut, Lebanon.

Correspondence: Riad Akoum MD. e-mail : [email protected] Fax: +961 1 200816

RÉSUMÉ • Objectifs : La récidive locale inflammatoire (RLI)après traitement conservateur du cancer du sein est associéeà un mauvais pronostic. Les facteurs de risque d’une tellerécidive ne sont pas encore bien connus. Méthodes : Entre2001 et 2010, douze patientes présentant une RLI après chi-rurgie conservatrice, chimio- et radiothérapie adjuvantes ontété traitées dans notre département. Nous avons réalisé uneanalyse comparative des caractéristiques clinico-pathologi-ques de ces patientes avec 24 patientes présentant une réci-dive locale non inflammatoire, 24 patientes présentant desmétastases à distance et 48 patientes en rémission persis-tante. La stratification a tenu compte de l’âge et de la date dudiagnostic. Résultats : La RLI survenait après une périodemoyenne de 10 mois. L’analyse univariée a montré une corré-lation significative des facteurs suivants: atteinte lymphatique(p < 0,05), multifocalité (p < 0,05), surexpression de c-erbB2(p < 0,05) et invasion des vaisseaux lymphatiques (IVL)(p < 0,001). La régression logistique conditionnelle a montréune corrélation significative entre RLI et IVL associée au hautgrade histologique [OR6,14; CI 95% pour la comparaison RLIavec récidive locale non inflammatoire et OR 3,05; CI 95%pour la comparaison RLI avec métastases à distance]. Lacomparaison de ces mêmes groupes de patients en consi-dérant la présence d’antécédents familiaux de cancer du seina donné des ORs plus significatifs : 5,47 et 5,62. Conclusion:Les femmes pré- et postménopausées présentant un cancerdu sein de haut grade histologique avec invasion lympho-vasculaire sont à risque élevé de développer une RLI sur-tout en présence d’antécédents familiaux de cancer du sein.L’identification des facteurs de risque de cette RLI pourraitcontribuer à l’élaboration d’une stratégie thérapeutique plusefficace.

ABSTRACT • Background: Inflammatory local recurrence(ILR) after breast-conserving surgery for noninflammatorybreast cancer (BC) is associated with dismal prognosis. Riskfactors for ILR are not well defined. Methods : Between 2001and 2010, twelve patients at our hospital developed ILR afterbreast-conserving surgery, adjuvant chemotherapy, and radio-therapy for BC. We compared their clinico-pathological char-acteristics to those of 24 patients with noninflammatory localrecurrence (non-ILR), 24 patients with distant metastases,and 48 disease-free controls, matched for age and observa-tion period. Results : The median time to ILR was 10 months.In univariate analysis, extent of lymph node involvement(p < 0.05), multifocality (p < 0.05), c-erbB2 overexpression(p < 0.05), and lymphovascular invasion (LVI) (p < 0.001)affected the risk of ILR. Conditional logistic regression analy-sis showed a significant association between ILR and com-bined LVI and high histopathological grade. The odds ratio(OR) for ILR versus non-ILR was 6.14 (95% confidence inter-val [CI] 1.48-25.38) and for ILR versus distant metastases itwas 3.05 (95% CI 0.09-97.83) when both LVI and high histo-pathological grade were present. Patients with family histo-ry of BC were more likely to present with ILR than non-ILR(OR 5.47; 95% CI 1.55-19.31) or distant relapse (OR 5.62;95% CI 0.26-119.95). Conclusions: Pre- and postmeno-pausal women with high-grade BC and LVI are at increasedrisk to develop ILR, especially in the presence of family histo-ry of BC. Identification of risk factors for this lethal form ofrecurrent BC may lead to more effective preventive treatmentstrategies in properly selected patients.

Keywords : Inflammatory breast cancer, local recurrence,inflammatory local recurrence, chemotherapy, radiotherapy,mastectomy

172 Lebanese Medical Journal 2015 • Volume 63 (4) R. AKOUM et al. – Inflammatory recurrence after breast surgery

inflammatory breast cancer, and is surgically and onco-logically difficult to control. Patients with ILR have onlya 14% rate of 5-year overall survival compared to 50%for those with noninflammatory local recurrence (non-ILR) [4,8,13-15].

Pathological risk factors for IBTR after BCS includeclose or positive surgical margins [1,3,12,16], presenceof extensive intraductal component (EIC) in and/oraround the tumor [1,17-18], presence of lymphovascularinvasion (LVI) with or without positive axillary lymphnodes [1,3,7,16,19], histological multifocality [7,16],high histopathological grade [16], negative hormonalreceptor status [7], and c-erbB2 oncogene overexpres-sion [20]. Clinical risk factors for IBTR include tumorsize [16-17], young age [2-3,18-19], and gross multicen-tricity [9]; however, a strong family history of breastcancer has been associated with late IBTR and whetherthese are true recurrences or actually new tumors is con-troversial [21]. Skin involvement has been reported in3% to 22% of breast recurrences in general [6,22].

ILR is diagnosed clinically, based on the appearanceof diffuse redness and edema of the breast. The histo-pathological examination frequently reveals cancer em-boli in the lymphatic vessels. However, the absence ofthis characteristic does not rule out the diagnosis. Whenthe clinical criteria are missing, and a dermal lymphaticinvasion is seen in the pathological specimen, the term“occult inflammatory breast cancer” has sometimes beenused. However, whether occult inflammatory breast can-cer carries as poor a prognosis as inflammatory breastcancer is still a matter of debate. Some studies report anoutcome similar to that reported for inflammatory breastcancer [23-24]; others describe “occult inflammatorybreast cancer” as a more benign process. This type ofrecurrence has been associated with multicentricity [7,10-11], multifocality [25-26], positive resection margins[7,16], high number of positive lymph nodes [7,25], andLVI [4,7-8,12-15].

The aim of this study was to determine whether iden-tifiable clinico-pathological characteristics in womenwith primary noninflammatory, invasive breast canceraffected their risk of ILR after BCS.

MATERIAL AND METHODS

Study designTo assess potential risk factors for ILR after BCS forbreast cancer, we designed a case-control study in a hos-pital-based population. Data on cases and controls wereextracted from the hospital radiotherapy registry. Allcases and controls had primary noninflammatory, inva-sive breast cancer and were treated between 2001 and2010 with BCS, doxorubicin and taxane chemotherapy,and radiation therapy at the same dose. Patients with apositive estrogen and/or progesterone receptor receivedhormonal therapy. Patients with c-erbB2 positive statustreated after 2007 received adjuvant Trastuzumab.

Patients who developed ILR were compared with three

control groups: • patients who presented with non-ILR(selected in a 2:1 ratio) • patients who presented withmetachronous distant metastases (2:1) and • patients withno evidence of disease at least three years after the initialtreatment (4:1). All controls received adjuvant chemo-therapy containing a taxane and doxorubicin, and all ofthem received adjuvant radiation therapy delivering 50 Gyto the whole breast in 25 sessions, followed by a boost of10 Gy in five sessions to the involved quadrant.

The following data were analyzed: age, tumor size,time to recurrence, pathological grading, lymph node sta-tus, surgical margins, lymphatic vessel invasion by tumorcells, multifocality, presence of EIC, and immunohisto-chemical staining for estrogen receptor (ER), progester-one receptor (PR), and c-erbB2 oncoprotein expression.Clinical data were also analyzed, including common co-morbidities: diabetes mellitus, obesity, and infectiousmastitis that had completely resolved on antibiotic thera-py. The last updated follow-up was in 2011. Internal mam-mary relapse was considered as metastatic disease. Re-section margins ≤ 1 mm in width were considered positivemargins.

Local recurrences in cases and controls were histo-pathologically confirmed and found to be identical to thecorresponding initial tumors with regards to grade andER, PR, and c-erbB2 status.

Statistical analysisThe clinical and pathological variables were comparedusing χ2 test, Fisher’s exact test, and Student’s t test. Fami-ly history, ER and PR status, menopausal status, c-erbB2status, LVI, EIC, and multifocality were coded as dichoto-mous variables. Tumor size, grade, and lymph node statuswere coded as categorical variables. Logistic regressionwas used to evaluate the association of each potentialrisk factor with the occurrence of ILR. A conditional lo-gistic regression model adapted for case-control studieswith 2 or 4 matched controls per case was applied. Thesignificance threshold was set at p < 0.05, and oddsratios (ORs) and 95% confidence intervals (CIs) weredetermined. Intercooled Stata software, version 9.0 (Stata-Corp, 2005), was used for the statistical analysis.

This retrospective review was conducted in accor-dance with institutional policy.

RESULTS

We identified 12 pre- or postmenopausal women withILR, mean age 53 years (range, 30-65 years). All ILRcases exhibited intradermal lymphatic vessel invasionthat was absent in their initial tumors (Figure 1). All ILRoccurred within 20 months from the completion of thera-py. The median time to recurrence was 10 months, signi-ficantly shorter than in the non-ILR and metastatic controlgroups (26 and 36 months respectively, p < 0.001).

Table I summarizes the baseline patient and tumorcharacteristics of cases and matched controls. Tumor size,EIC, negative ER and PR status, triple negative receptor

R. AKOUM et al. – Inflammatory recurrence after breast surgery Lebanese Medical Journal 2015 • Volume 63 (4) 173

a. Photomicrograph of histopathological specimen from the initial breast tumor of a patient with a typical infiltrating ductal carcinoma ofcolloid type (hematoxylin and eosin x 20). b. Photomicrograph of the inflammatory local recurrence from the same patient showingtumor emboli within the intradermal llymphatic vessels (hematoxylin and eosin x 20). c. A photograph of the skin involvement in thesame patient with inflammatory local recurrence despite chemotherapy and electron beam radiation to the skin.

TABLE IBASELINE CHARACTERISTICS of PATIENTS with INFLAMMATORY LOCAL RECURRENCE and MATCHED CONTROLS

Inflammatory NoninflammatoryDistant metastases NED plocal recurrence local recurrence (n = 24) (n = 48)(n = 12) (n = 24)

Mean age / years (range) 53 (30-65) 54 (29-65) 57 (32-66) 54 (29-67)Median TTR / months(range) 10 (3-20) 26 (4-130) 36 (8-130) < 0.001

I 0 3 (13%) 3 (13%) 7 (15%)Grade II 4 (33%) 13 (54%) 15 (63%) 33 (69%)

III 8 (67%) 8 (34%) 6 (23%) 8 (17%) 0.061 0 4 (17%) 1 (4%) 22 (46%)

T Class 2 9 (75%) 17 (71%) 18 (76%) 24 (50%)3 3 (25%) 3 (13%) 5 (21%) 2 (4%) NS0 1 (8%) 9 (38%) 2 (8.5%) 29 (62%)

N Class 1-4 3 (25%) 8 (34%) 15 (63%) 16 (34%)> 4 8 (67%) 7 (29%) 7 (29%) 3 (5%) < 0.05

Free 8 (67%) 19 (79%) 17 (71%)

Surgical margins Free after 2 (16%) 3 (13%) 5 (21%) 3 (5%)re-resection

Positive 2 (16%) 2 (9%) 2 (8%) 1 (2%) NS

Multifocality No 7 (58%) 22 24Yes 5 (42%) 2 (9%) 0 4 (9%) < 0.05

EIC No 6 (50%) 20Yes 6 (50%) 4 (17%) 7 (29%) 21 (44%) NS

ER Positive 22 19 (79%) 37 (76%)Negative 3 (25%) 2 (9%) 5 (21%) 11 (24%) NS

PR Positive 9 (75%) 19 14 (58%) 36 (74%)Negative 3 (25%) 5 (21%) 10 (42%) 12 (26%) NS

c-erbB2 Negative 2 (16%) 14 (58%) 14 (58%) 34 (70%)Positive 10 (84%) 10 (42%) 10 (42%) 14 (30%) < 0.05

Triple negative* Yes 2 (16%) 1 (4%) 0 5 (11%) NS

LVI Absent 0 16 17 43Present 12 (100%) 8 (34%) 7 (29%) 5 (11%) ≤ 0.001

Personal history Positive 0 1 (4%) 1 (4%) 1 (2%) NSFamily history Positive 1 (8%) 3 (13%) 1 (4%) 4 (9%) NSDiabetes Present 2 (16%) 2 (9%) 1 (4%) 1 (2%) NSObesity Present 3 (25%) 5 (21%) 4 (17%) 4 (9%) NSMastitis ** Present 0 6 (13%) < 0.0001

TTR< 20 months 12 (100%) 12 (50%) 17 (71%)> 20 months 12 (50%) 7 (29%) 0.05

NED: no evidence of disease TTR: time to recurrence NS: not statistically significant EIC: extensive intraductal component ER: estrogen receptorPR: progesterone receptor *Triple negative: negative ER, negative PR and negative c-erbB2 LVI: lymphovascular invasion**Mastitis: breast cellulitis complicating breast conservation surgery

FIGURE 1 a b c


status (ER, PR, c-erbB2), and positive surgical marginswere not significantly correlated with the type of recur-rence. The risk of ILR was significantly affected by theextent of lymph node involvement (p < 0.05), multifo-cality (p < 0.05), LVI (p < 0.001), and positive c-erbB2status (p < 0.05). Histopathological grade did not signi-ficantly distinguish between recurrence groups (p = 0.06).Six disease-free controls (12.5%) developed infectiousmastitis within 15 months from the BCS; none of thesepatients developed ILR, and none of the cases with ILR orthe controls with non-ILR developed this benign postop-erative condition (p < 0.0001). Obesity, diabetes mellitus,and personal and family history of breast cancer were notassociated with the type of recurrence.

Survival analysisComplete follow-up data were available on 100% of ILRpatients. All ILR patients received cisplatin-based chemo-therapy. Three patients developed distant metastasesshortly afterwards. Four patients underwent mastectomyafter a partial response to chemotherapy, and four patientsdeveloped extensive skin involvement; however, they alldied with metastatic disease within 18 months. One pa-tient remains alive with disease at current follow-up, four

years after the ILR, with extensive centrifugal skin inva-sion without distant metastasis.

Univariate logistic regressionHistopathological grade III was significantly associatedwith the occurrence of ILR rather than non-ILR (OR5.62; 95% CI 1.34-23.50; p = 0.02), distant metastases(OR3.89; 95%CI 1.00-15.00; p = 0.05), or long-term re-mission (OR 9.11; 95% CI 2.31-35.80; p = 0.002). Lym-phovascular invasion was significantly more likely to beassociated with ILR (p < 0.0001) than with other out-comes. Multifocal tumors were more likely to predict ILRthan to predict non-ILR (OR27.70; 95%CI 1.28-598.00;p = 0.001), distant metastasis (OR 11.00; 95% CI 1.75-69.00; p = 0.01), or no evidence of disease (OR 11.25;95% CI 2.44-51.70; p = 0.002). Positive margins, tumorsize, lymph node status, EIC, ER, PR, and c-erbB2 sta-tus, and family history of breast cancer were not signifi-cant predictors of ILR (Table II).

Unconditional multivariate logistic regressionIn comparing ILR cases with non-ILR controls usingmultivariate logistic regression, LVI, high grade, andmultifocality were predictors of ILR (data not shown).

EIC: extensive intraductal component ER: estrogen receptor LVI: lymphovascular invasion NS: not significantPR: progesterone receptor TTR: time to recurrence +: positive or positive findings **Variable “perfectly” predicts outcome

In comparing ILR cases with distant metastases controls,LVI predicted ILR perfectly, whereas multifocality andhigh grade were less significantly predictors of outcome.

Conditional logistic regressionBased on univariate regression models, tumor grade, LVI,multifocality, EIC, negative ER and PR status, positivec-erbB2 status and family history were entered into con-ditional logistic regression models. These variables gavebetter fits by comparing the log likelihood of the models.

Comparison of ORs in the univariate and multivariateconditional logistic regression analyses further clarifiedthe relationship of the association of two or more factorsin the prediction of ILR. Odds ratios remained signifi-cantly high for LVI and pathological grading as predic-tors of ILR (Table III).

Family history of breast cancer had no significant rolein predicting recurrence; however, patients with familyhistory of breast cancer were more likely to present withILR than with non-ILR (OR3.53; 95%CI 0.22-54.60) ordistant relapse (OR 48.40; 95% CI 0.07-31,287).

When all variables were included in the conditionallogistic regression analysis to tease out the effects ofeach factor, controlling for the others, only LVI andgrade III were strongly associated with ILR as comparedto non-ILR, distant metastasis, and no evidence of dis-ease (ORs: 6.14 [95% CI 1.48-25.38], 3.05 [95% CI0.095-97.83], and 28.43 [95% CI 2.33-346.17], respec-tively, when both LVI and grade III were present). When

family history of breast cancer was added to this combi-nation of risk factors, the ORs remained greater than 4(Figure 2).

DISCUSSION

This case-control study is the largest reported to date thatexamines risk factors for ILR that can be identified at thetime of treatment of the primary tumor. The results ofconditional multivariate regression demonstrated a con-sistent trend toward a higher likelihood of LVI and highgrade among women with unilateral, invasive breast can-cer who developed ILR compared with women matchedfor age, adjuvant treatment modalities, and diagnosisperiod who did not develop ILR. These characteristicswere also potent risk factors for the development of ILRin patients with a positive family history of breast cancer.Until new molecular prognostic factors and new markersof inflammatory breast cancer are found, the presence ofLVI and grade III could be signals of elevated risk ofILR.

Two previous studies attempted to correlate histologi-cal characteristics with the risk of ILR after BCS for non-inflammatory breast cancer [8,13]. One of them collectednine cases from 18 Japanese centers over a 9-year period[8]. The other collected seven cases from the New YorkPresbyterian Hospital over a 5-year period [13]. In bothstudies more than 10 risk factors were analyzed: tumorsize and location, histological grade and type, lymph node


EIC: extensive intraductal component ER: estrogen receptor LVI: lymphovascular invasion NS: not significant PR: progesterone receptorTTR: time to recurrence +: positive or positive findings **Log likelihood not sufficient to enable conditional regression for these variables


involvement, resection margins, LVI, ER and PR status,and adjuvant treatment. In the first study, Nishimura et al.,using a case-control design and unconditional logisticregression analysis, observed that LVI was the major pre-dictor of ILR, whereas lymph node involvement, positivesurgical margins, and lack of adjuvant radiation therapywere not predictors [8]. They concluded that ILR is main-ly due to “occult” inflammatory breast cancer, since allrecurrences occurred within 12 months. In the secondstudy, only lymph node involvement and ER status werecorrelated with the time to ILR and with the overall sur-vival time [13]. Unusual cases of ILR occurring 3 yearsafter BCS were reported.

In a review of immunohistological c-erbB2 expres-sion in 1,794 patients with primary breast cancer in Italy,Menard et al. found that ductal type, LVI, high nucleargrade, and negative ER and PR status were associatedwith high c-erbB2 expression [20]. Higher c-erbB2expression was also associated with higher relapse rates,even in node-negative patients. However, conditionallogistic regression in the present study did not highlightc-erbB2 expression as a major determinant in predictingILR.

Some histological prognostic factors, such as LVI andmultifocality, have not always been considered impor-tant enough to be incorporated into guidelines for moreaggressive adjuvant therapy. Lymphovascular invasion

has been associated with an increased risk of local fail-ure in some series [1,7,19] but not all [17]. The impor-tance of tumor grade in local recurrence is still contro-versial [7,16]. The current study highlights its role inassociation with LVI as a predictor of ILR. Hormonalreceptor and c-erbB2 status are relevant prognostic fac-tors [20], and multifocality has been associated withhigh risk of local recurrence [9,17], but these factorsseem to be unrelated to the inflammatory nature of thelocal recurrence in this study.

Lymphovascular invasion is a widely recognized prog-nostic factor in node-negative breast cancer. Its signif-icance as an independent prognostic factor in node-posi-tive breast cancer remains uncertain [27]. The presentcase-control study demonstrated that LVI is a strong pre-dictor of ILR in both situations, especially in coexistencewith high-grade tumor and family history of breast cancer.

Cutaneous and subcutaneous LVI by cancer cells havebeen seen in up to 80% of inflammatory breast cancer,whereas less than 2% of operable noninflammatorycases exhibit such pathological features [28-29]. Recentstudies have attributed these pathological findings towhat is described as “occult inflammatory breast can-cer,” while recurrent tumors with obvious clinical fea-tures of inflammatory breast cancer are described as“secondary inflammatory breast cancer” [28-29]. Assuch, LVI may be a selection factor for more aggressive

Figure 2. Results of multivariate conditional logistic regression analysis. Comparison of the odds ratios for inflammatory local recurrenceof breast cancer when histopathological grade III and lymphovascular invasion (LVI) were present. • Upper panel: No family history of breastcancer • Lower panel: Family history of breast cancer present in addition to grade III and LVI.

ILR: inflammatory local recurrence Non-ILR: noninflammatory local recurrence NED: no evidence of disease.


treatment for these patients up front in order to lowertheir risk of post-treatment recurrence.

Whether ILR is a secondary characteristic acquiredby the recurrent lesion or related to the primary cancerremains to be clarified. The findings of this study may beinterpreted to indicate that the coexistence of high histo-logical grade and LVI in a breast tumor suggests anaggressive behavior; thus, these findings could be usedto alert the oncologist to higher risk of ILR and may for-mulate the basis of a future study population.

The main methodological limitation to this study isthe relatively small number of cases, though this is thelargest reported series of patients with ILR. However,matching controls to cases on the year of diagnosis andthe age group minimized the effects of periodic or physi-ological differences. The study design, the homogeneityof therapeutic approaches, and the complete pathologicfollow-up enabled relatively in-depth analysis of thisrare condition.

Unfortunately, no therapeutic approach has beenproven to prevent ILR. Whether our patients with high-grade tumors, family history of breast cancer, and LVIshould have been initially more aggressively treatedraises a relevant question in the era of growing tendencytowards conservative surgery and sentinel lymph nodebiopsy in early-stage patients such as in the Z0011 trial[30].

Identification of patients at elevated risk of ILR couldlead to more effective strategies of treatment at the timeof therapy for the primary tumor, designed to lower theincidence of recurrence of this lethal disease. Next stepsmay include larger cohort and case-control studies oranalyses based on prospective datasets that include morebiochemical and molecular factors. These would identi-fy more precisely which group of patients are at risk andallow development of methods of reducing the risk ofsuch recurrences.

ACKNOWLEDGMENTS

The authors thank Faith Reidenbach for editorial assis-tance, which they sponsored, and declare to have no con-flicts of interest to disclose.

REFERENCES

1. Veronesi U, Marubini E, Del Vecchio M et al. Local re-currences and distant metastases after conservative breastcancer treatments: partly independent events. J NatlCancer Inst 1995; 87 (1): 19-27.

2. Fisher B, Anderson S, Redmond CK, Wolmark N,Wickerham DL, Cronin WM. Reanalysis and results after12 years of follow-up in a randomized clinical trial com-paring total mastectomy with lumpectomy with or with-out irradiation in the treatment of breast cancer. N Engl JMed 1995; 333 (22): 1456-61.

3. Borger J, Kemperman H, Hart A, Peterse H, van Dongen J,Bartelink H. Risk factors in breast-conservation therapy.J Clin Oncol 1994; 12 (4): 653-60.

4. Salvadori B, Marubini E, Miceli R et al. Reoperation

for locally recurrent breast cancer in patients previouslytreated with conservative surgery. Br J Surg. 1999; 86(1): 84-7.

5. Gentilini O, Botteri E, Rotmensz N et al. When can a sec-ond conservative approach be considered for ipsilateralbreast tumour recurrence? Ann Oncol 2007; 18 (3): 468-72.

6. van Tienhoven G, Voogd AC, Peterse JL et al. for theEORTC Breast Cancer Cooperative Group and the DanishBreast Cancer Cooperative Group. Prognosis after treat-ment for loco-regional recurrence after mastectomy orbreast conserving therapy in two randomised trials(EORTC 10801 and DBCG-82TM). Eur J Cancer 1999;35 (1): 32-8.

7. Nishimura R, Akizuki M, Tashima R, Ootao R. Investi-gation of factors related to periods to ipsilateral breasttumor recurrence after breast-conserving surgery andmeasures for preventing recurrence in early breast can-cer. Breast Cancer 2006; 13 (2): 152-8.

8. Nishimura R, Koyama H, Kasumi F et al. A case controlstudy on risk factors involved in inflammatory breast re-currence after breast-conserving surgery. Oncology 1998;55 (5): 391-9.

9. Kurtz JM, Jacquemier J, Brandone H et al. Inoperable re-currence after breast-conserving surgical treatment andradiotherapy. Surg Gynecol Obstet 1991; 172 (5): 357-61.

10. Fisher ER, Gregorio R, Redmond C, Vellios F, Sommers SC,Fisher B. Pathologic findings from the National SurgicalAdjuvant Breast Project (protocol no. 4). I. Observa-tions concerning the multicentricity of mammary cancer.Cancer 1975; 35 (1): 247-54.

11. Fisher ER, Sass R, Fisher B, Gregorio R, Brown R,Wickerham L. Pathologic findings from the NationalSurgical Adjuvant Breast Project (protocol 6). II. Rela-tion of local breast recurrence to multicentricity. Cancer1986; 57 (9): 1717-24.

12. Voogd AC, Peterse JL, Crommelin MA et al. Histologicaldeterminants for different types of local recurrence afterbreast-conserving therapy of invasive breast cancer.Dutch Study Group on local Recurrence after BreastConservation (BORST). Eur J Cancer 1999; 35 (13):1828-37.

13. Huston TL, Simmons RM. Inflammatory local recur-rence after breast-conservation therapy for noninflamma-tory breast cancer. Am J Clin Oncol 2005; 28 (4): 431-2.

14. Alpert TE, Kuerer HM, Arthur DW, Lannin DR, HafftyBG. Ipsilateral breast tumor recurrence after breast con-servation therapy: outcomes of salvage mastectomy vs.salvage breast-conserving surgery and prognostic factorsfor salvage breast preservation. Int J Radiat Oncol BiolPhys 2005; 63 (3): 845-51.

15. Casalini P, Carcangiu ML, Tammi R et al. Two distinctlocal relapse subtypes in invasive breast cancer: effect ontheir prognostic impact. Clin Cancer Res 2008; 14 (1):25-31.

16. Cowen D, Jacquemier J, Houvenaeghel G et al. Localand distant recurrence after conservative management of“very low-risk” breast cancer are dependent events: a10-year follow-up. Int J Radiat Oncol Biol Phys 1998; 41(4): 801-7.

17. Touboul E, Buffat L, Belkacemi Y et al. Local recur-rences and distant metastases after breast-conservingsurgery and radiation therapy for early breast cancer. IntJ Radiat Oncol Biol Phys 1999; 43 (1): 25-38.

18. Voogd AC, van Tienhoven G, Peterse HL et al. for theDutch Study Group on Local Recurrence after BreastConservation (BORST). Local recurrence after breastconservation therapy for early stage breast carcinoma:detection, treatment, and outcome in 266 patients. Cancer1999; 85 (2): 437-46.

19. Fourquet A, Campana F, Zafrani B et al. Prognostic fac-tors of breast recurrence in the conservative managementof early breast cancer: a 25-year follow-up. Int J RadiatOncol Biol Phys 1989; 17 (4): 719-25.

20. Menard S, Balsari A, Tagliabue E et al. Biology, progno-sis and response to therapy of breast carcinomas accord-ing to HER2 score. Ann Oncol 2008; 19 (10): 1706-12.

21. Turner BC, Harrold E, Matloff E et al. BRCA1/BRCA2germline mutations in locally recurrent breast cancerpatients after lumpectomy and radiation therapy: impli-cations for breast-conserving management in patientswith BRCA1/BRCA2 mutations. J Clin Oncol 1999; 17(10): 3017-24.

22. Gage I, Schnitt SJ, Recht A et al. Skin recurrences afterbreast-conserving therapy for early-stage breast cancer. JClin Oncol 1998; 16 (2): 480-6.

23. Bonnier P, Charpin C, Lejeune C et al. Inflammatory car-cinomas of the breast: a clinical, pathological, or a clini-cal and pathological definition? Int J Cancer 1995; 62(4): 382-5.

24. Liauw SL, Benda RK, Morris CG, Mendenhall NP. In-flammatory breast carcinoma: outcomes with trimodalitytherapy for nonmetastatic disease. Cancer 2004; 100 (5):920-8.

25. Silvestrini R, Daidone MG, Luisi A et al. Biologic andclinicopathologic factors as indicators of specific relapsetypes in node-negative breast cancer. J Clin Oncol 1995;13 (3): 697-704.

26. Marret H, Perrotin F, Bougnoux P et al. Histologic multi-focality is predictive of skin recurrences after conservingtreatment of stage I and II breast cancers. Breast CancerRes Treat 2001; 68 (1): 1-8.

27. Ragage F, Debled M, MacGrogan G et al. Is it useful todetect lymphovascular invasion in lymph node-positivepatients with primary operable breast cancer? Cancer2010; 116 (13): 3093-101.

28. Levine PH, Veneroso C. The epidemiology of inflamma-tory breast cancer. Semin Oncol 2008; 35 (1): 11-16.

29. Cristofanilli M, Valero V, Buzdar AU et al. Inflammatorybreast cancer (IBC) and patterns of recurrence: under-standing the biology of a unique disease. Cancer 2007;110 (7): 1436-44.

30. Giuliano AE, Hunt KK, Ballman KV et al. Axillary dis-section vs no axillary dissection in women with invasivebreast cancer and sentinel node metastasis: a randomizedclinical trial. JAMA 2011; 305 (6): 569-75.


INTRODUCTION

Salivary gland neoplasms are rare tumors and account for2-3% of all head and neck tumors. About 70% to 85% ofall salivary gland tumors are located in the parotid gland[1-5]. The majority (80%) of parotid masses are benignneoplasms with pleomorphic adenoma being the mostcommon type [1,3,5]. Parotidectomy is a common surgi-cal procedure for parotid tumors while some tumors canbe nonsurgically managed such as inflammatory masses,Warthin’s tumors that are not esthetically disturbing andsome malignant tumors (lymphoma and metastasis to theparotid gland).

Due to the possibility of major postoperative compli-cations such as facial nerve paresis or paralysis that can

occur in 10-30% of cases [6], especially with regard tothe treatment of benign tumors, the adequate extent ofsurgery is always the question.

The majority of the studies in the literature considersurgery as the main treatment option in the managementof parotid tumors. In our study, we presented two groupsof patients having a parotid mass; one managed surgical-ly and the other consisting of patients who didn’t under-go surgery in our center.

The objective of the study was to analyze the profile andthe management options of parotid tumors, and the diag-nostic value of fine needle aspiration cytology (FNAC) inour center.

METHODS

The study protocol was reviewed and accepted by theethics committee of Hôtel-Dieu de France hospital.

We reviewed retrospectively the medical and pathol-ogy records of all the patients who presented with a


AARRTTIICCLLEE OORRIIGGIINNAALL// OORRIIGGIINNAALL AARRTTIICCLLEEPAROTID TUMORS: IS SURGERY ALWAYS THE ONLY WAY?http://www.lebanesemedicaljournal.org/articles/63-4/original2.pdf

Chadi FARAH1, Claude GHORRA2, Charbel NASSIF1, Bassam TABCHY1, Amine HADDAD1

Farah C, Ghorra C, Nassif C, Tabchy B, Haddad A. Parotidtumors : Is surgery always the only way ? J Med Liban 2015 ;63 (4) : 179-184.

Farah C, Ghorra C, Nassif C, Tabchy B, Haddad A. Tumeursparotidiennes : La chirurgie est-elle toujours le seul traitement ?J Med Liban 2015 ; 63 (4) : 179-184.

RÉSUMÉ• Objectif de l’étude:Analyser les caractéristiquesdes tumeurs parotidiennes et leur prise en charge dans uncentre de soins tertiaires. Type d’étude : Étude de cohorterétrospective. Matériels et méthodes : Tous les cas de tu-meurs (ou masses) parotidiennes traités dans notre départe-ment entre le 1/1/1999 et le 31/12/2012, ont été inclus danscette étude. Les données démographiques, les caractéristi-ques cliniques, ainsi que les résultats histopathologiques et letype de prise en charge dans chaque cas ont été analysés. Enplus, la valeur diagnostique de la cytoponction à l’aiguille fineréalisée dans notre centre a été évaluée. Résultats : Parmiles 216 cas de tumeurs parotidiennes inclus dans l’étude, 164ont bénéficié d’une parotidectomie et 52 ont été traités non chi-rurgicalement ; 73,1% avaient une tumeur bénigne (tumeur deWarthin : 36,6%) et 16,7% une tumeur maligne. Dans notrecentre, la cytoponction à l’aiguille fine s’est révélée avoir unesensibilité pour rapporter une tumeur maligne de 71,4% et unespécificité d’éliminer la présence d’une tumeur maligne de100%. Conclusion: Dans notre série, la tumeur de Warthin aété la plus fréquente probablement en raison du taux élevé detabagisme. La prévalence de tumeurs malignes était relative-ment élevée. Les tumeurs malignes primaires de la parotide etles adénomes pléomorphes doivent toujours être traités chirur-gicalement. Par contre, les lymphomes, les métastases et lesmasses inflammatoires bénignes ainsi que certains cas detumeur de Warthin peuvent être traités médicalement.

ABSTRACT • Aim of the study : To analyze the characteris-tics and management of parotid tumors in a tertiary care cen-ter. Study design : Retrospective cohort study. Methods :All cases of parotid tumors (or masses) in our department be-tween Jan 1, 1999 and December 31st, 2012 were studied.Demographic data, clinical characteristics, histopathology andmanagement were analyzed. We also evaluated the diagnos-tic value of fine needle aspiration cytology (FNAC) in our cen-ter. Results: Of the 216 parotid tumors, 164 underwent pa-rotidectomy and 52 were not managed surgically; 73.1% hada benign tumor (36.6% had Warthin’s tumor) and 16.7% hada malignant one. In our center, FNAC was found to have asensitivity for reporting malignancy of 71.4% and a specificityto rule in malignancy of 100%. Conclusion : In our series,Warthin’s tumor was the most frequent mass probably relatedto the high tobacco use. The prevalence of malignant tumorswas relatively high in our series. Primary malignant tumorsand pleomorphic adenomas should always be treated surgi-cally, however, lymphomas, metastatic and benign inflamma-tory masses and cases of Warthin’s tumors could be man-aged nonsurgically.

Keywords : benign parotid tumors, malignant parotid tumors,parotidectomy, nonsurgical management, fine needle aspirationcytology

1 Department of Otolaryngology - Head and Neck, Hôtel-Dieude France University Hospital, St Joseph University Schoolof Medicine (HDF-USJ), Beirut, Lebanon.

2 Department of Pathology, HDF-USJ.Correspondence : Chadi Farah, MD.e-mail : [email protected]

180 Lebanese Medical Journal 2015 • Volume 63 (4) C. FARAH et al. – Management of parotid tumors

parotid mass to the Otolaryngology Department in Hôtel-Dieu de France Hospital, Beirut, Lebanon, a tertiary carecenter, between January 1st, 1999 and December 31st,2012.

These patients were divided into two groups in rela-tion with the different management options. We had agroup of tumors managed surgically and another groupnot managed surgically.

For the two groups we reviewed the medical andpathology records and noted the age, sex, tobacco use,clinical signs and symptoms and their duration, diagnos-tic tests, cytology results (if present), and location of theparotid tumor. For the surgical group we also noted thedate of the surgery, surgical technique, hospital stay, post-operative histopathology results, early and late complica-tions and recurrence.

Our results regarding epidemiology, clinical character-istics, diagnostic tests, management and follow-up, weretabulated for each group using Excel sheets. For the sur-gically managed group, we also calculated the percent-ages of the different postoperative complications and re-currence rates. As for the types of tumors, we noted theoverall percentages of benign and malignant ones and thespecific distribution of the different types in each group.

Concerning the diagnostic value of FNAC in our cen-ter, we used a 2 x 2 table to calculate the sensitivity forreporting malignancy, the specificity to rule in malignan-cy, the positive predictive value (PPV), the negative pre-dictive value (NPV) and the overall accuracy. Theseresults were calculated in the group of patients who un-derwent a FNAC in our center preoperatively, compar-ing the cytology results with the final histopathologyresults. All the non-diagnostic cytology results were notincluded.

RESULTS

PatientsFrom January 1st, 1999 to December 31st, 2012, 215 pa-tients complaining of a parotid mass presented to theOtolaryngology Department at Hôtel-Dieu de France hos-pital, a tertiary care center. One patient had a bilateralparotid mass that was managed in our center. Of the 216parotid masses studied, 164 were managed surgically and52 were not. In the study population, the age range wasbetween 2 and 92 years with male sex predominance.

In the nonsurgically managed group, 10 of the 52 pa-tients didn’t have a medical record as outpatients but allof them had pathology records and only 13 patients ofthe 42 were noted to have tobacco use.

The patients’ baseline characteristics of each group ofparotid masses are detailed in Table I.

Clinical presentationIn our study, the most common clinical presentation inthe two groups was the presence of a mass in the parotidregion (100%).

In 23 patients of the nonsurgically managed group,the duration of symptoms was not noted. The mean dura-tion of symptoms for the 19 remaining patients (withmedical records) was 34.6 months. In the surgicallymanaged group, the mean duration of symptoms wasapproximately 40 months. Table II shows the detailsrelated to clinical presentation in the two groups.

ManagementOf the 164 surgically managed parotid tumors in 163 pa-tients, 12 had surgeries elsewhere for the same tumor andwere reoperated in our center for recurrence or incom-plete resection. Four patients had bilateral parotid tumorsincluding one that was operated in our center in the studytime period and was included in the study. As of the otherthree patients: one was operated elsewhere, one wasoperated before 1999 in our center and the last one wasnot operated for the second tumor because he was diag-nosed as having a lymphoma. The mean postoperativehospital stay was 1.5 days.

Three patients were readmitted for total parotidecto-my after undergoing a superficial one and diagnosed ashaving a malignant tumor.

Seventy-three tumors had a preoperative pathologydiagnosis; 71 underwent a FNAC (only 31 in our center)and 2 had a revision of a previous pathology diagnosisdone elsewhere. Details related to other diagnostic testsare mentioned in Table III.

The most frequent surgical approach in our study wassuperficial parotidectomy (75%) followed by total (20.1%)and subtotal parotidectomy (4.3%) and one patient (0.6%)had an extensive mass that was managed by a debulking

TABLE IPATIENTS’ BASELINE CHARACTERISTICS

ManagementSurgical group Nonsurgical group

(n = 164) (n = 52)

Mean age (years) 51.1 54.1Sex ratio (M/F) 1.45 1.6Tobacco use (%) 51.2 31*Side: Right/Left 90/74 22/28 (2 unknown)

* 42 patients had medical records & tobacco use was noted for only 13 patients.

TABLE IICLINICAL PRESENTATION

Management

Surgical group Nonsurgical group(n = 164) (n = 52)

[only 42 had medical records]

Mean symptoms40 34.6 *

duration (months)Mass 164 (100%) 42 (100%)Recurrent swelling 1 (0.61%) 0Pain 14 (8.5%) 2 (4.76%)Facial weakness 9 (5.5%) 2 (4.76%)*Only in 19 cases of the 42 with records, the symptoms duration was noted.

C. FARAH et al. – Management of parotid tumors Lebanese Medical Journal 2015 • Volume 63 (4) 181

(Figure 1). Eighteen patients had a neck dissection, 10cases of malignant tumors received an adjuvant radiationtherapy, 1 an adjuvant chemotherapy and radiation thera-py and 1 only chemotherapy. One patient, who was diag-nosed with a metastasis of a papillary thyroid carcinoma,received postoperatively radioactive iodine therapy. All pa-rotidectomies were performed under intraoperative facialnerve monitoring (or nerve stimulation).

As for the location of the tumors in the surgical group;135 were in the superficial parotid lobe (82.3%) and 29were in the deep lobe (17.7%).

In the nonsurgically managed group, 50 patients hada FNAC in our center and the other 2 patients had a revi-sion of a previous pathology diagnosis done elsewhere.Three patients (2 pleomorphic adenomas and one atypi-cal cytology) didn’t return for surgery and the otherswere managed medically (inflammatory lesions, War-thin’s tumors, lymphomas, metastases). Pathology detailswill be discussed in the next section.

Of the 42 patients who had medical records, 20 hadonly a cervical CT scan, 4 only a cervical MRI, 4 a CTscan and a MRI and one had a PET scan.

Pathology resultsOf the 216 parotid tumors: 73.1% were benign withWarthin’s tumor being the most frequent one (36.6%),16.7% were malignant (4.2% lymphomas and 3.7% muco-epidermoid carcinomas), 5.1% consisted of an inflamma-

tory lesion and 5.1% had a normal or a nonspecificpathology result. The exact distribution of the pathologyresults for each group is detailed below (Table IV).

TABLE IIIDIAGNOSTIC TESTS in the SURGICALLY MANAGED GROUP

Surgically managed group (n = 164)

Fine needle aspiration cytology 71

Revision of pathology results 2

Cervical CT scan 110

Cervical MRI 42

Cervical ultrasound 15

PET scan 2

FIGURE 1. Surgical techniques

TABLE IVPATHOLOGY RESULTS in the TWO GROUPS of PAROTID TUMORS

Surgically Nonsurgicallymanaged managed

PATHOLOGY group (%) group (%)

Benign 131 (79.9) 27 (51.9)

Pleomorphic adenoma 55 (33.54) 3 (5.8)

Warthin’s tumor 58 (35.4) 21 (40.4)

Lymphoid tissue with follicular

hyperplasia (Warthin’s 0 2 (3.8)

tumor v/s lymph node)

Basal cell adenoma 3 (1.83) 0

Lymphoepithelial cyst 3 (1.83) 0

Epidermoid cyst 3 (1.83) 1 (1.9)

Lipoma 2 (1.2) 0

Cavernous hemangioma 1 (0.61) 0

Desmoid/Fibromatosis 1 (0.61) 0

Sebaceous cyst 1 (0.61) 0

Sabaceous lymphadenoma 1 (0.61) 0

Oncocytic lipoadenoma 1 (0.61) 0

Myoepithelioma 1 (0.61) 0

Neurofibroma plexiform 1 (0.61) 0

Malignant 29 (17.7) 7 (13.5)

Mucoepidermoid carcinoma 6 (3.7) 2 (3.8)

Salivary duct carcinoma 2 (1.2) 0

Basal cell carcinoma 0 1 (1.9)

Carcinoma ex pleomorphic 2 (1.2) 0

adenoma

Squamous cell carcinoma 3 (1.83) 0

Lymphoma 7 (4.3) 2 (3.8)

Cylindroma 2 (1.2) 0

(adenoid cystic carcinoma)

Acinic cell carcinoma 2 (1.2) 0

Epithelial myoepithelial1 (0.61) 0

carcinoma

Rhabdomyosarcoma 2 (1.2) 0

Carcinomatous lymphangitis 1 (0.61) 0

Metastasis 1 (0.61) 2 (3.8)

Inflammatory 2 (1.2) 9 (17.3)

Non specific inflammation/2 (1.2) 2 (3.8)

granulation tissue

Abscess 0 5 (9.6)

Granuloma 0 2 (3.8)

No specific lesion/Normal tissue 2 (1.2) 9 (17.3)

TOTAL 164 (100) 52 (100)

Number of parotid tumors


Diagnostic value of FNACThirty-one patients underwent a preoperative FNAC in ourcenter; 4 results (13%) were not diagnostic and 27 werediagnostic. In order to determine the diagnostic value ofFNAC in our center we took into consideration only thediagnostic results and we compared them with the finalhistopathology results. We had no false positive results andtwo false negative ones: one considered as a benign lym-phoid inflammatory reaction on cytology was found to bea lymphoma, and one considered as pleomorphic adenomawas found to be a cylindroma (Table V). The sensitivity forreporting malignancy was 71.4 % and the specificity torule in malignancy was 100%. PPV was 100%, NPV was90.9% and the overall accuracy was 92.6%.

Follow-upIn the surgically managed group, of the 164 cases in thisgroup, 27 were lost to follow-up and 137 were followedfor a mean period of 18.8 months.

The most frequent postoperative complication wasfacial weakness in 26 cases (15.8%); 8 presented a totalfacial palsy and 18 a facial paresis. Seven of these caseshad already preoperatively a facial weakness which per-sisted postoperatively and 17 (10.4%) had a newly onsetnerve dysfunction. Two cases who presented with facialweakness preoperatively had a complete resolution oftheir symptoms in the immediate postoperative period.(Figure 2). Six patients of the 24 who had a persistentfacial weakness were noted to have a total recovery dur-ing the follow-up.

The second most common complication was the sali-vary fistula found in 12 patients (8.8%) followed by areported Frey’s syndrome in 7 patients (5.1%) and a sali-vary collection in 5 patients (3.6%). The recurrence ratein this group was 2.2% (3 patients of the 137 cases whohad a follow-up), where one patient who had a cav-ernous hemangioma had a recurrence 11 years later andwas reoperated in our center, another one who had apleomorphic adenoma had a recurrence 11 years laterand didn’t return for surgery and the last one, who had anonspecific lesion on the pathology result, presented a

recurrence of his mass two years later and was operatedby a plastic surgeon in our center with a final diagnosisof pleomorphic adenoma.

In the nonsurgically managed group, and of the 42 pa-tients who had medical records, 17 were lost to follow-upand only 25 patients were followed for a mean period of14.8 months. Other than the three patients who didn’treturn for surgery, all the cases of lymphoma and meta-stasis (Table IV) were referred to the oncology and radi-ation therapy departments. As for the two cases of muco-epidermoid carcinomas (Table IV), one didn’t have amedical record and the other was lost to follow-up andthe 3rd case of pleomorphic adenoma didn’t have a med-ical record. The two cases diagnosed with a granuloma-tous disease were referred to the internal medicinedepartment.

DISCUSSION

In our study, we reviewed 216 cases of parotid tumorsover a 14-year period with two different managementapproaches in a tertiary care center. The majority of ourpatients were males with an age range of 2 to 92 years.

Parotid gland tumors typically present with a palpablemass, recurrent swelling, pain or facial nerve dysfunction.In our study, the most common clinical presentation was aslowly growing mass followed by pain and facial weak-ness, which is consistent with the literature results [4,7-8].

Parotid gland surgery nowadays is established asstandard therapy throughout the world. Superficial andtotal parotidectomy are recognized as a standard proce-dure for removal of parotid gland tumors [9].

TABLE VDIAGNOSTIC VALUE of FNAC*

Malignant Benign Totalhistopathology histopathology

Malignant FNAC 5 0 5Benign FNAC 2 20 22Total 7 20 27*4 cytology results were not diagnostic FNAC: fine needle aspiration cytology

FIGURE 2. Facial weakness

C. FARAH et al. – Management of parotid tumors Lebanese Medical Journal 2015 • Volume 63 (4) 183

In our surgical group, parotid tumors were most fre-quently found in the superficial lobe (82.3%) and the mostfrequent surgical approach was superficial parotidectomy(75%) followed by total parotidectomy in 20.1% of casesand subtotal parotidectomy in 4.3% of cases.

According to the literature, the preferred surgical man-agement of benign parotid tumors confined to the super-ficial lobe is the partial superficial parotidectomy, whilecomplete superficial parotidectomy or total parotidec-tomy were performed in cases where tumors were largeor recurrent tumors with involvement of the deep lobe[1,3,9]. Partial superficial parotidectomy was found toreduce postoperative morbidity especially facial nervedysfunction and Frey’s syndrome [9].

As for the malignant tumors, management requires acombination of treatment modalities. According to thestage and grade of the tumor, we have a choice of parotid-ectomy alone if the tumor was in an early stage and a lowgrade or parotidectomy plus adjuvant radiation therapyin the early stage and high grade or parotidectomy plusneck dissection and adjuvant radiation therapy in theadvanced stages [1]. In our study, malignant tumors weremanaged according to the stage and grade of the tumorwith combination of treatment modalities.

In the group of patients who were not managed surgi-cally, the majority had Warthin’s tumor (40.4%) that wasnot esthetically disturbing and the others were diagnosedof having lymphomas, metastases and inflammatory le-sions that were better managed medically. Six patientsshould have had a surgical treatment (3 had primarymalignant tumors and 3 had pleomorphic adenomas) but3 of them didn’t return for surgery and the other 3 werelost to follow-up.

Although pleomorphic adenoma of the parotid glandis considered as a benign disease, it should be managedsurgically because of the possibility of malignant trans-formation (carcinoma ex pleomorphic adenoma) anddistant metastases [10].

In a review of the literature, we found that most ofthe parotid tumors were benign accounting for approxi-mately 60.6 to 80% with pleomorphic adenoma beingthe most frequent one followed by Warthin’s tumor [1,3,5]. Twelve to seventeen percent of the tumors weremalignant [11] with mucoepidermoid carcinoma beingthe most frequent (3-4% of all tumors) [1,3,7,12]. Lym-phoma was reported to be found in 1 to 5% of parotidtumors [13-14]. In our study, the results of benign tu-mors were consistent with the literature (79.9% in thesurgical group) but we have found that Warthin’s tumorwas the most frequent tumor even in the surgically man-aged group (35.4%). This result is probably related tohigh tobacco use (51.2%) and male predominance inthis group, as it was shown by Cardoso et al. [15]. Asfor the malignant tumors, we have found a number inthe high reported ranges (17.7%) in the surgical groupwith a slight predominance of lymphoma (4.3% of allthe tumors and 24% of the malignant tumors in the sur-gical group) over mucoepidermoid carcinoma (3.7%)

but still consistent with the literature ranges.FNAC is considered by many authors as a safe and

accurate diagnostic test that helps the surgeon in hismanagement plan. Throughout the literature, FNAC isfound to have a sensitivity for reporting malignancy of64-99% and a high specificity to rule in malignancy of96.3-100% and an overall accuracy of 81-98% [1,3,16-17]. In our series, 31 patients had a preoperative FNACin our center with 13% of nondiagnostic results. Thesensitivity was 71.4% but the specificity was found to be100% (no false positive results) and the overall diagnos-tic accuracy was 92.6%. In our center, FNAC can be con-sidered as a reliable tool to rule in malignancy and to helpthe surgeon in his management plan.

Although FNAC is not reliable to diagnose lymphoma,it could guide the surgeon, when finding lymphoid cells inthe specimen, toward nonradical management of thesetumors (a biopsy instead of a parotidectomy).

Parotidectomy carries the risk of postoperative com-plications affecting the patient’s quality of life.

Facial nerve dysfunction is considered as the majorand the most frequent complication found transiently in10-68% of cases and permanently in 0-19% of cases[9,18]. In our experience it was encountered in 15.8% ofcases with only 10.4% of newly onset postoperativenerve dysfunction and 11.6% with permanent dysfunc-tion over the follow-up period.

Frey’s syndrome is always underestimated and is usu-ally noted if the patient complains. The reported Frey’ssyndrome in the literature accounts for 2 to 96% [9,19].In our series, we had only 5.1% of Frey’s syndromereported by the patients.

Salivary fistula (8.8%) and salivary collection (3.6%)were also found in our series and were consistent withthe literature results [9].

The use of a more conservative surgery especially forbenign tumors and intraoperative monitoring (or stimu-lation) of facial nerve, in our experience and in the liter-ature [20], helped us reduce the risk of postoperativecomplications especially nerve dysfunction.

In our study, we had a low recurrence rate (2.2%) withno malignant tumor recurrence in a mean follow-up periodof 18.8 months which is consistent with the published rates(0-12%) depending on extent of surgery and tumor histol-ogy [9], and even in some studies, 40 to 45% recurrencerates were noted for benign and malignant tumors [1,5,21].

Our study had some limitations regarding its design(retrospective) carrying the problem of information col-lection from nonstandardized medical records. We triedto limit the effect of this bias by collecting each patient’sinformation from three different sources; the hospitalmedical records, the private clinics medical records andthe pathology records. On another hand, in our series,sur-geries were mainly performed by two experiencedhead and neck surgeons but the fact that all of them weremade under facial nerve monitoring and in merely thesame surgical techniques according to standardized indi-cations, reduced the surgical technique-related bias.


CONCLUSION

Parotid tumors are rare neoplasms with a predominanceof benign over malignant ones. In our series, we foundWarthin’s tumor to be the most frequent tumor in the twogroups related most probably to high tobacco use andmale sex predominance. Malignant tumors were foundto be in the high reported ranges.

In our experience, parotid tumors were managed sur-gically or nonsurgically according to the tumor’s type andpatient’s preferences. The use of a surgical approach adapt-ed to the tumor’s type and extent and the use of intraoper-ative facial nerve monitoring have shown a reduced risk ofpostoperative complications and recurrence rate.

In our center, FNAC was found to be an effective toolin the diagnosis and treatment planning of parotid tu-mors and should be more frequently performed in orderto reduce non necessary radical treatments.

Although we tend to think of primary malignant neo-plasms in the context of a parotid gland mass, whichshould be treated radically, we should keep in mind the pos-sibility of having a lymphoma to guide our managementtoward a simple biopsy instead of a parotidectomy.

Moreover, benign inflammatory masses or metastaticmasses in the parotid gland should be treated medicallyby a multidisciplinary team.

Finally, we should also consider a nonsurgical man-agement of some cases of Whartin’s tumor especially inelderly patients having comorbidities and with no esthet-ical disturbance in order to avoid anesthetic and surgicalrisks in this special population.

ACKNOWLEDGEMENTS

This work was presented at the 10th Annual Middle EastUpdate in Otolaryngology Conference at Dubai, UnitedArab Emirates on April 21-23, 2013.

We would like to thank Simon Rassi, MD, and NaylaMatar, MD, for their contribution to our study by givingus valuable advice regarding the form and the contentsof this paper and access to their patients’ medical records.

REFERENCES

1. Lin CC, Tsai MH, Huang CC, Hua CH, Tseng HC, Huang ST.Parotid tumors: a 10-year experience. American Journalof Otolaryngology 2008; 29 (2): 94-100.

2. Spiro RH. Salivary neoplasms: Overview of a 35-yearexperience with 2,807 patients. Head & Neck Surgery1986; 8 (3): 177-84.

3. Upton DC, McNamar JP, Connor NP, Harari PM, HartigGK. Parotidectomy: ten-year review of 237 cases at asingle institution. Otolaryngology-Head & Neck Surgery2007; 136 (5): 788-92.

4. Mohammed F, Asaria J, Payne RJ, Freeman JL. Retro-spective review of 242 consecutive patients treated surgi-cally for parotid gland tumours. Otolaryngology-Head &Neck Surgery 2008; 37 (3): 340-6.

5. Bradley PJ. Pleomorphic salivary adenoma of the parotidgland: which operation to perform? Current Opinion inOtolaryngology & Head and Neck Surgery 2004; 12 (2):69-70.

6. Silvoniemi A, Pulkkinen J, Grénman R. Parotidectomy inthe treatment of pleomorphic adenoma - analysis of long-term results. Acta Oto-Laryngologica 2010; 130 (11):1300-5.

7. Lee YM, Choi HJ, Kim JW, Kim JH. Parotid glandtumors in a Korean population. Journal of CraniofacialSurgery 2012; 23 (3): e205-e209.

8. Sungur N, Akan ÌM, Ulusoy MG, Özdemir R, Kilinç H,Ortak T. Clinicopathological evaluation of parotid glandtumors: a retrospective study. Journal of CraniofacialSurgery 2002; 13 (1): 26-30.

9. Koch M, Zenk J, Iro H. Long-term results of morbidityafter parotid gland surgery in benign disease. The Laryn-goscope 2010; 120 (4): 724-30.

10. Andry G, Hamoir M, Locati LD, Licitra L, Langendijk JA.Management of salivary gland tumors. Expert Review ofAnticancer Therapy 2012; 12 (9): 1161-8.

11. Papadogeorgakis N, Goutzanis L, Petsinis V, Alexandridis C.Management of malignant parotid tumors. Oral and Maxil-lofacial Surgery 2012; 16 (1): 29-34.

12. Ungari C, Paparo F, Colangeli W, Iannetti G. Parotidglands tumours: overview of a 10-years experience with282 patients, focusing on 231 benign epithelial neoplasms.European Review for Medical and PharmacologicalSciences 2008; 12 (5): 321-5.

13. Mehle ME, Wood BG, Tucker HM, Lavertu P, Tubbs R,Kraus DH. Lymphoma of the parotid gland. The Laryn-goscope 1993; 103 (1): 17-21.

14. Tiplady C, Taylor P, White J, Arullendran P, Proctor S.Lymphoma presenting as a parotid tumour: a population-based study of diagnosis, treatment and outcome onbehalf of the Scotland and Newcastle Lymphoma Group.Clinical Oncology 2004; 16 (6): 414-19.

15. Cardoso S, do Nascimento Souza KC, de Faria P et al.Warthin’s tumor at the Brazilian National Cancer Insti-tute: Additional evidence of homogeneous sex preva-lence and association with other neoplasms. ORL 2008;70 (6): 339-43.

16. Awan MS, Ahmad Z. Diagnostic value of fine needleaspiration cytology in parotid tumors. JPMA 2004; 54(12): 617-19.

17. Piccioni L, Fabiano B, Gemma M, Saranria D, Bussi M. Fine-needle aspiration cytology in the diagnosis of parotid le-sions. Acta Otorhinolaryngologica Italica 2011; 31 (1): 1-4.

18. Yuan X, Gao Z, Jiang H et al. Predictors of facial palsyafter surgery for benign parotid disease: Multivariateanal-ysis of 626 operations. Head & Neck 2009; 31 (12):1588-92.

19. Singh N, Kohli M, Kohli H. Innovative technique toreduce incidence of Frey’s syndrome after parotidsurgery. The American Surgeon 2011; 77 (3): 351-4.

20. Olsen KD, Daube JR. Intraoperative monitoring of thefacial nerve: an aid in the management of parotid glandrecurrent pleomorphic adenomas. The Laryngoscope1994; 104 (2): 229-32.

21. Zbären P, Tschumi I, Nuyens M, Stauffer E. Recurrentpleomorphic adenoma of the parotid gland. The AmericanJournal of Surgery 2005; 189 (2): 203-7.

INTRODUCTION

Unfractionated heparin (UFH) plays a major role as anti-thrombotic agent in the treatment of acute coronary syn-drome (ACS) [1-3]. However, its use is associated withan unpredictable anticoagulant effect requiring frequentblood testing to adjust the dose, a tendency to rebound anda risk of thrombocytopenia. Thus, low molecular weight

heparins (LMWH), particularly enoxaparin, became morepopular, replacing UFH in many indications and manycenters around the world. In fact, enoxaparin has a higherbioavailability, is easily administered, and has more pre-dictable dose-response characteristics with no requiredlaboratory monitoring. Furthermore, enoxaparin has beenshown to be more effective than UFH in preventing coro-nary events in patient with ACS [4-6].

LMWHs duration of activity is longer than that ofUFH and only partially neutralized by protamine sulfate[6-7], with a potential risk of increased bleeding whenused around the time of a coronary artery bypass graft(CABG) surgery.

In this paper, we describe our experience with the useof enoxaparin prior to CABG, and its effect on bleeding


Abi Ghanem M, Saliba E, Irani J, Darwish R, Kawkabani N,Yared T, Naim C, Abu Khalil B. Effects of preoperative enoxa-parin on bleeding after coronary artery bypass surgery. J MedLiban 2015 ; 63 (4) :185-190.

Abi Ghanem M, Saliba E, Irani J, Darwish R, Kawkabani N,Yared T, Naim C, Abu Khalil B. Effets de l’énoxaparine reçueen préopératoire sur le saignement après chirurgie de pontagecoronarien. J Med Liban 2015 ; 63 (4) : 185-190.

Departments of Cardiothoracic Surgery1, Cardiac SurgeryAnesthesia3, Cardiology4, St Georges Hospital-University ofBalamand Medical Center, Beirut, Lebanon.

2Faculty of Medicine & Medical Sciences, University of Bala-mand, Beirut, Lebanon.

Correspondence: Jihad Irani, MD.email: [email protected]

ABSTRACT • Background : Low molecular weight heparinsare replacing unfractionated heparin in practice prior to cardiacsurgery. This study examines postoperative (post-op) bleedingindicators in patients who received enoxaparin and under-went elective isolated first time coronary artery bypass graft.Methods : A total of 125 consecutive patients who underwentthis procedure between 2009 and 2011 at one tertiary centerwere reviewed and divided into three groups : Group A (n = 50)received the last dose of enoxaparin between 12 and 24 hoursbefore surgery, Group B (n = 25) received the last dose before24 hours and Group C (n = 50) did not receive enoxaparin.Perioperative bleeding indicators and transfusion rates werecompared. Results : Preoperative patients’ characteristicswere comparable between the three groups. There were noperioperative deaths, return to the operating room for any rea-son, nor major bleeding. Post-op bleeding indicators were sim-ilar in the three groups. The average chest tube drainage at24 hours post-op was 880 mL, 695 mL and 830 mL in Group A,B and C respectively (p = 0.71). Transfusion rates of red bloodcells were not statistically different (Group A 56%, B 64% &C 62% ; p = 0.747). In multivariate analysis, female gender,older age, and preoperative clopidogrel intake (stopped 5 daysprior to surgery) were associated with higher transfusion rates.Conclusion : In elective first time coronary artery bypass graftpatients who had no aspirin or clopidogrel intake 5 days prior tosurgery, the use of enoxaparin up to 12 hours prior to skin inci-sion does not increase the risk of post-op bleeding.

Keywords : enoxaparin, coronary artery bypass, postop-erative hemorrhage

RÉSUMÉ•Contexte:Les héparines de bas poids moléculaireremplacent dans la pratique l’héparine non fractionnée avant lachirurgie cardiaque. Cette étude examine les indicateurs hé-morragiques postopératoires chez les patients ayant reçu del’énoxaparine et subi un pontage coronarien électif isolé pour lapremière fois. Méthodes : 125 patients consécutifs ayant subicette procédure entre 2009 et 2011 dans un centre tertiaire ontété examinés et divisés en trois groupes: groupe A (n = 50) ontreçu la dernière dose d’énoxaparine entre 12 et 24 h avant lachirurgie, groupe B (n = 25) ont reçu la dernière dose avant24 h, et le groupe C (n = 50) n’ont pas reçu d’énoxaparine. Lesindicateurs de saignements peropératoires et les taux de trans-fusion ont été comparés. Résultats : Les caractéristiques despatients préopératoires étaient comparables entre les troisgroupes. Pas de décès per opératoires et pas de retour au blocopératoire pour hémorragie majeure ou autre raison quel-conque. Les indicateurs de saignements postopératoires étaientsimilaires dans les trois groupes. Le drainage thoracique moyenà 24 heures postopératoires était de 880 ml, 695 ml et 830 mldans les groupes A, B et C respectivement (p = 0,71). Les tauxde transfusion de globules rouges n’étaient pas statistiquementdifférents (groupe A 56%, B 64% et C 62%, p = 0,747). Uneanalyse multivariée a montré que le sexe féminin, l’âge, et leclopidogrel en préopératoire (arrêté 5 jours avant la chirur-gie) ont été associés à des taux plus élevés de transfusion.Conclusion : Chez les patients opérés d’un pontage corona-rien électif pour la première fois et qui n’avaient pas prisd’aspirine ou de clopidogrel 5 jours avant la chirurgie, l’utilisa-tion de l’énoxaparine jusqu’à 12 heures avant l’incision de lapeau n’augmente pas le risque d’hémorragie postopératoire.

AARRTTIICCLLEE OORRIIGGIINNAALL// OORRIIGGIINNAALL AARRTTIICCLLEEEFFECTS OF PREOPERATIVE ENOXAPARIN ON BLEEDING AFTER CORONARYARTERY BYPASS SURGERYhttp://www.lebanesemedicaljournal.org/articles/63-4/original3.pdf

Moussa ABI GHANEM1, Elie SALIBA1, Jihad IRANI2, Roula DARWISH3, Nadine KAWKABANI3

Toni YARED1, Charbel NAIM4, Bassam ABU KHALIL1

in the postoperative (post-op) phase in general and accord-ing to its stopping time preoperatively.

MATERIAL AND METHODS

We retrospectively reviewed consecutive patients whounderwent elective first time CABG at our tertiary cen-ter between 2009 and 2011. The study was approved bythe Institutional Review Board (IRB) at the Universityof Balamand.

We included all first time isolated CABG patients andexcluded those who had hematologic diseases, elevatedcreatinine (Cr) levels (Cr > 1.2), those on heparin andthose who were on anti-platelet therapy in the last fivepreoperative (pre-op) days. Patients with elevated cre-atinine were excluded since, according to the manufac-turer, enoxaparin associated risk of bleeding is increasedwith the degree of renal impairment.

We divided eligible patients into three groups: GroupA received enoxaparin until 12 to 24 hours beforesurgery, Group B received the last dose of enoxaparinmore than 24 hours before surgery and Group C consist-ed of patients who did not receive enoxaparin at all (con-trol). The dose of enoxaparin given was 1 mg/kg twice aday. The enoxaparin used in our center was Lovenox®

(Sanofi-Aventis, Canada).We collected the following data on all eligible pa-

tients: baseline hematologic characteristics, creatininelevel, duration of enoxaparin treatment, time betweenthe last enoxaparin dose and surgery and pre-op use oforal anti-platelets agents (aspirin, clopidogrel). Dataregarding the CABG procedure, that was undertaken bythe same senior surgical team, included the duration ofcardiopulmonary bypass (CPB), number of grafts, acti-vated clotting time (ACT) and the volume of blood andblood products transfused during the operation. Hema-tologic and transfusion data collected included perioper-ative hematocrit (Hct) measurement as well as the quan-tity of packed red blood cells (PRBC), fresh frozen plasma(FFP) and platelets administered intraoperatively andduring the first 72 hours after surgery.

All CABG procedures were performed by the samesurgical team, including the general anesthesia, radial ar-tery cannulation, median sternotomy, central cannula-tion, non-pulsatile CPB with moderate systemic hypo-thermia (32-33°C) and antegrade and retrograde coldblood cardioplegia. All patients received tranexamic acid10 mg/kg at induction of the general anesthesia and then250 mg on CPB. Anticoagulation was established by astandard dose of UFH (4 mg/kg) via the central line.Additional UFH was given in order to maintain an ACTgreater than 480 sec. After the discontinuation of bypasssupport, UFH was reversed by administration of 1 mgprotamine per 100 U of UFH to achieve an ACT of 130-140 sec. The cardiopulmonary bypass prime was 1.55 L:0.5 L of sodium bicarbonate 1.4%, 0.5 L of Haes sterile6% (Fresenius Kabi, India), 0.5 L of Ringer (Taj pharma-ceuticals, India) and 0.05 L of albumin 20%.

In post-op, we collected the chest tube output for thefirst 24 and 48 hours in addition to the amount of bloodproducts administration in the first 72 hours. All chesttubes were removed systematically on day 2 post-op.Patients were given PRBC to keep their Hct level above25%. FFP was given for fibrinogen levels below 100 orfor an international normalized ratio (INR) more than1.7. Platelets were given for platelet count below 50000.Post-op hematologic characteristics were also recorded.

Statistical AnalysisAt first, we used means, medians and standard devia-tions (SD) to describe our quantitative data, and propor-tions to describe our qualitative data. We used tradition-al statistical tests to compare between variables, Studentt-test, ANOVA for normally distributed variables andMann-Whitney and Kruskall-Wallis tests when appro-priate. We used chi-square test to compare qualitativedata, and Fisher exact test when appropriate. We ran alogistic regression model to analyze the factors associat-ed with the outcome variable ‘need for PRBC transfu-sion’, and we included in the model all the variables thatwere associated to this outcome in the bivariate analysis,in addition to the variable ‘group belonging’. Model cal-ibration was assessed using Hosmer-Lemeshow test. Thec-statistic was calculated as a measure of discrimination.All p-values below 0.05 were considered statisticallysignificant. We used SPSS v16.0 and MedCalc v 12.3 forstatistical analysis.

186 Lebanese Medical Journal 2015 • Volume 63 (4) M. ABI GHANEM et al. – Preoperative enoxaparin in CABG and bleeding

FIGURE 1Total number of patients included

and excluded in the study

166 patientsundergoing

isolated CABG

Excluded Included

29 patients(Cr > 1.2)

5 patients(UFH used within 12

hours of the surgery)

125 patients

3 patientsRedo surgery

3 patientsOff pump CABG

1 patient(History of coagulopathy

before surgery)

RESULTS

Among 166 patients who had the procedure betweenJune 2009 and October 2011, 125 met our selection cri-teria (Figure 1).

We analyzed the data on all 125 patients. Collectedpatients’ demographics (such as age, sex, etc.) wereidentical and equally distributed between the threegroups (Table I). The average age of the studied samplewas 65.8 years. Most of our patients were male (80%).Base-line hematocrit, platelets count, creatinine levels,and diabetic status were comparable between the differ-ent groups. The prothrombin time (PT) and partial

thromboplastin time (PTT) were similar, as well as theaverage bypass time. The pre-op ACT was slightly moreelevated in the Group A compared to Group B (mean 146vs. 139, p = 0.010 post-hoc analysis).

Seventy-five patients (60%) received one or moreunit of PRBC during the observed period. Only 17 pa-tients (13.6%) needed FFP, and 10 patients (8%) neededplatelet transfusion. In order to estimate the severity ofbleeding, we analyzed the “need for PRBC transfusion”as an outcome variable. In the three groups, the need forPRBC transfusion was 56% in Group A (28 patients),64% in Group B (16 patients) and 62% in Group C(31 patients) (p = 0.747; Table II).

M. ABI GHANEM et al. – Preoperative enoxaparin in CABG and bleeding Lebanese Medical Journal 2015 • Volume 63 (4) 187

There were no significant differences in PRBC trans-fusion nor in chest tube drainage post-op among thethree groups (Table II), (Figure 2).

The following factors were significantly associatedwith receiving more PRBC: being a female, older ageand having taken clopidogrel in pre-op independently ofwhat group they belong to (Table III). Eighty-eight per-cent of female patients needed a transfusion versus 53%of male patients (p = 0.001). The mean age of patientsreceiving PRBC transfusion was 68 years compared to amean of 62 years in those who did not receive any PRBCtransfusion (p < 0.001). Of those patients who were onclopidogrel in pre-op (26 patients - stopped 5 days pre-opon average), 85% needed a PRBC transfusion comparedto 53% of those who never had clopidogrel (p = 0.004).

After performing a logistic regression analysis, in-cluding in the model the following variables: sex, age,aspirin receipt in pre-op, clopidogrel in pre-op, pre-opHct and UFH in per-op (Table IV), the factors ‘age’,‘sex’ and ‘prior use of clopidogrel’ remained statistical-ly associated with the outcome ‘need for PRBC’.

When examining drains’ output of the first 24 hourspost-op, the following factors were found to be associat-ed with a higher output: number of bypasses (more thantwo grafts) and being a male (Table V).

DISCUSSION

Anticoagulation has been considered to be the corner-stone of the management of ACS [1]. Since the 1970’s,UFH was shown to benefit patients with acute myocar-dial infarction by exerting an activity against factor Xaand thrombin [1-2].

New formulation of LMWH, which have a greateractivity against Xa, have been used in treating ACS [6].In contrast to UFH, these molecules produce more pre-dictable anticoagulation response, have a better bioavail-ability and do not require laboratory monitoring [6].enoxaparin, which is a LMWH, has shown to be supe-rior to UFH in the treatment of ACS, however its long

duration of activity and its partial neutralization by pro-tamine have raised concerns regarding bleeding compli-cations in patients undergoing CABG treated by LMWHprior to surgery [6-7].

According to the latest guidelines concerning perioper-ative management of antithrombotic therapy published bythe American College of Chest Physicians (ACCP) in2012 [8], it is suggested continuing aspirin around thetime of surgery in patients requiring CABG instead ofstopping it 7 to 10 days before, and administering the lastdose of LMWH 24 hours before surgery instead of 12hours before. Those recommendations are grade 2C whichmeans very weak recommendations with low-quality evi-dence; that’s why those recommendations remain a sug-


Group

Units of PRBC received per group

FIGURE 2. Amount of PRBC transfused per group

enox 12-24h enox > 24h None

None 1-2 units 3-4 units 5 units and more

gestion and other alternatives may be equally reasonable. In order to elucidate this matter, many reports studied

the incidence of bleeding complications in patientsundergoing CABG treated prior to surgery by LMWH.

Some of these studies showed no significant risk ofbleeding in those patients compared to subjects receiv-ing UFH [9-10]. In fact, Renda et al. found that trans-fusional requirements of blood and plasma was similarin patients receiving either UFH or LMWH (stopped 12hours prior to surgery) but their study had sample sizelimitation [9]. Medalion et al. found also that adminis-tration of enoxaparin more than 8 hours before surgerydid not increase bleeding or products’ transfusion rate [10].

However, these findings contrast with what Jones et al.found: a significant increase in incidence of re-explorationfor post-op bleeding in a group of patient receiving enoxa-parin [11]. Their study was done on a small size of patientswho had some significant differences in the baseline char-acteristics between the two groups in relation to age, sexand type of surgery. It showed a lack of correlationbetween blood products usage and increased risk of surgi-

cal re-operation, and failed to describe any temporal rela-tionship between the last dose of enoxaparin and post-opbleeding. The increased risk of bleeding and mediastinalre-exploration in a group treated by LMWH was also men-tioned by two studies: a study by Myhre et al. [12] in 2004and a study by McDonald et al. [13] in 2005. Kincaid et al.showed also that the preoperative use of enoxaparin lessthan 12 hours before CABG is associated with lower post-op hemoglobin values and higher rates of transfusion whencompared to continuous use of UFH [14].

In our study, we showed that the use of enoxaparin inpre-CABG does not affect some bleeding parameters,

M. ABI GHANEM et al. – Preoperative enoxaparin in CABG and bleeding Lebanese Medical Journal 2015 • Volume 63 (4) 189


when stopped earlier than 12 hours prior to the surgery.Enoxaparin is usually administered twice daily and

since our patients had all elective surgery, we stoppedenoxaparin at two different intervals prior to surgery. Inone of the groups, enoxaparin was stopped 12 hoursprior to surgery while in the other it was stopped 24hours prior to surgery. We found that whether usingenoxaparin and stopping it late (12-24 h prior) or early(24 h prior) did not affect bleeding parameters (tubedrainage and transfusion rate) in comparison to the con-trols. Our series showed lower bleeding rates in gen-eral and that could be related to the use of tranexamicacid.

All patients who received enoxaparin did not havepre-op higher PT, PTT or ACT levels than the controlgroup. There is no data in the literature about heparinresistance and pre-op use of enoxaparin.

The heparin use on bypass in this study was similarbetween all groups (p = 0.258; Table I). This impliesabsence of any heparin resistance, even though it is doc-umented in the literature in around 26% of the cases[15]. Based on our findings, administering enoxaparinup to 12 hours prior to surgery appears safe.

The findings that the number of bypasses grafts andbeing a male increase chest tube drainage is interestingbut not related to the use of enoxaparin (Table V).

On the other hand a multivariate analysis showed thatfemale gender, age and administration of clopidogrelmore than five days prior to surgery increase the needfor transfusion. This also is not related to the use ofenoxaparin, even though they are important findings andneed to be affirmed in larger studies.

The fact that being of female gender was an indicatorof increased need for transfusion with however lowerabsolute amounts of chest tube drainage is intriguing, butcan be explained when looking to their smaller body sur-face area (BSA). In fact the female patients had a lowermean BSA (1.75) than male patients (1.91) and this mayexplain the higher transfusion rate in them even thoughthey had lower chest tube drainage.

Our study has its own limitations; it is a one-centerretrospective study, reflecting the practice of one groupwhich might affect the generalizability of the findings.The sample size is relatively small and that might affectthe ability to detect differences that really exist (type-2error). Large multicenter studies would be required inorder to shed more lights on this issue.

CONCLUSION

In elective first time CABG patients who had no aspirinor clopidogrel intake five days prior to surgery, the useof enoxaparin up to 12 hours prior to skin incision is safeand does not increase the risk of bleeding postoperative-ly or the transfusion rate.

However, the use of clopidogrel (more than five daysprior to surgery), age and being a female may predict ahigher transfusion rate.

REFERENCES

1. Telford AM, Wilson C. Trial of heparin versus atenolol inprevention of myocardial infarction in intermediate coro-nary syndrome. Lancet 1981; 1: 1225-8.

2. Théroux P, Ouimet H, McCans J et al. Aspirin, heparin,or both to treat acute unstable angina. N Engl J Med1988; 319: 1105-11.

3. Oler A, Whooley MA, Oler J et al. Adding heparin toaspirin reduces the incidence of myocardial infarction anddeath in patients with unstable angina: a meta-analysis.JAMA 1996; 276: 811-15.

4. Antman EM, McCabe CH, Gurfinkel EP et al. Enoxa-parin prevents death and cardiac ischemic events in unsta-ble angina/non-Q-wave myocardial infarction. Results ofthe thrombolysis in myocardial infarction (TIMI) 11Btrial. Circulation 1999 Oct 12; 100 (15): 1593-601.

5. Goodman SG, Cohen M, Bigonzi F et al. Randomizedtrial of low molecular weight heparin (Enoxaparin) ver-sus unfractionated heparin for unstable coronary arterydisease: One-year results of the ESSENCE study. J AmColl Cardiol 2000; 36: 693-8.

6. Weitz JI. Low-molecular-weight heparins. N Engl J Med1997; 337: 688-98.

7. Holst J, Lindblad B, Bergqvist D et al. Protamine neutral-ization of intravenous and subcutaneous low-molecular-weight heparin (tinzaparin, logiparin). An experimentalinvestigation in healthy volunteers. Blood Coagulation &Fibrinolysis 1994; 5: 795-803.

8. Douketis JD, Spyropoulos AC, Spencer FA et al. Peri-operative management of antithrombotic therapy: Anti-thrombotic therapy and prevention of thrombosis, 9th ed:American College of Chest Physicians evidence-basedclinical practice guidelines. Chest 2012; 141 (2 suppl):e326S-e350S.

9. Renda G, Di Pillo R, D’Alleva A et al. Surgical bleed-ing after pre-operative unfractionated heparin and lowmolecular weight heparin for coronary bypass surgery.Haematologica 2007; 92: 366-73.

10. Medalion B, Frenkel G, Patachenko P et al. Preoperativeuse of enoxaparin is not a risk factor for postoperativebleeding after coronary artery bypass surgery. J ThoracCardiovasc Surg 2003 Dec; 126 (6): 1875-9.

11. Jones HU, Muhlestein JB, Jones KW et al. Preoperativeuse of enoxaparin compared with unfractionated heparinincreases the incidence of re-exploration for postopera-tive bleeding after open-heart surgery in patients whopresent with an acute coronary syndrome: clinical inves-tigation and reports. Circulation 2002 Sep 24; 106 (12Suppl 1): I19-I22.

12. Myhre U, Stenseth R, Karevold A et al. Bleeding followingcoronary surgery after preoperative low-molecular-weightheparin. Asian Cardiovasc Thorac Ann 2004; 12: 3-6.

13. McDonald SB, Renna M, Spitznagel EL et al. Preopera-tive use of enoxaparin increases the risk of postoperativebleeding and re-exploration in cardiac surgery patients.J Cardiothorac Vasc Anesth 2005; 19: 4-10.

14. Kincaid EH, Monroe ML, Saliba DL et al. Effects of pre-operative enoxaparin versus unfractionated heparin onbleeding indices in patients undergoing coronary arterybypass grafting. Ann Thorac Surg 2003; 76: 124-8.

15. Ranucci M, Isgro G, Cazzaniga A et al. Predictors forheparin resistance in patients undergoing coronary arterybypass grafting. Perfusion 1999; 14: 437-42.

INTRODUCTION

Parmi les personnes âgées fragiles, le risque de déclinfonctionnel lié à une hospitalisation est augmenté, pro-bablement autant par la rupture de l’équilibre global quepar la pathologie même qui motive la prise en charge hos-pitalière [1-2]. L’inadéquation entre le profil de fragilité

et les structures traditionnelles de l’hôpital d’une part, etla méconnaissance des indicateurs de ce profil d’autrepart, exposent la population âgée aux risques de pertesde capacités fonctionnelles, de placement en institutionet de réadmissions à l’hôpital [3-5].

Ces risques peuvent être infléchis favorablement si,dès l’admission, les indicateurs de ces risques sont prisen compte et que des interventions ciblées selon le mo-dèle de l’évaluation gériatrique standardisée sont misesen œuvre durant le séjour avec une planification pro-active de la sortie [6-10].

C’est avec cette conception de soins que les servicesspécialisés en gériatrie, appelés aussi unités de courtséjour gériatrique (UCSG), ont été mis en place dansplusieurs pays depuis plus de 30 ans [11].

La principale compétence des unités de gériatrie setrouve dans la capacité de l’équipe hospitalière pluridis-ciplinaire à répondre à trois défis :

– La prise en charge de la polypathologie, en quoi elle

Journal Médical Libanais 2015 • Volume 63 (4) 191

AARRTTIICCLLEE OORRIIGGIINNAALL// OORRIIGGIINNAALL AARRTTIICCLLEEPROFIL GÉRIATRIQUE selon l’INSTRUMENT IDENTIFICATION SYSTÉMATIQUE des AÎNÉSà RISQUE (ISAR) au SERVICE des URGENCES d’un CENTRE HOSPITALIER UNIVERSITAIREhttp://www.lebanesemedicaljournal.org/articles/63-4/original4.pdf

Rita EL-HAYECK1, Rafic BADDOURA2, Patricia FADEL1, Amine WEHBÉ1, Antoine ZOGHBY3, Marc BERTHEL4

El-Hayeck R, Baddoura R, Fadel P, Wehbé A, Zoghby A, Berthel M.Profil gériatrique selon l’instrument Identification systématiquedes aînés à risque (ISAR) au service des urgences d’un centrehospitalier universitaire. J Med Liban 2015 ; 63 (4) : 191-197.

El-Hayeck R, Baddoura R, Fadel P, Wehbé A, Zoghby A, Berthel M.[Geriatric profile according to the Identification of Seniors AtRisk (ISAR) tool in the emergency department in a teachinghospital]. J Med Liban 2015 ; 63 (4) :) : 191-197.

1Département de Gériatrie, Faculté de Médecine, Université Saint-Joseph, Beyrouth, Liban.

2Service de Rhumatologie, Centre hospitalier universitaire Hôtel-Dieu de France (CHU-HDF), Beyrouth, Liban.

3Service des Urgences, CHU-HDF. 4Pôle de Gériatrie, Hôpitaux Universitaires de Strasbourg, France.

Correspondance : Dr Rita El-Hayeck. e-mail : [email protected]

RÉSUMÉ • Contexte : Avec l’augmentation prévisible dunombre des personnes âgées au Liban, il est important demettre en place des services adaptés à leurs besoins spé-cifiques. Le questionnaire Identification of Seniors At Risk(ISAR) permet d’identifier, aux urgences, les personnes âgéesfragiles à risque d’évènements défavorables et susceptiblesde bénéficier d’une prise en charge gériatrique spécifique.Objectifs : i) Évaluer le profil gériatrique selon le score ISAR ;ii) corréler le score ISAR aux suites deux mois après l’accueilaux urgences. Méthodologie: Étude prospective sur deuxmois : aux urgences du centre hospitalier universitaire del’Hôtel-Dieu de France, nous avons interrogé 273 personnesâgées de 70 ans et plus (ou l’aidant de la personne en ques-tion) en utilisant le questionnaire ISAR. Un suivi par téléphonea été fait deux mois plus tard. Résultats: La prévalence desujets avec un score ISAR ≥ 2 et donc susceptibles de béné-ficier de la mise en place d’un service de gériatrie est de 70,7%(IC 95% : 64,9-76,0). Les sujets âgés reçus aux urgences avecun score ISAR ≥ 2 sont plus à risque d’être hospitalisés (51,6%)que les sujets de même âge et sexe avec un score ISAR < 2(36,9%) (p = 0,034). Au bout de 2 mois de suivi, le risque deréadmission à l’hôpital et le risque de décès étaient significa-tivement associés au score ISAR (p = 0,0005) Conclusion:Le pourcentage de personnes âgées susceptibles de bénéfi-cier d’une prise en charge par une filière gériatrique spécialiséedès leur accueil aux urgences d’un centre hospitalier universi-taire est significativement élevé [70,7% (IC 95% : 64,9-76,0)].

Mots-clés : personnes âgées, services des urgences, dépistage,services sanitaires gériatriques, Liban

ABSTRACT • Background : With the anticipated increasein the number of elderly people in Lebanon, it is important todevelop services tailored to their specific needs. The Identi-fication of Seniors At Risk (ISAR) tool identifies, in emergen-cy setting, frail elderly people at risk of adverse outcomes,who are more likely to benefit from a geriatric approach.Objectives : i) Assess the geriatric profile according toISAR score ; ii) correlate the score to outcomes two monthsafter Emergency Department (ED) visit. Methods : A two-month prospective study : at the ED of Hôtel-Dieu de FranceHospital, we interviewed 273 people aged 70 years andolder (or their caregiver) using the ISAR tool. Telephonefollow-up was done two months later. Results : Theprevalence of subjects with ISAR score ≥ 2 and thus likely tobenefit from the establishment of a geriatric service is70.7% (95% CI : 64.9-76.0). Elderly patients admitted to EDwith ISAR score ≥ 2 are more likely to be hospitalized(51.6%) than subjects of the same age and sex with ISARscore < 2 (36.9%) (p = 0.034). After two months of follow-up,the risk of hospital readmission and the risk of death wassignificantly associated with the ISAR score (p = 0.0005)Conclusion : The percentage of elderly people likely tobenefit from specialized geriatric care network upon admis-sion to the ED of a university hospital is significantly high[70.7% (95% CI: 64.9-76 .0)].

192 Journal Médical Libanais 2015 • Volume 63 (4) R. EL-HAYECK et al. – Utilité de l’instrument ISAR dans les soins gériatriques

se rapproche de sa sœur aînée, la médecine interne,mais se distingue des spécialités médicales d’organe.

– Le repérage des fragilités par la démarche validéedans la littérature internationale de l’évaluation gé-riatrique standardisée [8, 12-13].

– La capacité à préserver voire à améliorer lescapacités fonctionnelles, ce qui permet de réduirel’entrée en institution et favorise ainsi le maintien àdomicile des patients âgés [12-13].

Au Liban, sur l’ensemble du territoire, seulementdeux UCSG ont été mises en place jusqu’à cette date[14]. Actuellement, la plupart des patients âgés fra-giles sont encore admis dans des services d’hospitalisa-tion traditionnels.

Les projections démographiques montrent que d’ici2035, la proportion des personnes âgées de 65 ans etplus va doubler, passant de 7,3% à 14% [15]. Il est doncimportant de mettre en place des UCSG adaptées auxbesoins spécifiques et prévisibles de cette tranche d’âge ;mais ce qui est encore plus pertinent c’est de biencibler la clientèle gériatrique qui bénéficiera le plus deces unités afin d’utiliser à bon escient nos ressourcessanitaires [14].

Dans cette perspective, l’utilisation du questionnaireIdentification of Seniors At Risk (ISAR) permet d’identi-fier, aux urgences, les personnes âgées fragiles à risqued’évènements défavorables et susceptibles de bénéficierd’une prise en charge gériatrique [16-18]. Le question-naire ISAR comprend six items (Tableau I) évaluant laprésence de facteurs de risque prédictifs d’évènementsdéfavorables tels que : hospitalisation, déclin fonctionnel,entrée en institutions, voire réadmissions aux urgencesà brève échéance [18]. Un patient est considéré à risque desubir un événement défavorable lorsqu’il a plus de 2 ré-ponses positives avec une sensibilité de 73% et une spé-cificité de 51% [18-19].

D’autres instruments de dépistage existent tels quele Triage Risk Screening Tool (TRST) [20], l’HospitalAdmission Risk Profile (HARP), et le Care ComplexityPrediction Instrument (CCPI) [21]. Cependant, deuxétudes ont démontré que le test ISAR performe mieuxque le TRST lorsqu’il a été appliqué dans des servicesd’urgences belges et canadiens [22-23]. Une étude

comparant les trois instruments HARP, CCPI et ISARa démontré la supériorité prédictive de ce dernier surle risque de déclin fonctionnel [23]. D’autres étudesont également montré les bonnes performances du testISAR dans différents contextes de soins [17-19, 22,24]. De plus le test ISAR est le plus facile à adminis-trer [25].

Dans la perspective de création d’une UCSG à l’Hôtel-Dieu de France qui est le Centre hospitalier universitaire(CHU) de l’Université St Joseph (USJ), et pour répondreà la question « Quel est le nombre de patients âgés quinécessiteraient une prise en charge par une filière géria-trique ? », cette étude a été mise en place. Elle avait doncpour objectifs :

– Évaluer le profil gériatrique selon le score ISARparmi les personnes âgées qui se présentent auxurgences de l’Hôtel-Dieu de France.

– Estimer la prévalence des personnes âgées ayant unscore ≥ 2 .

– Identifier des facteurs de risque potentiels d’avoirun score ISAR ≥ 2.

– Corréler le score ISAR aux suites après l’accueilaux urgences.

MATÉRIELS ET MÉTHODES

La population étudiée incluait l’ensemble des person-nes âgées de 70 ans et plus qui se sont présentées auxurgences de l’Hôtel-Dieu de France, pendant un mois, du4 juin au 4 juillet 2012.

Le patient, ou un membre de sa famille qui l’accom-pagnait, a été interrogé aux urgences par les internes deseptième année de Médecine, avec le questionnaire ISAR.Les personnes qui n’ont pas pu être interrogées aux ur-gences, ont été rappelées dans leur chambre durant lesdeux premiers jours de leur admission. Les données col-lectées comprenaient outre les items du score ISAR :l’âge, le sexe, l’existence d’une couverture médicale,l’admission ou non à l’hôpital et, en cas d’admission,dans quel service hospitalier le sujet avait été admis.

Pour les personnes qui n’ont pas été admises à l’HDF,elles ont été rappelées à domicile deux mois plus tarddurant la période du 1er septembre au 24 octobre 2012.

TABLEAU IQUESTIONNAIRE ISAR*

QUESTIONS

1. Avant cette admission aux urgences, aviez-vous besoin d’aide à domicile ? Oui / Non

2. Depuis le début des symptômes qui vous ont amené aux urgences, avez-vous eu besoin deOui / Nonplus d’aide à domicile ?

3. Avez-vous été hospitalisé pour 1 ou plusieurs jours ces 6 derniers mois ? Oui / Non

4. Dans la vie quotidienne souffrez-vous de problèmes de vue ? Oui / Non

5. Dans la vie quotidienne souffrez-vous de problèmes de mémoire ? Oui / Non

6. Prenez-vous plus de 3 médicaments par jour ? Oui / Non

* Questionnaire de dépistage des patients âgés à risque d’évènements indésirables. Un patient est considéré à risque d’évènement indésirable (déclin fonctionnel, réadmission) avec plus de 2 réponses positives.

R. EL-HAYECK et al. – Utilité de l’instrument ISAR dans les soins gériatriques Journal Médical Libanais 2015 • Volume 63 (4) 193

Le questionnaire utilisé comprenait les informationssuivantes : âge, sexe, score ISAR, système de couverturemédicale, si le patient a regagné son domicile ou s’il aété transféré dans un autre hôpital et dans ce dernier caspourquoi et dans quel service il a été admis, et finale-ment l’état du malade deux mois après sa présentationaux urgences.

L’analyse statistique a été réalisée avec le logicielSTATA version 9. L’hypothèse d’une association entre unscore ISAR ≥ 2 et des facteurs de risque potentiels a ététestée par un χ2 pour l’analyse univariée et logistiquepour une analyse multivariée. Le degré de significationstatistique est de 5%. Les différentes estimations sontcalculées avec leur intervalle de confiance à 95%.

RÉSULTATS

Du 4 juin au 4 juillet 2012, 2575 consultations ont étéréalisées aux urgences parmi lesquelles 374 concernaientdes personnes âgées de 70 et plus, soit une prévalence de14,5% de l’effectif total des consultations aux urgences(Figure 1). Ces 374 consultations correspondaient effec-tivement à 355 sujets car certaines personnes se sont pré-sentées aux urgences à plusieurs reprises durant cettemême période.

Caractéristiques sociodémographiquesde la population étudiée

Il y avait 184 hommes soit 51,8% de l’effectif, et 171femmes soit 48,2% de l’effectif (IC 95% : 42,9-53,5%).L’âge maximal était de 96 ans avec une médiane de 78 anset une moyenne (DS = 6,1) de 79 ans. Pas de différencesignificative entre hommes (79,1) et femmes (79,0). Des355 personnes reçues aux urgences, 181 (50,9%) ont étéhospitalisées à l’HDF. Nous avons pu interroger 128 per-sonnes hospitalisées (soit 70,7% de ceux qui ont été hos-pitalisés à l’HDF). Pour le reste : 22 personnes ont été

admises aux soins intensifs et n’ont pas pu être inter-rogées, et pour les 31 qui restaient les causes étaientmultiples : refus de participation, troubles de l’acuité au-ditive, départ de l’hôpital avant de les interroger.

Des 174 personnes qui n’ont pas été hospitalisées,nous avons pu en appeler 145 (soit 83,3%).

Pour le reste : • 14 : pas de réponses après plusieursappels • 8 : faux numéros de téléphone • 3 : refus derépondre • 3 : ligne occupée après plusieurs tentatives• 1 : en voyage) (Figure 1).

Douze personnes ont été transférées dans un autrehôpital (10/12 à cause de faute de place) lorsqu’elles ontquitté les urgences de l’HDF.

Au total, sur les 273 personnes que nous avons puinterroger, nous avons eu les résultats suivants :

Bénéfices d’un tiers payant86,8% des personnes interrogées bénéficiaient d’un

tiers payant (Tableau II).

TABLEAU IIPOURCENTAGE DE SUJETS BÉNÉFICIANT D’UN TIERS PAYANT

PAR CATÉGORIE

Tiers payant Effectif Fréquence

Pas de réponse 8 2,9%Assurance privée 84 30,8%CNSS 49 17,9%Assurance privée + CNSS 17 6,2%Mutuelle des fonctionnaires de l’État 31 11,4%Armée 8 2,9%Payant ordinaire 36 13,2%Caisse des Français à l’étranger 7 2,6%Autres 33 12,1%

Total 273 100,0%CNSS: Caisse Nationale de Sécurité Sociale

FIGURE 1. Population étudiée

128 personnes interrogées 145 personnes interrogées

181 personnes hospitalisées 174 personnes non hospitalisées

22 : admises aux soins intensifs 29 : pas de réponse,31 : refus de participation, départ de l’hôpital faux numéros, occupé,

avant de les interroger… refus de répondre...

2575 personnes admises aux urgences

355 personnes âgées de 70 ans et plus

Répartition des personnes hospitalisées à l’HDF selonles services d’admission à l’hôpital

La proportion de sujets hospitalisés est de 50,9%(IC 95% : 45,7%-56,3%) parmi l’ensemble des sujets re-çus aux urgences ; elle est de 49,2% parmi les hommesversus 54,1% parmi les femmes, différence statistique-ment non significative. (p = 0,35)

71% des personnes âgées admises à l’HDF ont étéhospitalisées dans des services de médecine, 14% dansdes services de chirurgie et 15% dans des services desoins intensifs.

Pourcentage de réponses « oui » aux questions duscore ISAR (cf. Figure 2)

Distribution du score ISAR dans la populationétudiée (cf. Figure 3)

Le score ISAR varie entre 0 et 6 avec une médiane de2 et une moyenne de 2,5 (DS = 1,5). Pas de différencesignificative entre hommes (2,5) et femmes (2,6) (p = 0,8).

Prévalence des sujets avec un score ISAR ≥ 2 et doncsusceptibles de bénéficier de la mise en place d’uneUCSG

Parmi l’ensemble des 273 personnes interrogées, lenombre de sujets avec un score ISAR ≥ 2 était de 193soit une prévalence de 70,7% (IC 95% : 64,9-76,0).

La prévalence était de 67,9% (IC 95% : 59,3-75,9)parmi les femmes et de 77,1% (IC 95% : 68,9-83,9)parmi les hommes ; (p = 0,094).

Parmi les personnes qui ont été hospitalisées à l’HDF,le taux de prévalence était de 77,3% versus 64,8% parmiles personnes non hospitalisées à l’HDF.

Corrélation entre le score ISAR et l’âgeLe score ISAR augmentait significativement avec

l’âge : le χ2 de tendance est de 10,26 avec un p de0,00136 ; la fréquence de sujets ayant un score ISAR ≥ 2augmentait aussi avec l’âge. (Tableau III).

Comparaison entre hospitalisés et non hospitalisés• La moyenne (DS) d’âge des sujets hospitalisés est

80,2 (6,2) ans versus 77,9 (5,7) parmi les sujets nonhospitalisés ; la différence est statistiquement signi-ficative (p = 0,0003).

24,81


TABLEAU IIIDISTRIBUTION EN NOMBRE ABSOLU (%) DU SCORE ISAR ET FRÉQUENCE DES SUJETS

AYANT UN SCORE ≥ 2 PAR TRANCHE D’ÂGE DE 5 ANS

Score ISAR 0 1 2 3 4 5 6 % Score ISAR ≥ 2

[70-75[ 6 (8,3%) 18 (25%) 15 (20,8%) 15 (20,8%) 14 (19,4%) 4 (5,5%) 0 66,7%

[75-80[ 6 (7,4%) 15 (18,5%) 26 (32,1%) 20 (24,7%) 10 (12,3%) 3 (3,7%) 1 (1,2%) 74,1%

[80-85[ 2 (3,3%) 16 (26,7%) 18 (30%) 6 (10%) 11 (18,3%) 4 (6,7%) 3 (5%) 70,0%

[85-90[ 2 (5,4%) 6 (16,2%) 5 (13,5%) 9 (24,3%) 10 (27%) 4 (10,8%) 1 (2,7%) 78,4%

90+ 0 2 (13,3%) 2 (13,3%) 3 (20%) 3 (20%) 4 (26,7%) 1 (6,6%) 86,7%

FIGURE 2. Pourcentage de réponses « oui » à chacune desquestions du score ISAR

80

70

60

50

40

30

20

10

0

Aide

à d

omici

le

Plus

d’a

ide

Anté

céde

nts

d’ho

spita

lisat

ion

Trou

bles

visu

els

Trou

bles

de

mém

oire

Plus

de

3m

édica

men

ts

35,7

78,5

46,2

24.8

%% de “oui”

25

20

15

10

5

00 1 2 3 4 5 6

FIGURE 3. Distribution en pourcentage du score ISAR dansla population étudiée.

6,027,14

2,26

18,05

20,321,43

%%

22.1

46,6

AG

E


• Le score ISAR des sujets hospitalisés est 2,8 (1,4)versus 2,3 (1,5) parmi les sujets non hospitalisés ; ladifférence est également statistiquement significa-tive (p = 0,01).

• Le score ISAR était significativement lié à la proba-bilité d’hospitalisation, en ajustant sur l’âge et lesexe ; OR (IC 95%) : 1,21 (1,01-1,44) ; p = 0,035(Tableau IV).

Comparaison des sujets ayant un score ISAR ≥ 2 avecceux ayant un score ISAR < 2

Il n’y a pas de différence significative entre les deuxgroupes selon l’âge, le sexe, le système de couverturemédicale. Par contre, ceux ayant un score ISAR ≥ 2 sontplus à risque d’être hospitalisés puisque le pourcentagede sujets hospitalisés parmi ceux avec un ISAR < 2 estde 36,9% versus 51,6% pour un ISAR ≥ 2 ; p = 0.034.

Suivi au bout de 2 mois• Parmi les 170 sujets non hospitalisés, 145 sujets ont

pu être réévalués deux mois plus tard : 29% ont étéréadmis plus tard à l’hôpital, 6% sont décédés et64,8% sont dans un état stationnaire.

• Le score ISAR est significativement lié à chacunede ces trois éventualités ; le score ISAR moyen estde 1,9 (1,3) parmi ceux dont l’état reste station-naire, 2,9 (1,6) parmi ceux qui sont réadmis à l’hô-pital et 3,2 (0,5) parmi ceux qui décèdent dans lesdeux mois qui suivent leur arrivée aux urgences. Ladifférence ainsi observée entre les trois groupes eststatistiquement significative (p = 0,0005)

DISCUSSION

Notre étude a montré que les personnes âgées de 70 anset plus constituent 14,5% de l’effectif total des admis auxurgences. Les données américaines montrent un tauxmoyen de 18% pour les personnes âgées de 65 ans et plus(allant de 11% à 23%) [26]. Les données italiennes sontsimilaires avec des taux de 21% en moyenne [27-28].Dans la plupart de ces pays ce taux est en train d’aug-menter due au vieillissement de la population [26-28].

Dans notre étude, 51% des personnes âgées qui sesont présentées aux urgences ont nécessité une hospitali-sation. Par rapport aux sujets jeunes, les personnes âgéesutilisent les services des urgences pour des problèmesplus urgents [29], nécessitent plus de ressources, et plusd’admissions hospitalières (30-50% vs 12% chez lesjeunes) [30]. Les personnes âgées sont aussi plus àrisque de visites à répétition aux urgences, de réhospita-lisation, d’entrée en institutions, de déclin fonctionnel etde mortalité après une visite aux urgences [30].

Le questionnaire ISAR a été développé pour identi-fier les personnes âgées le plus à risque d’évolution défa-vorable [18]. Il a été démontré qu’il performe bien dansles services des urgences [17-19, 22, 24].

C’est la première étude libanaise qui étudie le profilde fragilité selon le score ISAR chez les personnes âgées

qui se présentent aux urgences d’un CHU.Les résultats d’une étude récente faite en Grande-

Bretagne sont assez proches de la nôtre puisque 39%avaient besoin d’aide à domicile (versus 35,7% dans notreétude), 25% des problèmes visuels (versus 22% dans notreétude), 18% avaient des problèmes de mémoire (24,8%dans notre étude), 45% (46,2% dans notre étude) ont desantécédents d’hospitalisation dans les six mois précédents,78% (78,5% dans notre étude) prenaient plus de 3 médica-ments/jour. Par contre, beaucoup plus de personnes dansnotre étude ont eu besoin de plus d’aide à domicile depuisle début des symptômes ayant motivé le recours aux ur-gences (46% versus 25% dans l’étude anglaise) [31].

La prévalence du score ISAR ≥ 2 dans notre échantil-lon est de 70,7%. Ce chiffre est assez proche d’une autreétude récente faite en Italie qui a montré une prévalencede 68%. Cette étude ayant pris en considération les su-jets âgés à partir de l’âge de 65 ans alors que dans notreétude c’était à partir de 70 ans, mais la moyenne d’âgede l’étude italienne est supérieure à notre échantillon(81,7 ans vs 79 ans). C’est pour cela, sachant que la pré-valence du score ISAR > 2 augmente avec l’âge, quenous avons trouvé une prévalence assez proche de l’étudeitalienne [32].

Dans l’étude faite en Grande-Bretagne consacréecomme la nôtre aux sujets âgés de 70 ans et plus, 69%avaient un score ISAR ≥ 2 [31], ce qui est assez prochede nos résultats.

Donc, selon notre étude, plus des 2/3 des patients âgésde plus de 70 ans qui se présentent aux urgences néces-sitent une évaluation gériatrique standardisée et davan-tage, plus des 3/4, parmi ceux qui sont hospitalisés.

Il est illusoire d’admettre tous ces patients dans uneunité de court séjour gériatrique surtout que 30% ont étéadmis directement en chirurgie ou en soins intensifs.Mais ce qui est évident, c’est que ce sont des patientsqui, à défaut d’être admis dans une UCSG, peuvent aumoins profiter de la consultation d’une équipe mobile degériatrie.

Notre étude a démontré que les personnes qui se pré-sentent aux urgences avec un score ISAR ≥ 2 ont un ORde 1,21 d’être hospitalisés. Ces résultats sont en accordavec d’autres études qui ont aussi démontré un OR de2,68 [32] ou de 3,38 pour le risque d’hospitalisation[33]. Notre étude a aussi démontré qu’un score ISARélevé est relié à un risque plus élevé de mortalité. Ce quiest en accord avec plusieurs autres études [18, 25, 30, 32].

Presque tous les patients avaient un régime d’assu-rance-maladie ce qui ne représente pas vraiment l’en-

TABLEAU IVCORRÉLATION ENTRE LE SCORE ISAR ET LE RISQUE

D’HOSPITALISATION LORS DE L’ACCUEIL AUX URGENCES

Variable OR (IC 95%) P

Age 1,05 (1,01-1,09) 0,035

Sexe 1,85 (1,12-3,04) 0,016

ISAR 1,21 (1,01-1,44) 0,035


semble de la population âgée libanaise où seulementprès de 59% bénéficient d’une couverture médicale [34].Il y a donc un biais de sélection car le CHU est un hôpi-tal privé et ceux qui n’ont pas d’assurance se dirigentplutôt vers les hôpitaux publics. Donc, il est difficiled’extrapoler les résultats à l’ensemble des services d’ur-gence.

Les patients hospitalisés n’ont pas été interrogés dansun deuxième temps pour évaluer leur état deux mois plustard, à l’instar de ce qu’a été fait avec les sujets retour-nant à domicile.

Nous n’avons pas évalué la corrélation entre le scoreISAR et le risque de déclin fonctionnel mais ceci pourrafaire l’objet d’études ultérieures. Comme il sera intéres-sant d’étendre cette étude à d’autres services d’urgencede manière à avoir une population représentative de l’en-semble du pays et de pouvoir ainsi estimer approxima-tivement la charge potentielle de soins gériatriques auniveau national.

CONCLUSION

Bien que la littérature scientifique ait démontré l’exis-tence de plusieurs outils de dépistage des personnes âgéesfragiles aux urgences qui sont à risque d’évolution défa-vorable, ces outils ne sont pas encore utilisés dans les dif-férents services d’urgence au Liban. L’identification despersonnes âgées à risque d’évolution défavorable n’estque la première étape d’une prise en charge gériatriqueadaptée qui peut être réalisée soit dans un service de courtséjour gériatrique, soit par une équipe mobile de gériatrieauprès des différents services hospitaliers.

RÉFÉRENCES

1. Hoenig HM, Rubenstein LZ. Hospital-associated decon-ditioning and dysfunction. J Am Geriatr Soc 1991; 39:220-2.

2. Sager MA, Franke T, Inouye SK et al. Functional out-comes of acute medical illness and hospitalization inolder persons. Arch Intern Med 1996; 156: 645-52.

3. Hirsch C, Sommers L, Olsen A, Mullen L, Winograd C.The natural history of functional morbidity in hospitalizedolder patients. J Am Geriatr Soc 1990; 38: 1296-303.

4. Creditor MC. Hazards of hospitalization of the elderly.Ann Intern Med 1993; 118: 219-23.

5. Pfitzenmeyer P, Devaux-Hancer M, d’Athis P et al. Fac-teurs de risque de réhospitalisation en secteur de courtséjour gériatrique. Revue de Gériatrie 2000; 25: 243-50.

6. Boult C, Dowd B, McCaffrey D, Boult L, Hernandez R,Krulewitch H. Screening elders for risk of hospitaladmission. J Am Geriatr Soc 1993; 41: 811-17.

7. Fleming KC, Evans JM, Weber DC, Chutka DS. Practicalfunctional assessment of elderly persons: a primary-careapproach. Mayo Clin Proc 1995; 70: 890-910.

8. Rolland Y, Laroche-Decottignies F, Nourhashémi F,Lafont CV, Vellas B, Albarède JL. Evaluations et interven-tions gérontologiques standardisées. Revue de Gériatrie1999; 24: 109-12.

9. Winograd CH, Lindenberger EC, Chavez CM, Mauricio MP,

Shi H, Bloch DA. Identifying hospitalized older patientsat varying risk for physical performance decline: a newapproach. J Am Geriatr Soc 1997; 45: 604-9.

10. Winograd CH, Gerety MB, Chung M, Goldstein MK,Dominguez F Jr, Vallone R. Screening for frailty: criteriaand predictors of outcomes. J Am Geriatr Soc 1991; 39:778-84.

11. Latour J, Lebel P, Leclerc BS et al. Short-term geriatricassessment units: 30 years later. BMC Geriatrics 2010Jun 22; 10: 41.

12. Baztan JJ, Suarez-Garcia FM, Lopez-Arrieta J, Rodriguez-Manas L, Rodriguez-Artalejo F. Effectiveness of acutegeriatric units on functional decline, living at home, andcase fatality among older patients admitted to hospitalfor acute medical disorders: meta-analysis. BMJ 2009,338: b50.

13. Ellis G, Whitehead MA, Robinson D, O’Neill D,Langhorne P. Comprehensive geriatric assessment forolder adults admitted to hospital: meta-analysis of ran-domized controlled trials. BMJ 2011 Oct 27; 343: d6553.

14. El-Hayeck R. Unité de court séjour gériatrique : Pour-quoi et pour qui ? J Med Liban 2012; 60 (4): 249-51.

15. Population Division of the Department of Economic andSocial Affairs of the United Nations Secretariat, WorldPopulation Prospects: The 2010 Revision, http://esa.un.org/unpd/wpp/index.htm

16. Salvi F, Morichi V, Grilli A et al. Screening for frailty inelderly emergency department patients by using theIdentification of Seniors At Risk (ISAR). The Journal ofNutrition, Health & Aging 2012; 16 (4): 314-19.

17. Dendukuri N, Mccusker J, Belzile E. The identificationof seniors at risk screening tool: further evidence of con-current and predictive validity. J am Geriatr Soc 2004;52: 290-6.

18. McCusker J, Bellavance F, Cardin S, Trépanier S, Verdon J,Ardman O. Detection of older people at increased risk ofadverse health outcomes after an emergency visit: the ISARscreening tool. J Am Geriatr Soc 1999; 47: 1229-1237.

19. McCusker J, Bellavance F, Cardin S, Belzile E, Verdon J.Prediction of hospital utilization among elderly patientsduring the 6 months after an emergency department visit.Ann Emerg Med 2000; 36: 438-45.

20. Lee JS, Schwindt G, llangevin M et al. Validation of thetriage risk stratification tool to identify older persons atrisk for hospital admission and returning to the emer-gency department. J Am Geriatr Soc 2008; 56: 2112-17.

21. Hoogerduijn JG, Schuurmans MJ, Duijnstee MS, deRooij SE, Grypdonck MF. A systematic review of pre-dictors and screening instruments to identify older hospi-talized patients at risk for functional decline. J Clin Nurs2007 Jan; 16 (1): 46-57.

22. Moons P, De Ridder K, Geyskens K et al. Screening forrisk of readmission of patients aged 65 years and aboveafter the discharge from the emergency department: pre-dictive value of four instruments. Eur J Emerg Med2007; 14: 315-23.

23. Fan J, Worster A, Fernandes CM. Predictive validity ofthe triage risk screening tool for elderly patients in a Ca-nadian emergency department. Am J Emerg Med 2006;24: 540-4.

24. Salvi F, Morichi V, Grilli A et al. Predictive validity ofthe Identification of Seniors At Risk (ISAR) screeningtool in elderly patients presenting to two Italian emer-

gency departments. Aging Clin Exp Res 2009; 21: 69-75.25. Hoogerduijn JG, Schuurmans MJ, Korevaar JC,

Buurman BM, de Rooij SE. Identification of older hospi-talized patients at risk for functional decline, a study tocompare the predictive values of three screening instru-ments. J Clin Nurs 2010 May; 19 (9-10): 1219-25.

26. Roberts DC, McKay MP, Shaffer A. Increasing rates ofemergency department visits for elderly patients in theUnited States, 1993 to 2003. Ann Emerg Med 2008; 51:769-74.

27. Salvi F, Morichi V, Grilli A, Giorgi R, De Tommaso G,Dessi-Fulgheri P. The elderly in the emergency depart-ment: A critical review of problems and solutions. InternEmerg Med 2007; 2: 292-301.

28. Salvi F, Vita D, Morichi V, Fallani M, Dessi-Fulgheri P.Older adults use the emergency department appropriate-ly. Intern Emerg Med 2009; 4: 93-4.

29. Wolinsky FD, Liu L, Miller TR et al. Emergency depart-ment utilization patterns among older adults. J GerontolA Biol Sci Med Sci 2008; 63: 204-9.

30. Aminzadeh F, Dalziel WB. Older adults in the emer-gency department: A systematic review of pattern of use,

adverse outcomes, and effectiveness of interventions.Ann Emerg Med 2002; 39: 238-47.

31. Edmans J, Bradshaw L, Gladman JR et al. The Identi-fication of Seniors at Risk (ISAR) score to predict clini-cal outcomes and health service costs in older people dis-charged from UK acute medical units. Age Ageing 2013Nov; 42 (6): 747-53.

32. Di Bari M, Salvi F, Roberts AT et al. Prognostic strati-fication of elderly patients in the emergency department:a comparison between the ‘Identification of Seniors AtRisk’ and the ‘Silver Code’. J Gerontol A Biol Sci Med Sci2012; 67: 544-50

33. Salvi F, Morichi V, Lorenzetti B et al. Risk stratificationof older patients in the Emergency Department: Com-parison between the Identification of Seniors At Risk andTriage Risk Screening Tool. Rejuvenation Research2012; 15 (3): 288-94.

34. Boulos C, Salameh P, Barberger-Gateau P. The AMELstudy, a cross sectional population-based survey on agingand malnutrition in 1200 elderly Lebanese living in ruralsettings: protocol and sample characteristics. BMCPublic Health 2013; 13: 573.


INTRODUCTION

Acute rheumatic fever (ARF) is a post streptococcal seri-ous inflammatory illness due to its significant morbidityand cardiovascular death. It is associated with active car-ditis and rheumatic heart disease (RHD) manifested assubclinical carditis (SCC). RHD is reported to cause200,000-250,000 premature deaths every year in devel-

AARRTTIICCLLEE OORRIIGGIINNAALL// OORRIIGGIINNAALL AARRTTIICCLLEEFALSE AND TRUE ACUTE RHEUMATIC FEVERA Lebanese Experiencehttp://www.lebanesemedicaljournal.org/articles/63-4/original5.pdf

Najwa JAWHAR1,2, Imad CHAMI1,2, Zakhia SALIBA3, Ghassan CHEHAB1,2,3

Jawhar N, Chami I, Saliba Z, Chehab G. False and true acuterheumatic fever: a Lebanese experience. J Med Liban 2015 ;63 (4) : 198-202.

Jawhar N, Chami I, Saliba Z, Chehab G. Rhumatisme articu-laire aigu, vrai ou faux ? Étude libanaise. J Med Liban 2015 ;63 (4) : 198-202.

1Department of Pediatrics, Faculty of Medical Sciences, Leba-nese University, Greater Beirut, Lebanon.

2Department of Pediatrics, Sahel General Hospital, BeirutSouth Suburb, Lebanon.

3CHU Hôtel-Dieu de France, Beirut, Lebanon. Correspondence : Najwa Jawhar, MD.e-mail : [email protected]

RÉSUMÉ •Objectifs : L’objectif principal de cette étude est d’é-valuer l’importance de plusieurs signes cliniques et biologiquesprésentés par un grand nombre de patients en pédiatrie con-duisant les médecins à suspecter un rhumatisme articulaire aigu(RAA) et demander une échocardiographie visant à confirmer lediagnostic. Ces raisons comprennent des taux élevés d’ASO etune amygdalite récurrente qui mènent habituellement à la surcon-sommation d’antibiotiques. Un autre objectif est de souligner lerôle de l’échocardiographie pour confirmer le diagnostic de RAAen détectant la cardite rhumatismale (CR) dans les cas suspects.Matériels et méthodes : 401 cas présumés de RAA ont étéenregistrés au Registre national de cardiologie pédiatrique et con-génitale du 1er avril 2002 à fin mars 2013. Les raisons cliniquesqui mènent à la suspicion sont examinées et évaluées selonles critères de Jones modifiés. Les patients sont classés en deuxgroupes: 1) RAA confirmé et 2) autre diagnostic. Tous lespatients, venus de différentes régions géographiques du Liban etadressés par une variété de médecins, ont bénéficié d’uneéchocardiographie. Résultats: Sur un total de 401 cas suspects,le diagnostic de RAA a été confirmé chez 48 patients (12%). Dansce groupe, les raisons principales faisant soupçonner le diagnos-tic sont : souffle cardiaque : 29/48 (60,4%); arthrite : 28/48(58,3%); arthralgie avec réduction de l’activité physique : 15/48(31%) et chorée : 5/48 (10%). Dans le groupe ayant reçu undiagnostic différent, un taux d’ASO élevé 174/353 (49,3%),arthralgie isolée 91/353 (25,8%) et ASO élevé avec arthralgiesdans 44/353 (12,5%) sont les raisons prédominantes de soup-çon. L’échocardiographie Doppler a pu détecter 38/48 cas decardite active et cardite infraclinique (CIC). Conclusion: Soufflecardiaque, arthrite, arthralgie avec activité physique réduite etchorée sont les principales raisons cliniques qui mènent à soup-çonner le RAA. Des taux élevés d’ASO et/ou arthralgie ont été lesprincipales raisons observées dans le groupe ayant reçu un diag-nostic différent. Nous avons conclu qu’un taux d’ASO élevé etune vitesse de sédimentation normale avec arthralgies et uneactivité physique normale peuvent exclure ce diagnostic. L’écho-cardiographie Doppler est l’examen-clé pour confirmer la carditerhumatismale, CIC en particulier, et donc le RAA.

ABSTRACT • Objectives : The main aim of this study is toevaluate the significance of several clinical and laboratoryfindings presented by a large number of pediatric patientswho were suspected to have acute rheumatic fever (ARF) byphysicians who ordered echocardiogram for confirmation andfor avoiding overuse of unnecessary antibiotics. These sus-pecting reasons include elevated ASO titers and recurrenttonsillitis. Another aim is to stress the role of Doppler echo-cardiogram in confirming the diagnosis of ARF by detectingrheumatic heart disease (RHD) in suspected cases. Mate-rials and methods : 401 patients were registered in theNational Registry of the Pediatric and Congenital HeartDisease (RNCPC) from April 1, 2002 until the end of March2013 as suspected cases of ARF. The clinical reasons thatbrought suspicion are discussed and evaluated in accor-dance to modified Jones criteria. The patients are distributedinto two groups : 1) confirmed ARF and 2) different diagnosis.All patients underwent echocardiogram. Patients came fromdifferent geographic areas and were referred by a variety ofphysicians. Results : 48 patients out of 401 suspected casesturned to have confirmed ARF (12%). In this group, the pre-dominant suspecting reasons are reported as heart murmurin 29/48 (60.4%), arthritis in 28/48 (58.3%), arthralgia withreduced physical activity in 15/48 (31%) and chorea in 5/48(10%). In the different diagnosis group with a high ASO titer174/353 (49.3%), isolated arthralgia 91/353 (25.8%) andcombined elevated ASO with arthralgia in 44/353 (12.5%)were the predominant suspecting reasons. Doppler echocar-diography could detect 38/48 cases of active carditis and sub-clinical carditis (SCC). Conclusion : Heart murmur, arthritis,arthralgia with reduced physical activity and chorea are themain clinical reasons that lead to suspicion in the ARF group.High ASO titers, arthralgia, and combination of high ASO titerswith arthralgia are the main presenting symptoms in the groupwith a different diagnosis. We concluded that elevated ASOtiter and normal ESR with arthralgia not reducing physicalactivity can exclude ARF. Doppler echocardiography is foundto be a gold standard key to confirm acute rheumatic heartdisease and ARF.

Keywords : rheumatic fever, acute rheumatic fever, rheumaticheart disease, subclinical carditis, RNCPC

198 Lebanese Medical Journal 2015 • Volume 63 (4)

oping countries [1].WHO report in Geneva 2004 statesthat the incidence of ARF is directly related to socioeco-nomic risks that are mostly: poverty, overcrowding, andlimited access to medical centers and that ARF is one ofthe major problems in large cities of the third world.WHO also claims that the prevalence of ARF is decreas-ing in developed countries while it is endemic in mostdeveloping countries [2]. The annual incidence in USAapproximates 1 in 83 people or < 1 % [3]. In developingcountries, the incidence is 100-200 times greater thanthat in developed countries [4]. In general, the world-wide prevalence approximates 19 per 100,000 popula-tions ranging between 5-51 per 100,000 people. There arefew data about the incidence of rhumatic fever (RF) andits complications in the Middle East countries. Extra-polated statistics – not based on country specific datasources and only estimates having very limited relevanceto the actual incidence – are reported as 45,326 warningsin Lebanon. This is claimed by US Census Bureau, Pop-ulation Estimates 2004. Besides, studies about RF inLebanon are very limited. A Lebanese epidemiologicalsurvey reported in a prospective study from 1999 to year2005 states that rheumatic heart disease was the mostcommon acquired heart disease in children [5]. Compara-tively the mean annual incidence of ARF in Kuwait isreported as 23 per 100,000 people and is 11.2 per 100,000in Qatar, while it approximates 51 per 100,000 in India[6]. In Egypt, a field study on school children shows theprevalence of rheumatic fever and rheumatic heart dis-ease per 1000 is 1.3 and 1.2 respectively [7].

ARF being associated with heart problems remains asubstantial global health issue especially in developingcountries. This is emphasized through many studies in-cluding one reported by Dr Bitar et al. which was done on91 patients hospitalized in a tertiary hospital in Beirut.This study revealed carditis in 93% of cases, congestiveheart disease in 44%, arthralgia in 55% and fever in 62%[8]. Thus, early diagnosis and management is needed,based on evaluation of clinical, laboratory and echocar-diography findings in collaboration with Jones criteriathat are stated by the American Academy of Pediatrics(AAP) in 1992 and most recently updated in 2002 “mod-ified Jones criteria” [9]. Elevated ASO titers and recur-rent tonsillitis are recognized by many physicians aswarning reasons to suspect ARF. In fact, elevated ASOtiters denote group A streptococcal infection rather thanARF. According to “modified Jones criteria,’’ diagnosisof ARF does not require serologic evidence of recentgroup A strep infection, but documented strep infectionthrough a positive throat culture and that elevated ASOtiters are considered as minor criteria [9].

This study deals with patients coming from differentgeographic regions presenting with clinical and labora-tory manifestations that made physicians suspect ARF, sothey referred these patients and confirmed their diagno-sis by echocardiogram. We studied these different clini-cal and lab criteria in the group of patients who turned tohave ARF (true diagnosis) and in the group who turned

to have a different diagnosis. The role of elevated ASOtiters and repeated tonsillitis and its relevance to thediagnosis of ARF in our cases are also evaluated. In ad-dition, we aim to evaluate the role of Doppler echocar-diogram as a key tool for the diagnosis of ARF in theform of rheumatic carditis or subclinical carditis.

MATERIELS AND METHODS

The subjects of the retrospective study are children 5 to16 years old who were suspected to have ARF and automat-ically referred to the National Registry of the Pediatric &Congenital Heart Disease (RNCPC*), to undergo echocar-diogram for confirming the disease. This registry is a privatesector controlled and organized by a group of paediatriccardiologists. Children less than 5 years are not includedbecause ARF is reported as rare in this age group. Physicianswho referred the suspected patients are pediatricians,cardiologists, rheumatologists and general practitioners. Thegeographic locations where our patients were living arespecified. The duration of the study covered a period of 11 years from April 1st, 2002, until the end of March 2013.

The 401 patients are classified into 2 groups: the firstgroup proved to have true ARF based on clinical and lab-oratory criteria that coincide with Jones criteria, comple-mented with echocardiography findings. The secondgroup proved to have a different diagnosis since they hadnonconforming echocardiography and lack of fulfill-ment to Jones criteria. Clinical and lab reasons are dis-cussed and evaluated in all patients to estimate the sig-nificance of each criterion to diagnose ARF. All patientsunderwent Doppler echocardiography to identify if heartproblems are present, whether active carditis or subclini-cal carditis or other diagnosis. The role of Doppler echo-cardiography is stressed and evaluated.

RESULTS

Reviewing the clinical manifestations and echocardio-graphy findings of the heart in 401 suspected patients re-vealed true ARF diagnosis in 48/401 patients (12%) witha mean age of 7.7 years ranging between 7.2-16.8 years.Their distribution according to gender is M:F 1.3:1. Onthe other hand, patients with a diagnosis different from RFare 353/401 (88%), aged 1.7-15.2 yrs range with a meanof 7.6 years and distributed as M:F 1.2:1. Unlike the pa-tients with true diagnosis these patients do not fulfill Jonescriteria and their Doppler echocardiogram turned to benon suggestive of ARF (Table I).

TABLE IDISTRIBUTION of CASES in RELATION to GENDER and AGE

Number Male Female Age (years)*

Different diagnosis 353 (88%) 209 144 7.6 (1.7-15.2)ARF 48 (12%) 27 21 7.7 (7.2-16.8)

Total 401 236 165 7.6 (1.7-15.2)* Average and extremes ARF: acute rheumatic fever

*RNCPC: The National Registry of the Pediatric and Congenital Heart Disease of the Society of Cardiology in Lebanon, created in May 1999, Lebanese Order of Physicians. This registry is a private sector controlled and organized by a group of private pediatric cardiologists.

200 Lebanese Medical Journal 2015 • Volume 63 (4) N. JAWHAR et al. – Rheumatic fever in Lebanon

Physicians who referred suspected RF patients weremostly pediatricians. They referred 351/401 patients out ofwhich 45/351 proved to have the disease (11.2%). The re-maining 50/401 suspected patients were referred by otherphysicians: • 19/50 by cardiologists and only one turned tohave RF • 17/50 by rheumatologists out of which oneproved to have RF • 14/50 by general practitioners andagain one turned to have ARF (Table II).

The geographic distribution of the suspected patientsshows • 275 patients living in Beirut and its south sub-urbs region • 41 in Mount Lebanon • 52 in the Bekaa •13 in the North and • 21 in the South. The 48 ARFcases are geographically distributed as follows: • 27(56.2%) living in Beirut and its south suburbs region •7 (14.58%) in Mount Lebanon • 7 (14.58%) in the Bekaa• 5 (10.4%) in the North • and 2 (4%) in the South. Thesenumbers show that the majority of our ARF patients areliving in Beirut and its south suburbs region (56.2%)(Table III).

The clinical and lab findings that mostly concernedphysicians and made them suspect ARF and request aDoppler echocardiography are evaluated. In the group ofARF patients they are: • heart murmur in 29/48 patients(60.4%) • arthritis in 28/48 (58.3%) • arthralgia withreduced physical activity in 15/48 patients (31%) •chorea is evident in 5/48 patients (10%) and • prolongedfever in only 2/48 patients (4%). Recurrent tonsillitis,elevated ASO titers with or without arthralgia are notevident in the 48 ARF patients. Only 3 patients of theARF group were documented to have positive throat cul-ture of strep group A and were given penicillin G as pro-phylaxis (Table IV).

Other findings in the ARF group include elevatedESR in 42 patients (87.5%), fever in 37 cases (77%) andPR prolongation in 7 cases (15%). (Table V).

To summarize: heart murmur, arthritis, arthralgia withreduced physical activity and chorea are found to be themost predominant clinical reasons of suspicion in theARF group (Table IV). Also, elevated ESR and fever areadditional complementary reasons.

On the other hand, the suspected clinical and lab rea-sons that are predominant in the group with a differentdiagnosis include: • heart murmur in 16 out of 353 pa-tients (4.6%) • arthritis in 28/353 (7.9%) and • arthralgianot altering physical activity in 91/353 (25.8%)). Re-current tonsillitis is evident in 12/353 (3.4%), elevatedASO titers as an isolated finding and ASO associatedwith arthralgia are detected in 174/353 and in 44/353,(49.3%) and (12.5%) respectively. So, the major sus-

pecting reasons in this group are listed in decreasingorder: elevated ASO titers (49.3%), arthralgia (25.8%),elevated ASO combined with arthralgia (12.5%), arthri-tis (7.9%), heart murmur (4.6%) and recurrent tonsillitis(3.4%). (Table IV).

Carditis is diagnosed by echocardiography in 38/48cases (79%) in the ARF Group. These cases are distrib-uted as clinical carditis in 22/48 cases (46%), subclinicalcarditis in 16/48 (33%), mitral regurgitation (MR) in26/38 (68%), and MR combined with aortic regurgita-tion in 12/38 (32%) of cases (Table V).

DISCUSSION

The diagnosis of ARF is based on fulfillment to modi-fied Jones criteria complemented by Doppler echocar-diography. The files of 401 suspected patients are retro-spectively reviewed and 48 patients out of the 401 sus-pected cases found to have true ARF diagnosis (12%).The remaining 353/401 (88%) suspected patients werecomplaining of a different diagnosis since their clinicalfindings do not comply with Jones criteria and theirechocardiograms are either normal or suggest anotherdisease such as congenital heart disease or acquired heartdisease as Kawasaki. Both groups show male predomi-nance and nearly the same average age at presentation.Pediatricians are the predominant physicians who re-ferred most of the suspected cases, 351 patients out of401 (87.5%). The remaining physicians referred 50/401suspected cases (12.5%).

Regarding clinical findings, the cardiac murmur ispresented in 60.4% of the ARF group and in 16/353(4.6%) patients with a different diagnosis. This provesthat cardiac murmur is an important finding that bringssuspicion and mostly suggests carditis or subclinicalvalvular heart disease and thus echocardiography isneeded for confirmation. On the other hand, the murmurin the second group (patients with a different diagnosis)

TABLE IIDISTRIBUTION of REFERRED PATIENTS in RELATION

to REFERRING PHYSICIANS

Number (%) ARF Different diagnostic Total

Pediatricians 45 (11.2%) 306 (87.2%) 351

Cardiologists 1 (5%) 18 (95%) 19Rheumatologists 1 (6%) 16 (94%) 17General practitioners 1 (7%) 13 (93%) 14

*87.5% of total cases were referred by pediatricians & 12.5% by other physicians.

TABLE IIIGEOGRAPHIC DISTRIBUTION

Beirut + Southern Suburbs Mount-Lebanon Bekaa North South TotalNumber (%)

Acute Rheumatic Fever 27 (9.8%) 7 (17%) 7 (13%) 5 (39%) 2 (10%) 48 (12%)

Different diagnosis 248 (90.2%) 34 (83%) 45 (87%) 8 (61%) 19 (90%) 353 (88%)

TOTAL 275 41 52 13 21 401

N. JAWHAR et al. – Rheumatic fever in Lebanon Lebanese Medical Journal 2015 • Volume 63 (4) 201

may have a different etiology: a functional murmur or asign of a congenital heart disease. So, echocardiographyis needed, supplemented with clinical, lab and ausculta-tory findings for differentiation and for the diagnosis ofrheumatic heart disease [10].

Arthritis is detected in 28/48 (58.3%) patients of theARF group and in 28/353 (7.9%) patients of the groupwith a different disease. Arthritis is an important clinicalcriterion especially in the acute phase of ARF and it im-plies a first Doppler echocardiogram in children. Besides,arthritis is a significant major criterion in correlation toJones criteria [11-12]. Arthritis in the group of patientswith a different diagnosis may have a different etiologysuch as viral or bacterial infection or a connective tissuedisease.

Arthralgia as an isolated entity is detected in 15/48cases (31%) in the ARF group and in 91/353 (25.8%) inthe group of patients with a different diagnosis. Arthral-gia in ARF patients is more often associated with re-duced physical activity. Arthralgia associated with ele-vated ASO titer is not found in our ARF group, while itis present in 44/353 patients of the second group (12.5%),indicating that arthralgia could be a symptom of infectionas viral infection, bacterial infection as Streptococcusgroup A, systemic disease or a connective tissue disease.According to our series we can estimate that arthralgiaalone is not a significant index of suspicion that does notrequire echocardiography.

ASO titers were done 2-4 weeks after the onset ofsymptoms. They were not elevated in any of the ARFgroup, while elevated titers with or without arthralgiapresented in 44/353 (12.5%) and 174/353 (49.3%) pa-tients of the group of different disease respectively. Thus,in our series, ASO titer is found to be a poor reason tosuspect ARF. This is supported by the WHO report 2003in Geneva claiming that ASO titer is more an evidence ofgroup A streptococcal infection and is becoming a weakindicator to confirm ARF. It is also reported that elevatedtiters may be coincidental [13-14]. ASO titer is claimed to beelevated in many inflammatory diseases [15]. This explains the

elevated ASO titers in the group with a different diagnosis.Elia et al. evaluated ASO titers in patients with and

without ARF living in areas with high or low incidence ofRF. They concluded that non rheumatic patients living inan area with high incidence of RF did show elevated ASOtiters. This reflects more correctly a high prevalence ofstrep infections [16]. It is worth to mention here, thatmany physicians are still considering elevated ASO titersa warning sign to suspect ARF, thus leading to over diag-nosis. Regarding our series, elevated ASO titers aloneare not a significant reason to perform echocardiography.This is supported by the modified Jones criteria that con-sider elevated ASO is no more required as a serologicevidence of recent strep infection [9, 17].

Chorea is evident in 5 patients of our ARF group,5/48 i.e. (10%). The presence of chorea alone is enoughfor diagnosis and Doppler echocardiogram is not a mustexcept if carditis is suspected [11].

Recurrent tonsillitis is detected in 12/353 (3.4%) ofsuspected patients who turned to have a different diagno-sis, and not found in any of our ARF group. Strep infec-tion is not documented as a cause of tonsillitis by throatculture in our patients. Recurrent tonsillitis may resultfrom strep infection and any viral or bacterial infectionother than strep [11]. The important issue here is that re-current streptococcal tonsillitis is common in developingcountries and may lead to damage and complications incardiac valves that may pass unnoticed [6,18].

Prolonged fever is found in 2/48 patients of the ARFgroup (4%) and in 5/353 cases (1.4%) of the group ofdifferent diagnosis. It is a significant reason to suspectARF in the presence of other criteria that comply withJones criteria [4]. Increased erythrocyte sedimentation

TABLE VCRITERIA FOUND in the GROUP of CHILDREN with ACR

(Normal ESR in 3 patients with chorea, and 3 diagnosed few weeksafter the onset of the disease based on the medical history,

laboratory tests and evidence of carditis.)

Number %MAJOR CRITERIA

CarditisTotal 38 79Clinical 22 46Subclinical 16 33Mitral regurgitation [MR] 26 68MR + Aortic regurgitation 12 32

Arthritis 28 58.3 Erythema 6 12.5 Chorea 5 10.5 Nodules 1 2

MINOR CRITERIA

Elevated ESR 42 87.5Fever 37 77Arthralgia 15 31PR prolongation 7 15ACR: acute rheumatic fever ESR: erythrocyte sedimentation rate PR: interval between P & R in EKG

TABLE IVCLINICAL & LABORATORY FINDINGS that MAKE PHYSICIANS

SUSPECT ACUTE RHEUMATIC FEVER in PATIENTS

Reason for referring to Number of ARF Different

pediatric cardiologist patients (48) diagnosis (353)N % N %

Heart murmur 29 60.4 16 4.6Arthritis 28 58.3 28 7.9

Arthralgia 15 31 91 25.8Chorea 5 10 0 0

Prolonged fever 2 4 5 1.4Recurrent tonsillitis 0 0 12 3.4

Parental anxiety 0 0 7 2Elevated ASO 0 0 174 49.3

Elevated ASO + arthralgia 0 0 44 12.5

202 Lebanese Medical Journal 2015 • Volume 63 (4) N. JAWHAR et al. – Rheumatic fever in Lebanon

rate (ESR) in this study, is detected in 42/48 cases(87.5%) in the group of confirmed RF mainly in theacute phase of the disease and in the presence of carditisand arthritis. Normal ESR is reported in 3 patients withchorea and in 3 patients who were diagnosed few weeksafter the onset of the disease (Table V). ESR is elevatedin 156/353 (44%) patients of the group with a differentdiagnosis indicating a different inflammatory disease.

The application of the findings of our patients toJones criteria is shown in (Table V).

The role of Doppler echocardiography appeared to bean essential key tool for the diagnosis of rheumatic car-ditis and subclinical carditis (SCC) based on importantclinical criteria as heart murmur and arthritis. Echocar-diography could detect 38/48 cases of total carditis(79%) distributed as mentioned before in (Table V). Thisreflects that echocardiography is very needed to confirmthe diagnosis of rheumatic carditis and its long termcomplications as SCC [18-19]. In general, studies statethat the prevalence of rheumatic heart disease is muchhigher, around 10 times more diagnosed by echocardio-gram v/s clinical screening [18]. Performing the firstechocardiogram in children is important to the diagnosisof clinical and subclinical rheumatic carditis especiallythat during the past two decades, the presentation of RFhas changed markedly from that of an acute illness to amore subtle subacute form that is frequently missed.

CONCLUSION

• The prevalence rate in our limited study is 12% butthis does not reflect the prevalence rate in Lebanon.

• Heart murmur, arthritis and arthralgia reducing phy-sical activity are the predominant clinical reasons tosuspect ARF and thus to refer patients to Doppler echo-cardiography.

• Recurrent tonsillitis is found to have no value in sus-pecting and diagnosing RF, especially if strep infec-tion is not documented.

• Elevated ASO titers, according to our series, proved tohave a weak predictive value to suspect ARF and thusto do Doppler echocardiogram.

• Elevated ASO titers with normal ESR and with arthral-gia not reducing physical activity can exclude ARF.

• Doppler echocardiography is the most reliable key forthe diagnosis of rheumatic carditis and rheumatic heartdisease in children.

REFERENCES

1. Marijon E, Mirabel M, Celermajer DS, Jouven X. Rheu-matic heart disease. Lancet 2012 Mar 10; 379 (9819):953-64.

2. WHO. Rheumatic fever and rheumatic heart disease:report of a WHO Expert Consultation, Geneva: WorldHealth Organization, 2004.

3. Kaplan EL. Recent epidemiology of group A streptococ-cal infections in North America and abroad: an overview.Pediatrics. 1996 Jun; 07 (6 pt 2): 945-8.

4. Leiner S. Surveillance for rheumatic fever. New EnglandJournal of Medicine 1996; 334: 273-4.

5. Chehab G, Shalak W, Gerbaka B, Saliba Z. Inflammatoryheart diseases in childhood: Lebanese epidemiologicalsurvey. J Med Liban 2006 Jul-Sep; 54 (3): 124-31.

6. Das De Sudeep KS. The descriptive epidemiology ofacute rheumatic fever and rheumatic heart disease in lowand middle-income countries. American Journal ofEpidemiology and Infectious Disease 2013; 1 (4): 34-40.

7. Azzam A. Prevalence of rheumatic fever and rheumaticheart disease in a group of school children in GUAGovernorate and concomitant association with mitralregurgitation. J A C 2005 Dec; 16 (6).

8. Bitar FF, Hayek P, Obeid M, Gharzeddine W, Mikati M,Dbaibo GS. Rheumatic fever in children: a 15-year expe-rience in a developing country. Pediatr Cardiol 2000Mar-Apr; 21 (2): 119-22.

9. Ferrieri P. Jones Criteria Working Group, “Proceedingsof the Jones Criteria workshop”. Circulation 2002; (JonesCriteria Working Group) 106 (19): 2521-23.

10. Chehab G. Subclinical carditis during an initial attack ofacute rheumatic fever: contribution of colored Dopplerechocardiography and therapeutic advantages. J MedLiban 2001; 49 (6): 311-15.

11. Special writing group of the committee on rheumaticfever, endocarditis and Kawasaki disease of the Ameri-can Heart Association. Guidelines for the diagnosis ofRheumatic Fever: Jones criteria P 1992 update. JAMA1992; 268: 2069-73.

12. Chehab G. Assessment of the reasons prompting a firstechocardiography in children: a Lebanese multicenterstudy. J Med Liban 2007 Jan-Mar; 55 (1): 2-10.

13. Gerber MA. Acute rheumatic fever. In: Nelson, Book ofPediatrics, 18th edition, Philadelphia 2007: Saunders,1140-1144.

14. Shaikh N, Leonard E, Martin JM. Prevalence of strepto-coccal pharyngitis and streptococcal carriage in children:a meta-analysis. Pediatrics 2011; 26 (3): e557-e564.

15. Espinosa M, Gottlieb BS. Juvenile idiopathic arthritis.Pediatr Rev 2012 Jul; 33 (7): 303-13.

16. Ayoub EM, Nelson B, Shulman ST et al. Group A strepto-coccal antibodies in subjects with or without rheumaticfever in areas with high or low incidences of rheumaticfever. Clin Diagn Lab Immunol 2003 Sep; 10 (5): 886-90.

17. Gerber MA, Baltimore RS, Eaton CB, Gewitz M, RowleyAH, Shulman ST, Taubert KA. Prevention of rheumaticfever and diagnosis and treatment of acute Streptococcalpharyngitis: a scientific statement from the American HeartAssociation Rheumatic Fever, Endocarditis, and Kawa-saki Disease Committee of the Council on CardiovascularDisease in the Young, the Interdisciplinary Council onFunctional Genomics and Translational Biology, and theInterdisciplinary Council on Quality of Care and Out-comes Research: endorsed by the American Academy ofPediatrics. Circulation 2009 Mar 24; 119 (11): 1541-51.

18. Tubridy-Clark M, Carapetis JR. Subclinical carditis inrheumatic fever: a systematic review. Int J Cardiol 2007Jun 25; 119 (1): 54-8.

19. Marijon E, Ou P, Celermajer DS et al. Prevalence ofrheumatic heart disease detected by echocardiographicscreening. N Engl J Med 2007; 357 (5): 470-6.

20. Rheumatic fever in the new Millenium pediatrics, byguest Alyaa Kotby, April 16, 2014, Aapublications.org

INTRODUCTION

Speech has several acoustic and aerodynamic measuresthat relate to velopharyngeal function, among which isnasality. The assessment of nasality has traditionally beena perceptual task performed by speech language patholo-gists. The reliability of these perceptual ratings has alwaysbeen subject to criticism in view of the diversity in the rat-ing scales used and the presence of confounding factors inspeech that may mask the perception of nasality [1-4].Thus, despite the extensive information provided by thespeech therapists in the evaluation of patients with velo-pharyngeal incompetence and the relative validity of suchinformation, the need for other assessment tools has risen.

Nasal endoscopy and videofluoroscopy have been used asalternative or complementary diagnostic tests. However,both are rather invasive and one entails exposure to radia-tion therapy. The nasometer, a microcomputer-based in-strument devised by Kay Pentax to measure nasalance andcalculate a nasalance score [5], has been commonly usedas a noninvasive, reliable and objective device to measurethe extent of nasality [6-8]. The primary function of thenasometer is to provide a quantitative measure related toperceived hypernasality. The term nasalance is often usedto reflect the ratio of nasal acoustic energy to nasal-plus-oral acoustic energy in speech. The consensus remainsthat nasalance can be used as a reliable supplement in theevaluation of velopharyngeal competence. This is in viewof the results of extensive studies on the validity of thedevice that have shown a high correspondence betweenthe perceptual judgments of nasality and the nasalancescores [3,9-10].


AARRTTIICCLLEE OORRIIGGIINNAALL// OORRIIGGIINNAALL AARRTTIICCLLEENASALANCE SCORES IN LEBANESE ENGLISH-SPEAKING ADULTS USING NASOMETRIC ANALYSIShttp://www.lebanesemedicaljournal.org/articles/63-4/original6.pdf

Abdul Latif HAMDAN1, Georges ZIADE1, Jad JABBOUR2, Gebran KHNEIZER2, Issa KUTKUT2

Hamdan AL, Ziade G, Jabbour J, Khneizer G, Kutkut I.Nasalance scores in Lebanese English-speaking adults usingnasometric analysis. J Med Liban 2015 ; 63 (4) : 203-208.

Hamdan AL, Ziade G, Jabbour J, Khneizer G, Kutkut I. Scoresde nasalance pour adultes libanais de langue anglaise utilisantl’analyse nasométrique. J Med Liban 2015 ; 63 (4) : 203-208.

From American University of Beirut Medical Center, Lebanon,1Department of Otolaryngology, 2Medical student.

Correspondence: Abdul Latif Hamdan, MD. AUB-MedicalCenter, P.O.Box 110236, Beirut. Lebanon.

e-mail: [email protected]

ABSTRACT • Objective : To obtain normative data for nasal-ance scores in Middle Eastern English-speaking adult popula-tion. Design : Cross-sectional study. Participants : A total of102 subjects were recruited in the study, 26 were excluded,thus, 77 Middle Eastern English-speaking adults (mean age =23.77 ± 4.295 ; 39 males, 38 females) with normal speech andno hearing problems participated in the study. Interventions :Using Nasometer II 6450, nasalance scores were obtainedfor each participant’s readings of 3 passages: Zoo and Rain-bow passages and nasal sentences. Main outcome meas-ures : Mean nasalance score, standard deviation and range.Results : Mean nasalance scores for Zoo passage, Rainbowpassage and nasal sentences were 25.21 ± 11.07, 34.04 ±9.30 and 41.29 ± 9.87 respectively. Mean scores didn’t differsignificantly between males and females, though scores forthe Zoo passage were slightly higher among males (26.51 ±11.66) than females (23.87 ± 10.42). In within-session relia-bility testing, 86% of retests for all three passages fell within5 nasalance points of the previous test. In across-session re-liability testing, 93% of retests for all 3 passages were within5 points of initial test. Conclusions and Relevance: Thisstudy provides normative data for nasalance scores amongMiddle-Eastern adults, which can help make nasometer anddetermination of nasalance more clinically useful in this popu-lation.

Keywords: nasalance, nasality, nasometer, Middle-East

RÉSUMÉ•Objectif : Obtenir des données normatives pour lesscores de nasalance dans la population adulte anglophone duMoyen-Orient. Protocole : Étude transversale. Participants :Sur un total de 102 sujets recrutés, 26 ont été exclus ; 77adultes anglophones du Moyen-Orient (âge moyen = 23,77 ±4.295 ; 39 hommes, 38 femmes), ayant un discours normal etsans problèmes auditifs, ont participé à l’étude. Interventions:On a demandé aux participants de lire 3 passages chacun: lespassages Zoo et Rainbow et des phrases nasales; les scoresde nasalance ont été obtenus avec un Nasometer II 6450.Principaux critères de jugement: le score de nasalance,l’écart type et la gamme. Résultats : Les scores de nasalancepour les passages Zoo, Rainbow et les phrases nasales étaientde 25,21 ± 11,07, 34,04 ± 9,30 et 41,29 ± 9,87 respectivement.Les scores moyens ne diffèrent pas significativement entrehommes et femmes, bien que les scores pour le passage Zooaient été légèrement plus élevés chez les hommes (26,51 ±11,66) que chez les femmes (23,87 ± 10,42). Les tests de fia-bilité menés dans la même session ont montré que 86% desnouveaux essais pour les trois passages se trouvaient à pres-que 5 points de nasalance de l’essai précédent. Dans les testsde fiabilité entre sessions, 93% des nouveaux essais pour lestrois passages se trouvaient à moins de 5 points du test initial.Conclusions et pertinence: Cette étude fournit des donnéesnormatives pour les scores de nasalance pour les adultes duMoyen-Orient, ce qui peut aider à rendre le nasomètre et ladétermination de nasalance plus utile cliniquement dans cettepopulation.

204 Lebanese Medical Journal 2015 • Volume 63 (4) A.L. HAMDAN et al. – Nasalance scores in Lebanese English-speaking adults

In order for the nasometer to have clinical usefulness,normative data on normal speakers’ nasalance scores isneeded for reference. Normative data on nasalance havebeen published for white Americans, African Americans[11], Australian English-speaking children [12], Irish chil-dren [13], American English-speaking children [4], Ca-nadian children [14], Puerto Rican Spanish-speakingwomen [15], Dutch children [16], and many others.

These studies have indicated that nasalance scores varyacross languages and dialects [7], these could be lan-guage-specific, material-specific or dialect-specific. Dif-ferences in mean nasalance scores between the differentlanguages and dialects were attributed to the usage ofdifferent vowels, oral and nasal consonants across lan-guages [14,17] further substantiating that normative na-salance scores for each region and language are needed.So far, no study has examined the nasalance scores inEnglish-speaking Lebanese adults. The purpose of thisstudy is to obtain normative nasalance data for a largegroup of normal English-speaking Lebanese adults know-ing that dialectal variations have an effect on mean na-salance scores. The standard speech stimuli: the Zoo pas-sage, the Rainbow passage and nasal sentences will beused. It is not the purpose of this study to provide lan-guage specific material or stimuli or to look at the effectof language on nasalance scores.

MATERIALS AND METHODS

ParticipantsA total of 102 subjects were recruited in the study, 26 wereexcluded, therefore, 77 English-speaking Lebanese adultvolunteers were enrolled in the study after having readand signed the informed consent approved by the Insti-tutional Review Board at the American University ofBeirut Medical Center (AUBMC) in Lebanon. Exclusioncriteria included history of hearing loss, recent upper res-piratory tract infection, and history of nasal or velopha-ryngeal surgery. The data were collected over a course oftwo months at the “Hamdan Voice Unit” at the AUBMC.

Materials and ProceduresThree reading passages originally described by Fletcher[18] and utilized in most clinical and nonclinical studies onnasalance [7,11,19] were used in this study: 1) the Zoo pas-sage, which contains no nasal consonants, 2) the Rainbowpassage, in which the occurrence of phonemes is similar totheir occurrence in conversational speech, and 3) a set ofnasal sentences loaded with nasal consonants.

The Nasometer II 6450 from Kay Elemetrics [5] wasused for data collection and analysis. As previously de-scribed, this computer-based system includes headgearthat contains two microphones, one on either side of aseparator plate, which allows for distinct determinationof nasal and oral contributions in speech [7]. The samedevice, computer, and headgear were used for all sub-jects. Prior to each testing session, the nasometer wascalibrated according to the manufacturer’s instructions.

After informed consent was obtained and basic demo-graphic information (age, gender, and history of smokingand reflux) recorded for each participant, the headgear wasplaced using previously described techniques [20]. Theparticipant was then asked to read the Zoo passage,Rainbow passage, and nasal sentences, in that order, onetime each. All testing was carried out by one of two re-searchers, following the same protocol, in a quiet room,where only the participant and researcher were present.

Data AnalysisMean, minimum, and maximum nasalance scores foreach passage for each participant were computed usingthe “Calculate” function available in the Nasometer 6450software package. Descriptive analysis was used to reportthe mean nasalance scores, standard deviations and rangefor each stimulus. Unpaired t-tests and ANOVA were usedto assess differences across studies and between genderswithin this study.

Seven participants were randomly selected for reliabili-ty testing at a later stage. This involved following the sameprocedures as the initial test, but with each passage beingread twice. The comparison was made by comparing themeans of the two readings in the second session to thecorresponding value collected in the initial session. Thisallowed for comparisons both within and across sessions.

RESULTS

Demographic DataA total of 102 participants were recruited, however 26subjects were excluded because they didn’t fit the study’sinclusion criteria, a total of 77 participants were thus in-cluded in the final analysis.

The mean age was 23.77 ± 4.295 years old, with arange from 19 to 40. Thirty-nine participants were males,and 38 females, with no significant difference in agebetween the two groups (males: 23.41 ± 2.704, females:24.13 ± 5.418; p = 0.46). Fourteen participants (18.2%)were smokers and 14 (18.2%) reported having reflux(Table I).

TABLE IDEMOGRAPHIC INFORMATION for STUDY POPULATION (N = 77)

AgeMean ± SD 23.77 ± 4.295

Minimum 19Maximum 40

GenderMale 39 (50.6%)

Female 38 (49.4%)Smoking

Yes 14 (18.7%)No 61 (81.3%)

RefluxYes 14 (18.7%)No 61 (81.3%)

A.L. HAMDAN et al. – Nasalance scores in Lebanese English-speaking adults Lebanese Medical Journal 2015 • Volume 63 (4) 205

Nasal ValuesThe group mean nasalance scores for the Zoo passage,Rainbow passage and nasal sentences were 25.21 ± 11.07,34.04 ± 9.30 and 41.29 ± 9.87, respectively. As expected,the mean nasalance score was highest for the nasal sen-tences and lowest for the Zoo passage. For the Zoo pas-sage, the mean nasalance score was slightly higher inmales compared to females (26.51 ± 11.66 vs. 23.87 ±10.42), but the difference was not significant (p = 0.30).For the Rainbow passage and nasal sentences, the meannasalance scores were comparable for the male andfemale groups (Table II).

Reliability TestThe difference in the mean nasalance scores within ses-sions and between sessions varied with the reading stimuli.For within-session variability, at least 86% of retests forall three passages were < 5 nasalance points differentfrom the initial test. In the case of the Zoo passage andRainbow passage, there was one outlier participant rep-resenting the remaining 14% whose retest was 8 nasal-ance points different for the Zoo passage and 7 pointsdifferent for the Rainbow passage.

For across-session variability, at least 93% of retestsfor all three passages were < 5 nasalance points differ-ent, again with one test representing the remaining 7% inthe case of the Rainbow passage and nasal sentences.For the Zoo passage, 100% of the retests resulted in adifference in mean nasalance score within 5 nasalancepoints (Table III).

DISCUSSION

Multiple instruments have been used to estimate or com-pute the extent of nasalance in normal English-speakingadults and children. These instruments and techniqueshave also been relied upon for the evaluation of hyper-nasality and hyponasality in subjects with velopharyn-geal incompetence or cleft palate deformity. Since itsintroduction in 1986, the nasometer has been used as aresearch tool and as a noninvasive method for assessingnasal resonance related to velopharyngeal insufficiency[8,21] and nasal obstruction [22]. A nasalance score re-flects the dispersion of nasal acoustic energy in the over-all vocal tract. Its perceptual correlate is the perceivednasality in speech.

Several studies have evaluated the relationship be-tween perceptual ratings of nasality and nasalance usingcorrelation analysis and sensitivity and specificity mea-sures. The correlation between nasalance and nasalityhas varied between 0.02 and 0.82 [8]. This range in vari-ability has been attributed to differences in methodologyrelated to variations in speech stimuli, rating scales andnumber of listeners. Speech stimuli loaded with nasalconsonants tend to increase the correlation between per-ceptual rating and hyponasality whereas speech stimulidevoid of nasal consonants increase the correlation withhypernasality. Several studies indicate that nasalancescores vary across languages and dialects [17], an obser-vation that mandates the presence of separate normativedata to be used as cutoffs for what is normal and what is

TABLE IINASALANCE SCORES of the THREE PASSAGES for the TOTAL STUDY POPULATION

Zoo Passage Rainbow Passage Nasal Passage

Mean value ± SD [Range]

Overall population (N = 77) 25.21 ± 11.07 [9-49] 34.04 ± 9.30 [14-56] 41.29 ± 9.87 [20-73]

Males (N = 39) 26.51 ± 11.66 [9-49] 34.05 ± 10.46 [14-56] 41.21 ± 12.08 [20-73]

Females (N = 38) 23.87 ± 10.42 [9-43] 34.03 ± 10.08 [24-48] 41.37 ± 7.08 [32-55]

SD: standard deviation

TABLE IIICUMULATIVE FREQUENCIES in PERCENT of EACH of the SEVEN SUBJECTS' REPEATED READINGS

WITHIN SESSION (7 MEASUREMENTS) and ACROSS SESSIONS (14 MEASUREMENTS).

Difference Zoo Passage Rainbow Passage Nasal Passage

(Nasalance points) Within session Across sessions Within session Across sessions Within session Across sessions

< 1 0 7 0 14 14 14< 2 43 36 14 43 57 29< 3 71 57 71 64 71 79< 4 71 79 86 79 71 86< 5 86 100 86 93 100 93< 6 86 86 93 93< 7 86 86 93 100< 8 86 100 100< 9 100


abnormal in different cultures, languages and even forspeakers of the same language in different regions. Theusage of pre-existing normative data derived primarilyfrom Western countries may be misleading in the assess-ment and management of patients with nasalance prob-lems in other contexts. Thus, having normative data for agiven language or at least foreign-English-speaking indi-viduals is necessary in order to better evaluate and man-age speakers with resonance disorders. Very few studieshave looked at the effect of dialect on nasalance scores inEnglish-speaking adults [17]. Most of the studies havebeen done in North America. Dalston et al. reported dif-ferent cutoff scores in different regions of North America[17]. Speakers of American English have been reportedto have significant differences in their nasalance scoresacross different geographic regions in the United States[7]. Another study showed no significant differences be-tween white and African American men in the meannasalance scores for the Zoo passage [11]. The normativescores for the Zoo passage, Rainbow passage, and nasalsentences have been reported as 11.25 ± 5.63, 31.47 ±6.65, and 59.55 ± 7.96, respectively, for AmericanEnglish-speaking adults [7]. In other English-speakingcountries, Van Doorn and Purcell have reported norma-tive data for Australian English-speaking children, andfound that mean nasalance scores on the Zoo passagewere 2 points lower in their population compared toAmerican English-speaking children [12]. Sweeney et al.examined a group of Irish-speaking children and report-

ed the normative scores for the total speech sample, highpressure consonant sentences, low-pressure consonantsentences and nasal consonant sentences to be 26%, 14%,16% and 51%, respectively, with no significant differ-ences among genders [13].

In our study, the mean nasalance score for the Rain-bow passage was comparable to the scores reported in theliterature for different English dialects, whereas scores forthe Zoo passage and nasal sentences were quite differentfrom the results of other studies. For the Rainbow pas-sage, the nasalance score in English-speaking Lebaneseadults (34.04 ± 9.30) was well within the range of 31.69 ±5.47 to 35.2 ± 4.7 reported for Canadian adults [19] andAmerican adults from various regions [7]. For the Zoopassage, the results differed by 4-12 points in comparisonto the normative data reported for American English-Speaking adults from different regions [7], by almost 7points in comparison to White Americans and 8 pointscompared to African Americans [11]. All differences forthis passage were statistically significant, (p = 0.032 forstudy population compared to Mid-Atlantic U.S. adults,the group with the closest value). For the nasal sentences,the nasalance score was markedly lower than those report-ed by other studies (16-19 points). Similarly, all differ-ences were statistically significant (p < 0.0001 for studypopulation compared to any other group). Table IV dis-plays the mean nasalance scores obtained in this studyalongside with those from readings of the same passagesin previously published studies [7,11-12,19].

TABLE IVNASALANCE SCORES of the THREE PASSAGES for the STUDY POPULATION - FOREIGN ENGLISH-SPEAKING LEBANESE ADULTS,

SEVERAL ENGLISH SPEAKING ADULT POPULATIONS in the U.S. and CANADA, and AUSTRALIAN CHILDREN

Population N Mean age ± SD Zoo Passage Rainbow Passage Nasal Passage[Range]

Lebanon, Adults 77 23.77 ± 4.295 25.21 ± 11.07 34.04 ± 9.30 41.29 ± 9.87(39 M;38 F) [19-40]

*Mid-Western U.S., Adults1 148 15 ± 6 35 ± 5 62 ± 6

*Mid-Atlantic U.S., Adults1 148 21 ± 5 39 ± 6 65 ± 5

*Southern U.S., Adults1 148 13 ± 7 34 ± 6 61 ± 6

*Ontario, Canada, Adults1 148 12 ± 6 36 ± 7 61 ± 7

African American Adults2 80 23.2 ± 5.33 17.05 ± 5.2 57.74 ± 5.3(40 M;40 F)

White American Adults2 80 23.2 ± 5.33 18.24 ± 4.2 60.89 ± 6.6(40 M;40 F)

Toronto, Ontario, Adults3 76 26.5 ± 5.8 13.45 ± 5.9 31.69 ± 5.47 57.9 ± 6.69(25 M;51 F)

Australian Children4 245 6.5 ± 1.5 13.1 ± 5.9 59.6 ± 8.1(122 M;123 F)

For all the *subjects as one group, the mean age was 33.07 years with a range of 16.17-63.33 and the male (M) to female (F)ratio: 56 males and 92 females.1Seaver et al., 1991, using Nasometer 6200. 2Mayo et al., 1996, using Nasometer 6200. 3Bressman, 2005, using Nasometer 6200.4Van Doorn & Purcell, 1998, using Nasometer 6200.

A.L. HAMDAN et al. – Nasalance scores in Lebanese English-speaking adults Lebanese Medical Journal 2015 • Volume 63 (4) 207

The significant differences in nasalance scores for theZoo passage and nasal sentences may be attributed to sev-eral factors. First and foremost is the known difference invowels and consonants across different dialects in English,as previously discussed. This provide the most reasonableexplanation for why mean nasalance scores in this Leba-nese English-speaking population showed such a smallerrange (i.e. were significantly higher for the Zoo passageand significantly lower for the nasal sentences) comparedto other English dialects. A second potential factor is thedifference in the mean age group in the reported studies.Mean age in the studies by Seaver et al. (1991), Bressman(2005) and Van Doorn and Purcell (1998) [7,12,19] wereall significantly different from the mean age in this study(Table IV). The possible effect of such differences is un-clear. Trindade et al. (1997) found that children, comparedto adults, had significantly lower nasalance scores fornon-nasal passages [23]. At the same time, other studiesthat have compared adults and children have found no sig-nificant differences in nasalance scores between the twogroups [4,7], concluding that factors such as language ordialect more directly influence nasalance than does age[13]. A third potential effect relates to the nasometer used;the Nasometer 6450 was used in this study while the ver-sion used in all comparison studies was the Nasometer6200. In a comparison of the Nasometer 6400 series andNasometer 6200 series, Watterson et al. (2005) found thatthe former tends to yield slightly higher means than the lat-ter [19]. This difference was statistically significant, butprobably due mostly to within-participant performance in-consistency and variation in headgear placement [19].While this machine variation might help explain some ofthe discrepancy in our findings related to the Zoo passage,we would have anticipated such an (elevating) effect for allthree passages. Aside from the variations of mean nasal-ance scores in relation to language- and dialect-specificstimuli, gender may have an impact as well. Gender’seffect has been attributed to physiological differences,among other methodological discrepancies [24-25].Whether female speakers have significantly higher nasal-ance scores than male counterparts has been an issue ofdebate, with no clear consensus on whether such a differ-ence exists or if it is significant [7,23,26]. Many authorshave indicated that women have higher nasalance scoresfor language-specific stimulus material [7,26]. These dif-ferences were more pronounced on passages containingnasal consonants, as reported by Van Lierde et al. (2001)[26] and Seaver et al. (1991) [7]. Others have reported theopposite, indicating that men have higher mean nasalancescores than women [24,27]. Even when statistically signif-icant differences in the mean nasalance scores were presentbetween genders, these were in the range of only 2 scalarpoints and thus carry little clinical significance. Reflectingthe equivocal literature, there was no considerable differ-ence based on gender in our study. Scores were roughlyequal between males and females for the Rainbow passageand the nasal sentences, and there was a slight but insignif-icant difference (p = 0.3) for the Zoo passage.

CONCLUSIONS

This study helps fill a gap in clinical knowledge by pro-viding normative data for nasalance scores in a LebaneseEnglish-speaking population. Our data can subsequentlybe employed to direct the use of then nasometer as a clin-ical assessment tool in this population. Compared to nor-mative data from other English-speaking populations, theoverall range of mean nasalance scores in this populationwas smaller, with significantly higher values for the Zoopassage and significantly lower values for the nasal sen-tences. This is most likely due to differences in vowelsand oral and nasal consonants across different dialects.No significant difference in nasalance scores was foundbetween males and females. Similar studies could bedone in the future in order to corroborate our study find-ings in the Middle East population.

ACKNOWLEDGMENTS

This manuscript has no actual or potential conflict ofinterest.

Please note that Dr. Abdul Latif Hamdan had fullaccess to all the data in the study and takes responsibilityfor the integrity of the data and the accuracy of the dataanalysis.

REFERENCES

1. Bradford LJ, Brooks AR, Shelton RL Jr. Clinical judg-ment of hyper-nasality in cleft palate children. CleftPalate J 1964; 1: 329-35.

2. Fletcher SG. Theory and instrumentation for quantitativemeasurement of nasality. Cleft Palate J 1970; 7: 601-9.

3. Fletcher SG. “Nasalance” vs. listener judgements ofnasality. Cleft Palate J 1976; 13: 31-44.

4. Fletcher SG, Adams LE, McCucheon MJ. Cleft palatespeech assessment through oral-nasal acoustic measures.In: Bzoch KR, ed. Communicative Disorders Related toCleft Lip and Palate. Boston: Little, Brown 1989: 246-57.

5. Kay Elemetrics. Nasometer II Model 6450 InstructionManual. Lincoln Park, NJ: Kay Elemetrics, 1994.

6. Dalston RM, Warren DW, Dalston ET. A preliminaryinvestigation concerning the use of nasometry in indenti-fying patients with hyponasality and/or nasal airwayimpairment. J Speech Hearing Res 1991a; 34: 11-18.

7. Seaver EJ, Dalston RM, Leeper HA, Adams LE. A studyof nasometric values for normal nasal resonance. JSpeech Hear Res 1991; 34: 715-21.

8. Nellis JL, Nieman GS, Lehman JA. Comparison ofnasometer and listener judgments of nasality in theassessment of velopharyngeal functions after pharyngealflap surgery. Cleft Palate Craniofac J 1992; 29: 157-63.

9. Dalston RM, Warren DW. Comparison of Tonar II, pres-sure flow, and listener judgments of hypernasality in theassessment of velopharyngeal function. Cleft PalateCraniofac J 1986; 23: 108-15.

10. Hardin MA, Van Demark DR, Morris HL, Payne MM.Corrrespondence between nasalance scores and listenerjudgments of hypernasality and hyponasality. Cleft


Palate Craniofac J 1992; 29: 346-51.11. Mayo R, Floyd LA, Warren DW, Dalston RM, Mayo CM.

Nasalance and nasal area values: cross-racial study. CleftPalate Craniofac J 1996; 33: 143-9.

12. Van Doorn J, Purcell A. Nasalance levels in the speech ofnormal Australian children. Cleft Palate Craniofac J 1998;35: 287-92.

13. Sweeney T, Sell D, O’Regan M. Nasalance scores fornormal-speaking Irish children. Cleft Palate Craniofac J2004; 41: 168-74.

14. Leeper HA, Rochet AP, Mackay IRA. Characteristics ofnasalance in Canadian speakers of English and French. In-ternational Conference on Spoken Language Processing,Banff, Alberta, Canada, October 1992: 49-52. Abstract.

15. Anderson RT. Nasometric values for normal Spanish-speaking females: a preliminary report. Cleft PalateCraniofac J 1996; 33: 333-6.

16. Van der Heijden P, Hobbel HH, Van der Laan BF, Korsten-Meijer AG, Goorhuis-Brouwer SM. Nasometry normativedata for young Dutch children. Int J Pediatr Otorhino-laryngol 2011; 75: 420-4.

17. Dalston RM, Neiman GS, Gonzalez-Landa G. Nasomet-ric sensitivity and specificity: a cross-dialect and cross-culture study. Cleft Palate Craniofac J 1993; 30: 285-91.

18. Fletcher SG: Diagnosing Speech Disorders from CleftPalate, New York: Grune and Stratton, 1978: 92-157.

19. Bressman T. Comparison of nasalance scores obtainedwith the Nasometer, the NasalView, and the OroNasalSystem. Cleft Palate Craniofac J 2005; 42: 423-33.

20. Watterson T, Lewis K, Brancamp T. Comparison of na-salance scores obtained with the Nasometer 6200 and theNasometer II 6400. Cleft Palate Craniofac J 2005; 42:574-9.

21. Dalston RM, Warren DW, Dalston ET. Use of nasometryas a diagnostic tool for identifying patients with velopha-ryngeal impairment. Cleft Palate Craniofac J 1991b; 28:184-9.

22. Dalston RM, Warren DW, Dalston ET. The identificationof nasal obstruction through clinical judgments of hypo-nasality and nasometric assessment of speech acoustics.Am J Orthod Dentofac Orthop 1991c; 34: 11-18.

23. Trindade IEK, Genero KF, Dalston RM. Nasalancescores of normal Brazilian Portuguese speakers. Braz JDysmorphol Speech Hear Disord 1997; 1: 23-34.

24. Litzaw LL, Dalston RM. The effect of gender upon na-salance scores among normal adult speakers. J CommunDisord 1992; 25: 55-64.

25. Zajac D, Lutz R, Mayo R. Microphone sensitivity as asource of variation in nasalance scores. J Speech HearRes 1996; 39: 1228-31.

26. Van Lierde KM, Wuyts FL, De Bodt M, Van Cauwenberge P.Nasometric values for normal nasal resonance in thespeech of young Flemish adults. Cleft Palate Craniofac J2001; 38: 112-18.

27. Kavanagh JL, Fee EJ, Kalinowski J, Doyle PC, Leeper HA.Nasometric values for three dialectal groups within theAtlantic Provinces of Canada. J Speech Lang PatholAudiol 1994; 18: 7-13.

INTRODUCTION

The human’s voice perceptual quality has many perspec-tives. It is, not only, an indicator of health and sickness,but also carries information on the speaker’s identity,culture, language, dialect and intellectual status [1]. As ameans of communication, the acoustic perceptual quali-ties of the speaking voice are influenced by race, lan-guage, intonation and cultural effects. As such, theseseem to be major variables in determining the speakingfrequencies and carry differences in the vocal character-istics as well.

The cross-linguistic differences in the acoustic para-meters of the speaking voice, namely the formant fre-quencies and their dispersion, have been attributed notonly to the cross-cultural differences in patterns ofstressing and intonation while speaking, but also to the

different morphological features of the speakers [2].Because formants are affected by the shape and configu-ration of the vocal tract and since these are under theinfluence of genetic and cultural differences, we antici-pate that groups of different languages, cultures and eth-nicities to have different acoustics in terms of frequen-cies and formant characteristics [3-5]. Yet listeners ingeneral and speech language pathologists in particularbuild a perceptual evaluation of what is abnormal basedon their inherent understanding and knowledge of whatis a normal voice and what is deviant from a normalvoice, a judgment that is mostly founded on normativedata collected invariably from Western population. Otherfactors that might affect their judgment are their own cul-ture and manner in communication [6].

With the presence of empirical data in the literaturedemonstrating some differences in the spectral measure-ments and speaking frequencies across different culturesand ethnicities, and with the assumption that vowels maynot consistently carry acoustic data in a universal fashion,we elected in this study to report the formant character-istics of English-speaking Lebanese men during steadyprolongation of vowels /a/ and /i/.


AARRTTIICCLLEE OORRIIGGIINNAALL// OORRIIGGIINNAALL AARRTTIICCLLEEFORMANT CHARACTERISTICS OF ENGLISH-SPEAKING LEBANESE MENhttp://www.lebanesemedicaljournal.org/articles/63-4/original7.pdf

Abdul Latif HAMDAN1, Georges ZIADE1, Zeid AL-ZAGHAL2, Dollen TABRI3, Solara SINNO3

Rami SAADE2, Jad JABBOUR3, Jihad NASSAR2

Hamdan AL, Ziade G, Al-Zaghal Z, Tabri D, Sinno S, Saade R,Jabbour J, Nassar J. Formant characteristics of English-speakingLebanese men. J Med Liban 2015 ; 63 (4) : 209-212.

Hamdan AL, Ziade G, Al-Zaghal Z, Tabri D, Sinno S, Saade R,Jabbour J, Nassar J. Caractéristiques des formants d’hommeslibanais anglophones. J Med Liban 2015 ; 63 (4) : 209-212.

Departments of 1Otolaryngology, 2Otolaryngology-Head &Neck Surgery, 3Medical student, American University ofBeirut Medical Center, Lebanon.

Correspondence: Abdul Latif Hamdan, MD. AUB - MedicalCenter, P.O.Box 110236. Beirut. Lebanon.

e-mail: [email protected]

ABSTRACT • Objective : To report the formant characteris-tics of English-speaking Lebanese men during steady prolon-gation of vowels /a/ and /i/. Design : Cross-sectional studyinvolving volunteer participants. Participants : Fifty healthyEnglish-speaking males with a mean age of 32 ± 4.027 yearsand a range of 42 years (18-60). Materials and Methods : Atotal of 50 healthy males between the age of 18 and 60 wererecruited for the study. Each subject was asked to phonate asustained /a/ and /i/ sound at a comfortable pitch and intensi-ty level. Measures were made in real-time and formant fre-quencies across F1, F2, F3 and F4 were determined usingthe Real-time Spectrogram VP 3950 (Kay Elemetrics, NewJersey). Main outcome measures : The formant frequen-cies across F1, F2, F3 and F4 were recorded. Results: Forthe vowel /a/, the mean values of F1, F2, F3 and F4 were622.86 ± 61.293Hz, 1264 ± 78.602Hz, 2610.90 ± 206.359Hzand 3483.56 ± 206.833 Hz, respectively. For the vowel /i/, thecorresponding values were 378.88 ± 51.825 Hz, 2210.34 ±124.077 Hz, 2847 ± 168.770 Hz and 3576.82 ± 242.760 Hz,respectively. Conclusion: Formant characteristics vary amongcultures and ethnic groups.

Keywords : formant characteristics, vowel, male

RÉSUMÉ • Objectifs : Signaler les caractéristiques des for-mants d’hommes libanais anglophones au cours de la pro-longation stable des voyelles /a/ et /i/. Protocole : Étudetransversale. Participants: Cinquante hommes anglo-phones avec un âge moyen de 32 ± 4,027 ans et un éven-tail de 42 ans (18-60). Interventions : Il a été demandé àchaque sujet de vocaliser les voyelles /a/ et /i/ et de lessoutenir pour une certaine durée avec un ton et un niveaud’intensité confortables. Les fréquences en temps réel ontété enregistrées, et les formants à travers F1, F2, F3 et F4ont été déterminés et ce en utilisant le spectrogramme entemps réel VP 3950 (Kay Elemetrics, New Jersey). Princi-paux critères de jugement : Les fréquences des formantsà travers F1, F2, F3 et F4 ont été enregistrées. Résultats :Pour la voyelle /a/, les valeurs moyennes de F1, F2, F3 etF4 étaient 622,86 ± 61,293Hz, 1264 ± 78,602Hz, 2610,90 ±206,359Hz et 3483,56 ± 206,833Hz, respectivement. Pour lavoyelle /i/, les valeurs correspondantes étaient de 378,88 ±51,825 Hz, 2210,34 ± 124,077 Hz, 2847 ± 168,770 Hz et3576,82 ± 242,760 Hz, respectivement. Conclusion : Lescaractéristiques des formants varient selon les cultures etles groupes ethniques.

210 Lebanese Medical Journal 2015 • Volume 63 (4) A.L. HAMDAN et al. – Formant characteristics of Lebanese men

MATERIALS AND METHODS

ParticipantsA total of 50 healthy males between the age of 18 and 60were recruited for the study. Subjects with recent historyof upper respiratory tract infection, laryngeal manipula-tion or surgery, or history of chronic disease known toaffect the voice were excluded from the study. All sub-jects were informed about the study, which was approvedby the Institution Review Board of the American Uni-versity of Beirut, Lebanon.

Data Collection and AnalysisAfter having read and signed the consent form, each sub-ject was seated in a quiet room in front of a unidirec-tional condenser microphone at a constant mouth-to-microphone distance of 10 cm. Each subject was askedto phonate a sustained /a/ and /i/ sound at a comfortablepitch and intensity level. Measures were made in real-time, and formant frequencies across F1, F2, F3 and F4were determined using the Real-time Spectrogram VP3950 (Kay Elemetrics). The Sono Match module, usingthe Real-Time FFT window, was used to confirm theresults obtained by the Real-Time Spectrogram. The cur-sor was placed at the centermost point of the steady-stateformant band when looking at the spectrogram.

The mean, standard deviation (SD) and range valueswere derived for all variables. Outliers were detected bythe means of a box-plot. Statistical analyses were carriedout with SPSS (Version 17). Statistical significance wasset at p-value < 0.05.

RESULTS

Demographic DataAll subjects were males with a mean age of 32 ± 4.027years and a range of (18-60 years). The mean weight andheight were 80.684 ± 13.01 Kg and 179.84 ± 7.147 cm,respectively (Table I).

Formant Frequencies of Vowels /a/ and/i/For the vowel /a/, the mean values and ranges of F1, F2, F3and F4 were 622.86 ± 61.293 Hz (403-735), 1264.16 ±78.602 Hz (1067-1446), 2610.90 ± 206.359 Hz (2110-3153)and 3483.56 ± 206.833 Hz (3035-3841), respectively.

For the vowel /i/, the corresponding values were

378.88 ± 51.825 Hz (285-474), 2210.34 ± 124.077 Hz(1897-3134), 2847.52 ± 168.770 Hz (2537-3177) and3576.82 ± 242.760 Hz (2533-4149), respectively (Table II).

DISCUSSION

A review of the literature revealed numerous reportsdemonstrating that listeners can invariably identify thespeaker’s age, gender and dialect when the requiredspeaking tasks and duration of voice sample are con-trolled [7]. Likewise, racial and phonetic characteristicsof the speaker could be identified with good reliabilityduring steady vowel prolongations [7-10].

These perceptual evaluations have been substantiatedby other acoustic studies documenting the differences infundamental frequencies and formant values among var-ious ethnic groups. In a study by Andrianopolus et al.,fundamental frequency (F0) and formant frequenciescharacteristics during sustained vowels varied in relationto race and gender [11-12]. Along the same line ofthought, Altenberg and Ferrand have reported highermean speaking fundamental frequency (SF0) duringconnected speech in English-speaking Russian femaleswhen speaking English and Russian compared to themean SF0 of the monolingual English-speaking controlgroup [13]. Similarly, there has been numerous reportsshowing no consensus regarding the presence of signifi-cant differences in the acoustic values between nativeand English-speaking adults [14-15]. In comparison toAmericans, it appears that for the vowel /a/, the secondformant F2 is consistently lower in the English-speakingLebanese men whereas formant F3 is rather comparable

TABLE IDEMOGRAPHIC DATA

Mean value ± SD[Range]

Age (years)32.0 ± 4.027

[18-60]

Height (cm)179.84 ± 7.147

[163-196]

Weight (kg)80.684 ± 13.0181

[59.1-123.4]

TABLE IIFORMANT FREQUENCIES of VOWELS / a / and / i /

Mean value ± SD [Range]

(Hz) F1 F2 F3 F4

Vowel /a/622.86 ± 61.293 1264.16 ± 78.602 2610.90 ± 206.359 3483.56 ± 206.833

[403-735] [1067-1446] [2110-3153] [3035-3841]

Vowel /i/378.88 ± 51.825 2210.34 ± 124.077 2847.52 ± 168.770 3576.82 ± 242.760

[285-474] [1897-3134] [2537-3177] [2533-4149]

A.L. HAMDAN et al. – Formant characteristics of Lebanese Men Lebanese Medical Journal 2015 • Volume 63 (4) 211

[16-17]. According to Hillenbrand et al. [16] formant F1is lower in Americans compared to English-speakingLebanese men, whereas based on Lee et al. study F1 ishigher [18]. Likewise for the vowel /i/, F1 is consistent-ly higher and F2 and F3 are consistently lower in English-speaking Lebanese men compared to Americans, eventhough the differences are not significant [16-17]. Severalexplanations can be proposed to account for the cross-lingual differences in speakers’ formant frequencies. Theseinclude vocal tract configurations and linguistic factors.The intrinsic relationship between vocal tract configura-tion and formant frequencies has been clearly confirmedin the literature [19]. Initially due to the limitations ofimaging studies the focus was on the vocal tract lengthbut more recent studies using acoustic reflection andmagnetic resonance imaging have helped in further dis-secting the relationship between the two [2].

Vowel formant frequencies are one of many frequent-ly used acoustic parameters for the analysis of a speak-er’s speech signal. They are also intimately related to theshape and configuration of the vocal tract which againvaries with culture, language, dialect and other geneticand social factors. By altering the shape of the vocaltract, we can change the desired frequencies of vibrationor the preferred resonating frequencies, known as for-mants. These are also affected by the position of thetongue, with F1 for instance being lower when thetongue is at the roof of the mouth, whereas F2 is morerelated to the frontness and backness of the highest partof the tongue [19]. F3 seems to be more important indetermining the phonemic quality of a speech sound, andF4 and F5 reflect more the quality of voice. Aside fromthe inherent factors such as race, culture and other phys-iological factors that are major determinants of our vocalfrequencies, are the acquired pathological changes and/or the surgical induced alterations in the vocal tract. Theeffect of vocal tract alterations on voice has been report-ed by numerous studies investigating formant character-istics in subjects undergoing surgery of upper aero-digestive system. Reports on formants frequencies fol-lowing partial laryngectomy, glossectomies, tonsillec-tomies, palatal surgeries and other interventions [20-22],substantiate further the intrinsic relationship betweenvocal tract configuration and voice.

In order to draw objective conclusions on what isnormal or abnormal, non-biased formant frequencyanalysis among groups of different cultures, ethnicitiesand native languages must be taken into consideration.The results of our study clearly indicate that there aredifferences in formant characteristics between English-speaking Lebanese men and Americans for the vowels/a/ and /i/. These differences must be considered in theanalysis of normal and/or pathological voice samples.This study carries certain limitations. One is the limitednumber of vowels examined; it would have been moreinformative to have the formant characteristics of theremaining vowels as well. Second is the fact that ourdata is limited to men only.

CONCLUSION

Formant characteristics vary among cultures and ethnicgroups. The cross-ethnic and cross-cultural variationsought to be taken into consideration by speech languagepathologists. The formant frequencies of English-speak-ing Lebanese men vary in comparison to Americans.Examining the differences between formant frequenciesof vowels /a/, /i/ and /u/ of trilingual English-, French-and Arabic-speaking subjects compared to monolingualArabic-speaking subjects will be the core investigationof the future study.

ACKNOWLEDGMENTS

This manuscript has no actual or potential conflict ofinterest.

Please note that Dr. Abdul Latif Hamdan had fullaccess to all the data in the study and takes responsibil-ity for the integrity of the data and the accuracy of thedata analysis

REFERENCES

1. Titze IE. Workshop on Acoustic Voice Analysis.Summary Statement. Iowa City, IA: National Center forVoice and Speech. Wendell Johnson Speech & HearingCenter, University of Iowa; 1995. Accessed from http://tsvoice.net84.net/media/sumstat.pdf on Nov 13, 2013.

2. Xue SA, Hao GJ, Mayo R. Volumetric measurements ofvocal tracts for male speakers from different races. ClinLinguist Phon 2006; 20 (9): 691-702.

3. Titze IE. Physiologic and acoustic differences betweenmale and female voices. J Acoust Soc Am 1989; 85:1699-707.

4. Majewski W, Hollien H, Zalewski J. Speaking funda-mental frequency of Polish adult males. Phonetica 1972;25:119-25.

5. Hudson Al, Holbrook A. A study of the reading funda-mental frequency of young black adults. J Speech LangRes 1981; 24: 197-201.

6. Aronson A. Introduction to clinical voice disorders. In:Aronson, editor. Clinical Voice Disorders. New York,NY: Thieme, 1990: 1-9.

7. Lass NJ, Tecca JE, Mancuso RA, Black WI. The effect ofphonetic complexity on speaker race and sex identifica-tions. J Phonet 1979; 7: 105-18.

8. Peterson GE, Barney HL. Control methods used in astudy of vowels. J Acoust Soc Am 1952; 24: 175-84.

9. Bricker PD, Pruzanski S. Effects of stimulus content andduration on talker identification. J Acoust Soc Am 1966;40:1442-9.

10. Assmann PF, Neary TM, Hogan JT. Vowel identification:orthographic, perceptual, and acoustic aspects. J AcoustSoc Am 1982; 71: 975-89.

11. Andrianopoulos MV, Darrow KN, Chen J. Multimodalstandardization of voice among four multicultural popu-lations: Fundamental frequency and spectral characteris-tics. J Voice 2001; 15 (2):194-219.

12. Andrianopoulos MV, Darrow KN, Chen J. Multimodalstandardization of voice among four multicultural popu-

212 Lebanese Medical Journal 2015 • Volume 63 (4) A.L. HAMDAN et al. – Formant characteristics of Lebanese men

lations: Formant structures. J Voice 2001; 15 (1): 61-77.13. Altenberg EP, Ferrand CT. Perception of individuals with

voice disorders by monolingual English, bilingual Can-tonese-English, and bilingual Russian-English women.J Speech Lang Hear Res 2006; 49 (4): 879-87.

14. Walton JH, Orlikoff RF. Speaker race identification fromacoustic cues in the vocal signal. J Speech Hear Res 1994;37: 738-45.

15. Mayo R, Manning WH. Vocal tract characteristics ofAfrican-American and European-American adult males.Tex J Audiol Speech Pathol 1994; 20: 33-6.

16. Hillenbrand J, Getty LA, Clark MJ, Wheeler K. Acousticcharacteristics of American English vowels. J AcoustSoc Am 1995; 97 (5 Pt 1): 3099-111.

17. Alotaibi YA, Husain A. Formant based analysis of spo-ken Arabic vowels. Lecture Notes in Computer Science2009; 5707: 162-9.

18. Oliveira Barrichelo VM, Heuer RJ, Dean CM, SataloffRT. Comparison of singer’s formant, speaker’s ring, andLTA spectrum among classical singers and untrained nor-mal speakers. J Voice 2001; 15 (3): 344-50.

19. Lee S, Potamianos A, Narayanan S. Acoustics of chil-dren’s speech: Developmental changes of temporal andspectral parameters. J Acoustic Soc Am 1999; 105 (3):1455-68.

20. Kazi R, Prasad VM, Kanagalingam J et al. Analysis offormant frequencies in patients with oral or oropharyn-geal cancers treated by glossectomy. Int J Lang CommunDisord 2007; 42 (5): 521-32.

21. Ilk HG, Erogul O, Satar B, Ozkaptan Y. Effects of tonsil-lectomy on speech spectrum. J Voice 2002; 16 (4): 580-6.

22. Akpinar ME, Kocak I, Gurpinar B, Esen HE. Effects ofsoft palate implants on acoustic characteristics of voiceand articulation. J Voice 2011; 25 (3): 381-6.

INTRODUCTION AND OBJECTIVES

Developing academic physicians is one of the majorchallenges for healthcare centers and physicians at thesame time. Healthcare centers are interested in develop-ing academic healthcare professionals to improve quali-ty care to patients and quality teaching and training tomedical students and postgraduate trainees; whereas

academic physicians are required to fulfill multifacetedroles in teaching, research, clinical, and administrativeservices, and hence, it is essential for them to improvetheir skills in these four main areas. In this respect, it wassuggested that it is important for academicians to keeptheir growth and productivity at all career levels, in spiteof their challenges in maintaining a well-balanced role inresearch, teaching and service responsibilities [1]. Onthe other hand, it is argued that the three traditional rolesof academic physicians are competing with each other,considering the dominant compensation structure thatadds pressure on physicians to earn their own income,

Lebanese Medical Journal 2015 • Vol 63 (4) 213

MMIISSEE AAUU PPOOIINNTT// IINN--DDEEPPTTHH RREEVVIIEEWWTOWARDS DEVELOPING A SUSTAINABLE FACULTY DEVELOPMENT PROGRAM An Initiative of an American Medical School in Lebanonhttp://www.lebanesemedicaljournal.org/articles/63-4/review1.pdf

Boushra RAHAL1, Nabil MANSOUR1, Ghazi ZAATARI1,2

Rahal B, Mansour N, Zaatari G. Towards developing a sus-tainable faculty development program: An initiative of anAmerican medical school in Lebanon. J Med Liban 2015 ;63 (4) : 213-217.

Rahal B, Mansour N, Zaatari G. Vers l’élaboration d’unprogramme de développement durable pour les membres de lafaculté: Une initiative d’une faculté de médecine américaineau Liban. J Med Liban 2015 ; 63 (4) : 213-217.

1Faculty Affairs Office, Faculty of Medicine and MedicalCenter, 2Department of Pathology and Laboratory Medicine,American University of Beirut, Beirut, Lebanon.

Correspondence: Boushra Rahal, MPH, EMBANabil Mansour, MSc

e-mail: [email protected] & [email protected]

ABSTRACT • Introduction and Objectives : The AmericanUniversity of Beirut Faculty of Medicine (AUB-FM) strategy isto develop faculty members (fm) skills by sponsoring local andinternational scientific activities has been in place for overthree decades, and remains dependent on individuals’ efforts.In 2011-2012, Faculty Development Program (FDP) was in-troduced to develop faculty leadership, business skills in med-icine, fulfill personal and professional goals, followed by afive-year plan to cover five themes: Management/Leadership,Marketing, Finance, Strategic Planning and Communicationswith the purpose of integrating these themes in medical prac-tice. Methods: A survey was sent to all departments at AUB-FM in 2011 to assess needs and determine themes. Nine work-shops were conducted, followed by post-workshop evaluation.Results: 117 fm responded to needs assessment surveys.Respondents had on average 15 years in clinical practice,50% with extensive to moderate administrative experience;71% assumed administrative responsibilities at least once,56% in leadership positions. Faculty attendance dropped mid-way from 69 to 19, although workshops were rated very goodto excellent. Discussion: Although faculty were interested inFDP, the drop in attendance might be attributed to: challengesto achieve personal and professional goals while strugglingto fulfill their roles, satisfy promotion requirements and gener-ate their income. Recommendations: FDP has to be alignedwith FM strategic goals and faculty objectives, be complimen-tary to a faculty mentoring program, provide rewards, and besupported by a faculty progression tool.

Keywords : faculty development, mentoring, career growth,faculty progression

RÉSUMÉ • Introduction et objectifs : À l’Université améri-caine de Beyrouth la stratégie de la Faculté de médecine(AUB-FM) réside dans le développement des compétencesdes membres de la faculté (fm) en sponsorisant des activitésscientifiques locales et internationales ; mise en place depuisplus de trois décennies, elle reste corrélée aux efforts indivi-duels. En 2011-2012, le Programme de développement de lafaculté (PDF) a été introduit pour développer le leadership de lafaculté, les compétences relatives aux affaires en médecine etpour réaliser des objectifs personnels et professionnels. LePDF a été suivi par un plan quinquennal englobant les thèmessuivants: gestion/leadership, commercialisation, finances, pla-nification et communications stratégiques dans le but d'intégrerces thèmes en médecine. Méthodes : Un questionnaire a étéenvoyé à tous les départements de la FM en 2011 pour éva-luer les besoins et déterminer les thèmes. Neuf ateliers ont étéorganisés.puis évalués par les participants. Résultats : 117 fmont répondu à des enquêtes d’évaluation des besoins. Lesrépondants ont en moyenne 15 années de pratique clinique,50% avec une expérience comprise entre vaste et modérée enadministration; 71% ont assumé des responsabilités adminis-tratives au moins une fois et 56% ont occupé des postes deleadership. La présence fm a chuté de 69 à 19, bien que lesateliers aient été jugés très bons ou excellents. Discussion :Bien que les professeurs aient été intéressés par le PDF, lachute de la présence pourrait être attribuée à des défis visant àatteindre des objectifs personnels et professionnels tout en s’ef-forçant d’assumer leurs rôles, de satisfaire aux exigences depromotion et de créer leurs revenus. Recommandations : LePDF doit être en ligne avec les buts et les objectifs de la facultéde médecine, complémenter le programme de mentorat envigueur, offrir des récompenses et disposer d’un outil de pro-gression de faculté.

and whether they can still fulfill their traditional triplerole successfully [2]. It was also stressed that mentoringwas important in developing the career of future aca-demic physicians. The latter argument was supported bya group of researchers who reported several disadvan-tages that academic physicians suffer for not havingproper mentorship or “academic parenting,” such aslower salaries and scholarly productivity, and the timespent on learning things, compared to their peers whohad mentors [3].

Physicians who opt to pursue career in academic medi-cine face obstacles in achieving their new goals, mainlybecause their training prepared them to be medical doctors,but not tutors or researchers [4]. This notion alludes to cer-tain gaps and needs that could hinder the establishmentof a successful academic career by the physicians, whichmight be bridged by proper mentorship. Successful men-toring depends however, on the skills of the mentors andmentees, as well as on the institutional supporting environ-ment [5], and on finding a qualified mentor [6].

It is the strategy of the American University of BeirutFaculty of Medicine (AUB-FM) to develop faculty skillsthrough continuing education by sponsoring local activ-ities, and funding the attendance of an internationalmeeting for all faculty members. Such initiative has beenin place for over three decades and remains dependenton individuals’ efforts. To promote these initiatives fur-ther, and in an attempt to bridge the gaps of skills re-quired by academic physicians to accomplish their mul-tiple roles, AUB-FM introduced in 2011-2012 a year-round Faculty Development Program (FDP) [7], aiming

to develop the leadership and business skills in medi-cine, and to help them fulfilling their personal and pro-fessional goals at once.

The purpose of this study is to evaluate the FDP ini-tiative and its impact on the attendees, and to conclude alesson from this experience as a step towards establish-ing a sustainable faculty development or mentorship pro-gram at AUB-FM.

METHODS

An online survey was sent to all faculty members in the19 academic and clinical departments at AUB-FM in2011 to assess needs and determine themes (See Annex).

Nine workshops were conducted covering business inmedicine topics: • How to financially manage your prac-tice • Selection and Appraisal of personnel • Type of cor-poration: What’s in it for you? • Leadership in Medicine• Financial Investments and Liability • How to positionyourself in a competitive practice • Business Planning •Strategic Financial Management and • Conflict Manage-ment and Resolution. These workshops were also openedto senior administrators and residents. Post workshop eval-uation survey was sent to the attendees. IRB approval isnot required for this study.

RESULTS

Out of 296 faculty members, 117 responded to the first sur-vey of needs assessment at the Faculty of Medicine. Distri-bution of respondents by department is shown in Table I.

TABLE IDISTRIBUTION of RESPONDENTS by DEPARTMENT

Total Number Number % of Respondents % of Respondents by of of by Department from Total

DEPARTMENT Faculty Members Respondents Department Number of Respondents

Anesthesiology 16 7 44% 6% Biochemistry & Molecular Genetics 4 2 50% 2%

Dermatology 10 7 70% 6%Diagnoslic Radiology 8 6 75% 5%Emergency Medicine 4 4 100% 3%

Family Medicine 16 16 100% 14%Anatomy, Cell Biology & Physiology 12 5 42% 4%

Internal Medicine 71 31 44% 26%Experimental Pathology, 4 2 50% 2%

Inmunology & MicrobiologyObstetrics & Gynecology 17 7 41% 6%

Ophthalmology 13 7 54% 6%Otolaryngology, Head & Neck Surgery 14 8 57% 7%

Pathology & Laboratory Medicine 16 3 19% 3%Pediatrics & Adolescent Medicine 28 1 4% 1%

Pharmacology & Therapeutics 4 0 0% 0%Psychiatry 4 4 100% 3%

Radiation Oncology 5 2 40% 2%Surgery 50 5 10% 4%

TOTAL 296 117 40% 100%

The respondents were characterized by having an aver-age of 15 years in clinical practice, 50% had an extensiveto moderate administrative experience, and 71% of themhad assumed administrative responsibilities at least once,56% of which in leadership positions. The distribution ofrespondents by rank is shown in Figure 1.

The results of the needs assessment survey revealedthe interest and need of faculty members to developingskills in the following areas: leadership, strategic plan-ning, marketing, financial management, recruitment &selection and conflict management. The topics of work-shops were selected and conducted according to the re-sults of the needs assessment.

Overall, the attendance per workshop ranged from 69

to a minimum of 19. On average, the faculty attendancewas the highest compared to residents and administrators.However, the attendance of faculty dropped midwaythrough almost after the sixth workshop out of nine, asshown in Figure 2.

DISCUSSION

Developing faculty members and physicians is one ofthe main goals of medical centers, which provides a bet-ter career management and decrease physicians over-turn [8]. It was reported that for healthcare centers togrow, academic physicians should be developed andretained [9].

B. RAHAL et al. – Towards developing a sustainable faculty development program Lebanese Medical Journal 2015 • Vol 63 (4) 215

FIGURE 1. Percent distribution of respondents by academic rank.

How toFinancially

Manage YourPractice

Selection andAppraisal ofPersonnel

Type ofCorporation:

What’s in it foryou?

Leadership inMedicine

FinancialInvestments and

Liability

How to positionYourself in aCompetitive

Practice

Business Planning

StrategicFinancial

Management

Conflict Management and

Resolution

FIGURE 2. Attendants of the workshops by category of attendees.

Past Administrative ExperienceProfessorial Rank

Faculty member Resident Administrator Total

13%

34%

26%

27%

13%

37%

30%

20%

216 Lebanese Medical Journal 2015 • Volume 63 (4) B. RAHAL et al. – Towards developing a sustainable faculty development program

The workshops conducted took into consideration theadult learning framework [10-11] that divides the gapand needs at both organizational and individual levels.The topics were selected according to the choice of fac-ulty members themselves and their needs. However, theworkshop attendance declined gradually mostly amongjunior faculty, while the topics and lecturers were rated“very good” to “excellent.”

Although faculty were genuinely interested in FDP, asverified by the needs assessment and program evaluation,this drop in attendance might be attributed to one or moreof the factors highlighted below, to be further investigated:

A. Academic physicians are expected to dedicate 80%of their time and effort in clinical care and teachingwith the remaining 20% of their time to conduct re-search and be engaged in different administrativeduties. This structure seems to be challenging forthem to satisfy their multifaceted roles and expec-tations, and yet attend to their personal develop-ment and career progression.

B. Promotion requirements are demanding and addanother burden to academic physicians, which re-quire excellence in two areas per academic track tobe achieved within six years in rank: teaching, re-search, clinical and administrative services.

C. Compensation structure is the major burden overand above the previous two challenges. Physiciansare expected to generate two-third of their incomethrough clinical practice, which confirms the con-clusions of a group of researchers [2] that physiciansare pressured to earn their own income and fulfillthe expectations with limited resources and time.Therefore organizational support should be offeredin terms of resources, information and individual[11].

Accordingly, these academic physicians are chal-lenged to achieve personal and professional goals, whilethey are struggling to fulfill their roles. Our findings arein agreement with that highlighted by Garand et al. [1]that it is crucial for academicians to maintain theirgrowth and productivity at all career levels, in spite oftheir challenges in maintaining a well-balanced academ-ic and clinical role. Concurrently, Borges et al. [2], andMilner et al. [12] concluded that pursuing a career inacademic medicine is challenging, mainly in the absenceof institutional support, mentoring and development pro-gram. Moreover, the research findings of Kashiwagi etal. [13] and Gray et al. [14], support the importance ofmentoring junior academic physicians and developingtheir skills, although the forms and goals of such pro-grams vary from one academic center to another.

RECOMMENDATIONS

To ensure sustainability, commitment and success, werecommend that the Faculty Development Program hasto be:

• Aligned with the institutional strategic goals as well

as the academic physicians’ objectives.• Complimentary to a well-structured and institution-

alized faculty mentoring program, specifically tar-geting junior academic physicians.

• Provide protected time and academic rewards andrecognitions that would ensure continuous growthand development.

• Supported by a faculty progression tool for contin-uous monitoring of junior faculty academic pro-gress [1].

REFERENCES

1. Garand L, Matthews JT, Courtney KL et al. Developmentand use of a tool to guide junior faculty in their progres-sion: Toward promotion and tenure. Journal of Profes-sional Nursing 2010; 26 (4): 207-13.

2. Borges NJ, Navarro AM, Grover A, Hoban JD. How,when, and why do physicians choose careers in academ-ic medicine? A literature review. Academic Medicine2010; 85 (4): 680-6.

3. Jackson VA, Palepu A, Szalacha L, Caswell C, Carr PL,Inui T. “Having the right chemistry”: A qualitative studyof mentoring in academic medicine. Academic Medicine2003; 78 (3): 328-34.

4. Harris DL, Krause KC, Parish DC, Smith MU. Academiccompetencies for medical faculty. Family Medicine-Kansas City 2007; 39 (5): 343-50.

5. Sambunjak D, Marusic A. A systematic review of quali-tative research on the meaning and characteristics ofmentoring in academic medicine. Journal of GeneralInternal Medicine 2010; 25 (1): 72-8.

6. Chew LD, Watanabe JM, Buchwald D, Lessler DS.Junior faculty’s perspectives on mentoring. AcademicMedicine 2003; 78 (6): 652.

7. American University of Beirut, http://www.aub.edu.lb/FM/FAO/FD/Pages/purpose.aspx. Retrieved on Decem-ber 20, 2013.

8. Bickel J, Brown AJ. Generation X: Implications for facul-ty recruitment and development in academic health cen-ters. Academic Medicine 2005; 80 (3): 205-10.

9. Thorndyke LE, Gusic ME, George JH, Quillen DA,Milner RJ. Empowering junior faculty: Penn State’s fac-ulty development and mentoring program. AcademicMedicine 2006; 81 (7): 668-73.

10. Gilbert TF. Human competence-Engineering worthy per-formance. NSPI Journal 1978; 17 (9): 19-27.

11. Chevalier RD. When did ADDIE become addie?Performance Improvement 2011; 50 (6): 10-14.

12. Milner RJ, Gusic ME, Thorndyke LE. Perspective:toward a competency framework for faculty. AcademicMedicine 2011; 86 (10): 1204-10.

13. Kashiwagi DT, Varkey P, Cook DA. Mentoring programsfor physicians in academic medicine: A systematic re-view. Academic Medicine 2013; 88 (7): 1029-37.

14. Gray J, Armstrong P. Academic health leadership:Looking to the future. Proceedings of a workshop held atthe Canadian Institute of Academic Medicine meetingQuebec, Que., Canada, Apr. 25 and 26, 2003. Clinicaland Investigative Medicine 2003; 26 (6): 315-26.

B. RAHAL et al. – Towards developing a sustainable faculty development program Lebanese Medical Journal 2015 • Vol 63 (4) 217

ANNEX

ONLINE SURVEY DISTRIBUTED TO ALL FACULTY MEMBERS AT AUB-FACULTY OF MEDICINE

MMIISSEE AAUU PPOOIINNTT// IINN--DDEEPPTTHH RREEVVIIEEWWMANAGEMENT OF METASTATIC COLORECTAL CANCER Current Treatments and New Therapieshttp://www.lebanesemedicaljournal.org/articles/63-4/review2.pdf

Anna-Maria HENAINE1-2, Georges CHAHINE4, Pascale SALAMEH5, Edouard ELIAS4, Marcel MASSOUD4

Daniel HARTMANN1-2, Gilles AULAGNER1,2,3, Xavier ARMOIRY2,6

Henaine A-M, Chahine G, Salameh P, Élias E, Massoud M,Hartmann D, Aulagner G, Armoiry X. Management of meta-static colorectal cancer: Current treatments and new therapies.J Med Liban 2015 ; 63 (4) : 218-227.

Henaine A-M, Chahine G, Salameh P, Élias E, Massoud M,Hartmann D, Aulagner G, Armoiry X. Prise en charge du can-cer colorectal métastatique : Traitements standards et nouveau-tés thérapeutiques. J Med Liban 2015 ; 63 (4) : 218-227.

1Université de Lyon, Claude Bernard Lyon 1, Lyon, France. 2UMR CNRS 5510 MATEIS, Lyon, France3Hospices Civils de Lyon, Groupement Hospitalier Est, Service pharmaceutique, Bron, France.4Hôtel-Dieu de France (Beyrouth) & Hôpital Notre-Dame de Secours (Byblos), Liban. 5Université libanaise, Laboratoire clinique et épidémiologique, Faculté de Pharmacie, Hadath, Liban. 6Hospices civils de Lyon, Cellule Innovation/Délégation à la Recherche Clinique et à l’Innovation, Av. du Doyen Lépine, France.

Correspondence : Anna Maria HENAINE, PharmD. UMR CNRS 5510 MATEIS. 8 Avenue Rockefeller. 69373 Lyon. France.e-mail : [email protected]

RÉSUMÉ • Le cancer colorectal représente 8% des cas decancers métastatiques. Pendant des décennies, la chimio-thérapie associant le 5-Fluorouracil (5-FU) et l’acide foliniquea constitué le traitement conventionnel de référence. L’arri-vée de l’irinotécan, de l’oxaliplatine et des formes orales de5-FU au cours des années 90 a été considérée comme unegrande avancée thérapeutique. À partir de 2004, la thérapieciblée, avec les anticorps monoclonaux visant les récepteursEGF et l’angiogenèse (VEGF), a permis un bénéfice en termesde survie même si le cancer colorectal métastatique reste, àce jour, incurable. Nous présentons la prise en charge ac-tuelle du cancer colorectal métastatique, le développementclinique de ces traitements émergents et leur positionnementdans le système de santé libanais.

ABSTRACT • Colorectal cancer represents 8% of metastaticcancers. For decades, the gold standard therapy has beeninfusional chemotherapy with 5-Fluorouracil (5-FU) associ-ated to folinic acid. The discovery of irinotecan, oxaliplatinand oral forms of 5-FU in the nineties is considered a mile-stone in the treatment of this disease. Since 2004, targetedtherapy with monoclonal antibodies including anti-EGFR andangiogenesis inhibitors showed superiority in terms of mor-tality compared to conventional therapy. Metastatic colorectalcancer, however, remains an incurable disease. We presentthe current treatments of metastatic colorectal cancer, theclinical development of these emerging treatments, and theirposition in the Lebanese health care system.

Keywords: metastatic colorectal cancer, chemotherapy, mono-clonal antibodies

5-FU : 5-Fluorouracil5-FULV : 5-Fluorouracil/LeucovorinAPC : Protein codingASPECCT : A Study of Panitumumab Efficacy and safety Compared to

Cetuximab TrialBRAF : V-Raf murine sarcoma viral oncogene homolog B1CORRECT : Regorafenib monotherapy for previously treated metastatic

colorectal cancerCRC : ColoRectal CancerCRYSTAL : Cetuximab Combined With Irinotecan in First-Line Therapy

for Metastatic Colorectal CancerCYP3A4 : Cytochrome PDNA : DeoxyriboNucleic AcidEGF : Epidermal Growth FactorEGFR : Epidermal Growth Factor ReceptorFGFR : Fibroblast Growth Factor ReceptorFIRE : Trial comparing Cetuximab and bevacizumabFOLFIRI : 5-FULV + IrinotecanFOLFOX : 5-FULV + OxaliplatinFOLFOXIRI : 5-FULV + Oxaliplatin + IrinotecanGSK3B : Glycogen Synthase Kinase 3BHER : Human Epithelial Growth Factor ReceptorIg : ImmunoglobulinKRAS : Kirsten RAt Sarcoma 2 viral oncogene homologMAPK : Mitogen Activated Protein Kinase

mCRC : metastatic colorectal cancer NRAS : Neuroblastoma RAS viral oncogene homologOPUS : Oxaliplatin and Cetuximab in First-Line Treatment of Metastatic

Colorectal CancerOR : Overall responseORR : Overall response rateOS : Overall survivalPDGF : Platelet Derived Growth FactorPEAK : A Randomized, Multicenter Phase II Study of Panitumumab + Modified

Fluorouracil, Leucovorin, and Oxaliplatin (mFOLFOX6) or Bevacizumab +mFOLFOX6 in Patients With Previously Untreated, Unresectable, Wild-Type KRAS Exon 2 Metastatic Colorectal Cancer

PFD : Progression free diseasePFS : Progression free survivalPI3K/Akt : Phosphoinositide 3-kinasePIGF : Placental growth factorPRIME : Panitumumab Randomized trial In combination with chemotherapy

for Metastatic colorectal cancer to determine EfficacyRR : Response rateUGT1A : UDP-glucuronosyltransferase 1AVEGF : Vascular Endothelial Growth FactorVEGFR : Vascular Endothelial Growth Factor ReceptorVELOUR : Phase III study of aflibercept and FOLFIRI in patients with

metastatic colorectal cancer after failure of an oxaliplatin regimen.WT : Wild

LIST OF ABBREVIATIONS


BACKGROUND

Metastatic colorectal cancer (mCRC) is a major cause ofmorbidity and mortality throughout the world. It is thefourth most common cancer diagnosed in the U.S. and thesecond leading cause of cancer death in 2014. In Leba-non, colorectal cancer is the third leading cause of deathfrom cancer and the most common cancer both sex com-bined [1-2]. In 2012, according to national cancer statis-tics, the number of eligible new cases diagnosed in thehospital system was 2035, 979 men and 1056 women [3].

About 75% of patients with CRC have sporadic diseasewith no apparent evidence of having inherited the disorder.The remaining 25% of patients have a family history ofCRC suggesting a hereditary contribution, common expo-sures among family members, or a combination of both.

Numerous factors are found important in the develop-ment of CRC including high risk luminal environment,inflammation, as well as lifestyle factors such as diet,tobacco and alcohol consumption. In recent years, focushas turned towards the genetics and molecular biology ofCRC and several interesting and promising correlationsand pathways have been discovered. The major geneticpathways of CRC are the chromosome instability path-way representing the pathway of sporadic CRC throughthe KRAS, APC and p53 mutations, and the microsatelliteinstability pathway of hereditary non-polyposis throughmutations in mismatch repair genes [4].

Metastases are observed in 40-60% of cases of CRC ofwhich 25% are synchronous. Among patients withoutmetastases at diagnosis, the risk of metachronous is 30-40% [5]. Without any treatment, patients with mCRChave a median survival of seven months.

In some cases, the metastases reveal the cancer (pre-cessive metastases) whose primary site is not known andshould be identified. Approximately 30-50% of CRC pa-tients develop liver metastases, responsible alone for twothirds of the deaths. Being able to consider their resectionis a major challenge since the resection significantly im-proves the survival of patients. The extra-hepatic loca-tions are rare and of poor prognosis.

Currently, a multitude of clinical evidences have sheda new light on the issues such as the selection and dura-tion of therapy, the associations with targeted agents andtreatments tailored towards the clinic and molecular fac-tors. Moreover, knowledge of prognostic and predictivebiomarkers such as mutation in KRAS, NRAS, and BRAF,is of growing interest to allow a better selection of drugsand treatment strategies.

The aim of this paper is to review the updates to themanagement of mCRC.

THERAPEUTIC STRATEGIES in the TREATMENT of mCRC

The objectives of the management of patients withmCRC are generally palliative rather than curative, withthe exception of a small proportion of patients with re-sectable liver metastases and controllable by radiofre-

quency (or cryotherapy) or locoregional treatment bydirect injection of radioactive spheres [6].

Recent advances in chemotherapy regimens have in-creased the median overall survival (OS) from 6 monthsto over 30 months through the targeted therapy and dou-ble or triple combination protocols [7].

The three active agents for mCRC, known as conven-tional chemotherapy, are fluoropyrimidines (5-FU or itsoral prodrug), irinotecan and oxaliplatin. The associationsFOLFOX/XELOX and FOLFIRI are commonly used inthe Lebanese hospitals [8] with no evidence from litera-ture of differences between both regimens in terms of OS[9]. Triple associations such as FOLFOXIRI induce amedian OS without disease progression with greater re-sponse rate than FOLFIRI regimen, but are associatedwith a poor safety profile [10]. In fact, the FOLFOXIRIhas been compared to FOLFIRI in two randomized clini-cal trials in unresectable patients: in both studies, FOLFO-XIRI led to an increased in R0 secondary resection rates:6% vs 15% (p = 0.033) in the Gruppo Oncologico NordOvest (GONO) trial and 4% vs 10% (p = 0.08) in theGastrointestinal Committee of the Hellenic OncologyResearch Group (HORG) trial. In a follow-up study of theGONO trial, the 5-year survival rate was higher in thegroup receiving FOLFOXIRI (15% vs 8%) with a medianOS of 23.4% vs 16.7% month (p = 0.026) [11-13].

The addition of targeted therapies to these four cyto-toxic regimens has shown better outcomes in the major-ity of cases. However, the best strategy sequence and therespective roles of each combination are still debated.

PRINCIPAL AGENTS USED in the TREATMENT of mCRC

Table I summarizes the principal drugs used to treatmCRC.

Cytotoxic chemotherapy drugs Cytotoxic drugs induce acute cell death (necrosis), pro-grammed cell death (apoptosis), growth arrest or differen-tiation. Many antineoplasic drugs have multiple actionson the cell:

5-FU is activated to 5-fluorodeoxyuridylate, which,in the presence of a reduced folate co-factor, inhibits theenzyme thymidylate synthase. This blocks the produc-tion of thymidine phosphate which is required for DNAsynthesis. The 5-fluorodeoxyuridylate may also be fraud-ulently incorporated into DNA causing a form of DNAdamage [14].

Capecitabine is an orally bioavailable 5-FU prodrugthat undergoes sequential hydrolysis and deaminationreactions in the liver to produce 5’-deoxy-5-fluorouri-dine. It is converted to 5-FU by thymidine phosphorylase(also known as platelet-derived growth factor). As thisenzyme is abundant in tumor tissue there is some tumorspecificity in the patient’s exposure to 5-FU.

The adverse effects, such as hand and foot syndrome,of capecitabine resemble those of 5-FU when given byprotracted infusion.

A.-M. HENAINE et al. – Metastatic colorectal cancer: new therapies Lebanese Medical Journal 2015 • Volume 63 (4) 219

220 Lebanese Medical Journal 2015 • Volume 63 (4) A.-M. HENAINE et al. – Metastatic colorectal cancer: new therapies

Other oral 5-FU prodrugs (e.g. S-1, UFT) havebeen the subject of extensive clinical trials [15].These are mostly combinations of 5-FU prodrugswith uracil, which is an inhibitor of dihydropyrimi-dine dehydrogenase, a ubiquitous enzyme that rapid-ly degrades 5-FU.

Oxaliplatin is a diaminocyclohexane platinum deriva-tive and a bifunctional alkylator capable of reactingwith adjacent guanine residues in DNA. It provides eitherintra- or inter-strand DNA cross-links, which interferewith DNA processing.

Oxaliplatin was evaluated as the first line metastaticcombination with 5-FULV (FOLFOX4). Three trialsshowed a significant improvement in the objective re-sponse and a progression-free survival, but no significantdifference in term of OS [16]. The two limiting toxici-ties of FOLFOX4 were neutropenia and the specificreversible sensory neuropathy of oxaliplatin. The resultsof the OPTIMOX 1 study [17] showed that a “stop andgo” approach using oxaliplatin free-interval resulted in adecrease in neurotoxicity without affecting the OS.Therefore, the panel recommends adjusting the schedule/timing of the administration of this drug as a means oflimiting this adverse effect. The current clinical dataare insufficient to support the routine use of calcium/magnesium infusions to prevent the oxaliplatin relatedneurotoxicity.

Irinotecan is a topoisomerase I inhibitor used to treatseveral solid tumor types, especially in combination withother chemotherapeutic agents in the treatment of mCRC.Inhibition of topoisomerase I by irinotecan and its activemetabolite, SN-38, prevents re-ligation of single strandedDNA breaks induced during DNA synthesis phase of cel-lular replication. Cell death ultimately occurs. Adverseeffects include severe diarrhea, myelosuppression andneutropenia, likely induced by inefficient metabolism and

excretion of SN-38, which undergoes glucuronidation pri-marily in the liver by UGT1A prior to excretion throughthe kidneys. UGT1A locus is alternatively spliced to pro-duce nine isoenzymes and the UGT1A1 isoform is solelyresponsible for the phase II metabolism of bilirubin, nu-merous endogenous hormones and pharmacologic com-pounds, including irinotecan. Thus, genetic variation inUGT1A correlates with adverse events caused by irinote-can toxicity [18].

Evidence indicates that, at relatively high irinotecanlevel (> 250 mg/m2), patients who are homozygous forthe UGT1A1*28 variant experience a greater risk ofclinically important neutropenia, especially with partic-ular chemotherapy agents/oxaliplatin [19]. That’s why itis reasonable to determine the UGT1A genetic back-ground to assist in toxicity management. Unfortunately,this molecular testing is not present in our Lebanese hos-pitals to date.

Irinotecan is also an inhibitor of acetyl cholinesteraseand patients may experience an acute onset of cholinergic-like symptoms including lacrimation, sweating, abdomi-nal cramping and diarrhea.

Irinotecan, in combination with fluorouracil(FOLFIRI), in the second-line, is considered a goodstandard in the palliative treatment of mCRC. Severalstudies [20-21] have shown a significant increase inthe response rate, PFS and OS with FOLFIRI com-pared to 5-FULV alone.

MODULATORS of the TARGETED BIOLOGICAL RESPONSE

As observed in a retrospective cohort study conductedin Lebanese hospitals [8], approximately 75% of pa-tients with mCRC receive a biological agent duringtheir treatment course targeting one of two main path-ways: the EGFR cascade and the VEGF signaling.

TABLE ISUMMARY of APPROVED THERAPIES for METASTATIC COLORECTAL CANCER in LEBANON

Drug Class Mode of Pathway administration

CHEMOTHERAPY5-Fluorouracil (5-FU®) Antimetabolite (pyrimidine antagonist) IV DNA / RNA damage

Capecitabine (Xeloda®) Antimetabolite (pyrimidine antagonist) PO DNA / RNA damage

Oxaliplatine (Eloxatine®) Alkylating agent (Platinum compound) IVDNA / RNA damage(nonspecific cell cycle)

Irinotecan (Campto®) Topoisomerase inhibitor IV DNA damageTARGETED AGENTS

Cetuximab (Erbitux®) Monoclonal antibody (IgG1 chimeric) IV PI3K / Akt, MAPKPanitumumab (Vectibix®) Monoclonal antibody (fully human) IV PI3K / Akt, MAPKBevacizumab (Avastin®) Monoclonal antibody (recombinant humanized) IV Angiogenesis

Aflibercept (Zaltrap®) Recombinant fusion protein IV Angiogenesis

Regorafenib (Stivarga®) Tyrosine kinase inhibitor POAngiogenesis, oncogenesistumor microenvironment

Ig: immunoglobulin MAPK: mitogen-activated protein kinase PI3K/Akt: phosphoinositide 3-kinase IV: intravenous PO: per os


Triggering cell death via cancer-directed immunother-apy or immunomodulation with the aim to overcomemajor mechanisms of immune resistance is a newlyarising field [22-23]. We will discuss the majorapproaches and current therapies in the treatment ofmCRC.

Inhibitors of angiogenesisAngiogenesis is a central mechanism in human CRCdevelopment and growth [24]. In particular, VEGF isclosely associated with the induction of neovasculariza-tion in human colon cancer and is one of the most impor-tant endogenous ligands of receptors present on the en-dothelial cell plasma membrane; it’s binding leads tointracellular signaling and, ultimately, gene transcriptionthat promotes endothelial proliferation, migration, andtube formation of endothelial cells resulting in tumorgrowth and metastasis.

Given the crucial importance of neoangiogenesis inmCRC, anti-VEGF pathway therapies have been inten-sively investigated. The monoclonal antibody bevacizu-mab is the first antiangiogenic agent to be approved inmCRC treatment, because of response rate and survivalbenefit [25-26].

The three antiangiogenic agents available withinLebanese hospitals are presented in Table II.

Bevacizumab is a recombinant humanized IgG1monoclonal antibody (7% murine) directed againstVEGF-A.

After showing its efficacy in several murine xenograftmodels of human tumors, bevacizumab alone or in com-bination with other cytotoxic agents, was shown to bewell tolerated in phase I trials apart from some sideeffects such as thromboembolic events, bleeding, hyper-tension and proteinuria.

The association of bevacizumab with 5-FU in-creased the OS (12.9 months for FOLFOX4/beva-cizumab vs 10.8 months for FOLFOX4 alone) and thedisease free progression (DFP) significantly (7.3 monthsvs 4.7 months respectively). In November 2012, beva-cizumab was approved in combination with chemo-therapy based on fluoropyrimidine, irinotecan or oxali-platin for the treatment of mCRC disease. This approvalallows the patients, who received first-line treatmentwith bevacizumab plus irinotecan-based chemotherapyor with oxaliplatin, to continue receiving bevacizumabplus other chemotherapy based on oxaliplatin or irino-tecan as second-line treatment after progression of thedisease. In the Lebanese hospitals, the practice is topreferably use the protocol FOLFOX-Bevacizumab asfirst-line [8].

Aflibercept also known as VEGF Trap, is a fullyhuman soluble recombinant protein composed of extra-cellular domains from both VEGFR-1 and VEGFR-2fused to the Fc (a) region of human IgG1 [27]. It wasdesigned to act as a decoy VEGF receptor, bindingVEGF-A, VEGF-B, and PlGF, subsequently preventingtheir interactions with VEGFR-1 and VEGFR-2 [28-29].

TAB

LE II

SU

MM

ARY

of th

eAN

TIAN

GIO

GEN

IC D

RU

GS

CU

RR

ENTL

YAP

PRO

VED

form

CR

C T

REA

TMEN

T:BE

VAC

IZU

MAB

, AFL

IBER

CEP

T&

REG

OR

AFEN

IBB

evac

izum

abA

flibe

rcep

tR

egor

afen

ibA

ppro

val d

ate

2004

Augu

st 2

012

Sept

embe

r 201

2R

egis

trat

ion

in L

eban

on20

0620

1420

14O

rigin

Gen

ente

ch In

c., S

outh

San

Fra

ncis

co, C

A, U

SASa

nofi-

Aven

tis, B

ridge

wat

er, N

J, U

SABa

yer A

G, L

ever

kuse

n, G

erm

any

Mul

tikin

ase

inhi

bito

r aga

inst

VEG

FR1,

VEG

FR2,

Mec

hani

sm o

f act

ion

Targ

ets

VEG

F-A

(onl

y)VE

GFR

dec

oy b

indi

ng V

EGF-

A,VE

GF-

B&

PIG

FVE

GFR

3, P

DG

FR-β

, FG

FR1,

TIE

2, K

IT, R

ETas

wel

l as

a si

gnal

tran

sduc

tion

inhi

bito

r of

the

RAF

/MEK

/ER

K pa

thw

ayEx

trac

ellu

lar/

Intr

acel

lula

rEx

trace

llula

rEx

trace

llula

rIn

trace

llula

r

Indi

catio

n1st

/2nd

line

mC

RC

+ c

ombi

natio

n ch

emot

hera

py

2ndlin

e in

mC

RC

+ F

OLF

IRI

3rdlin

e m

CR

C a

s si

ngle

age

nt, i

n re

fract

ory/

prog

ress

ive

dise

ase

Dos

ing

5-10

mg/

kg IV

eve

ry 2

wee

ks in

com

bina

tion

with

4 m

g/kg

IV e

very

2 w

eeks

in c

ombi

natio

n w

ith

160

mg

daily

for 2

1 da

ys o

f an

ever

y 28

d cy

cle

FOLF

OX

or F

OLF

IRI

FOLF

IRI

Com

mon

& c

lass

ical

lyH

yper

tens

ion,

pro

tein

uria

, thr

ombo

embo

lism

,Hy

perte

nsio

n,pr

otei

nuria

, dia

rrhea

, bon

e m

arro

w H

yper

tens

ion,

ast

heni

a, g

astro

inte

stin

al

side

effe

cts

gast

roin

test

inal

per

fora

tion

(rare

),su

ppre

ssio

n (n

eutro

peni

a), d

elay

woun

d he

alin

g,

perfo

ratio

n (ra

re),

hem

orrh

age,

han

d-fo

ot s

yndr

ome,

hem

orrh

age,

del

ay w

ound

hea

ling,

RPL

Sfis

tula

, RPL

S, th

rom

boem

bolis

mra

sh, h

epat

otox

icity

, RPL

S

Bla

ck b

ox w

arni

ngs

Arte

rial &

ven

ous

thro

mbo

embo

lism

, W

ound

hea

ling

impa

irmen

t, he

mor

rhag

e,H

epat

otox

icity

(rar

e), R

PLS,

deh

ydra

tion,

hype

rtens

ion,

wou

nd h

ealin

g im

pairm

ent.

gast

roin

test

inal

per

fora

tion

(rare

)ca

rdia

c di

seas

e

Ren

al fu

nctio

n, p

rote

inur

ia, b

lood

pre

ssur

e,R

enal

func

tion,

CBC

, inf

ectio

n, p

rote

inur

ia,

Bloo

d pr

essu

re, l

iver

func

tions

, ele

ctro

lyte

sM

onito

ring

para

met

ers

blee

ding

bloo

d pr

essu

re, b

leed

ing,

deh

ydra

tion

(if d

iarrh

ea),

INR

(if t

he p

atie

nt is

taki

ng w

arfa

rin),

callo

us in

the

hand

s or

feet

FGFR

:fibr

obla

st g

row

th fa

ctor

rece

ptor

P

DG

FR: p

late

let-d

erive

d gr

owth

fact

or re

cept

or

PIG

F: p

lace

ntal

gro

wth

fact

or

VEG

F: v

ascu

lar e

ndot

helia

l gro

wth

fact

or

VEG

FR: v

ascu

lar e

ndot

helia

l gro

wth

fact

or re

cept

orFO

LFIR

I: 5-

FULV

/Irin

otec

an

mC

RC

: met

asta

tic c

olor

ecta

l can

cer

RPL

S: re

vers

ible

pos

terio

r leu

koen

ceph

alop

athy

syn

drom

e

CB

C: c

ompl

ete

bloo

d co

unt


Aflibercept was studied in combination with FOLFIRIin a prospective, randomized, double-blind, multicenter,study on patients with mCRC previously treated with anoxaliplatin-based chemotherapy regimen. Patients wererandomized either to FOLFIRI plus aflibercept (4 mg/kgintravenously every 2 weeks) or FOLFIRI plus placebo.The aflibercept arm showed significant improvement inthe primary endpoint of OS (13.50 months vs 12.06months). PFS and ORR were also increased (6.90 monthsvs 4.67 months and 19.8% vs 11.1% respectively). Insubgroup analyses of patients with or without prior-bevacizumab regimen, PFS was significantly improvedin both. OS was significantly improved in the bevaci-zumab-naïve subset, but did not achieve significance inthe prior-bevacizumab subset.

Based on these results, the FDA approved, on Au-gust 2012, the use of aflibercept in combination withFOLFIRI for the treatment of patients with mCRC thatis resistant to or that has progressed after an oxaliplatin-containing regimen. Both antiangiogenic-associatedand chemotherapy-related adverse effects were seenwith greater incidence in the aflibercept arm (83.5%of grade 3 or 4 adverse events). The antiangiogenic tox-icities described were hypertension, hemorrhage, andarterial and venous thromboembolism, diarrhea, andasthenia.

Regorafenib is an oral small-molecule multikinase in-hibitor, inhibiting multiple oncogenic kinases, PDGF,FGFR and VEGFR-1. Its mechanism is not clearly under-stood, but studies have shown that it induces expression ofPUMA, p53 target and critical mediator of apoptosis incolorectal cells after inhibition of ERK and activation ofGSK3B through the cascade of NF-kB [30-31].

Regorafenib was approved for the treatment of mCRCpatients previously treated with fluoropyrimidine, oxali-platin, or irinotecan-based chemotherapy regimens, withprior anti-VEGF and/or anti-EGFR (KRAS wild) thera-pies. Regorafenib was evaluated in the Phase III COR-RECT trial [32] on patients with mCRC who had pro-gressed during or within three months after the abovetherapies. Patients on the study were randomized to re-ceive either regorafenib (160 mg/day/21 days) or placebo.Treatment with regorafenib was associated with a modestbut significant increase in survival (6.4 months vs 5.0months). The CORRECT trial showed a survival benefitwith regorafenib in an extensively treated patient popula-tion, also about 40% of patients achieved disease stabili-zation. Adverse events associated with regorafenib arehand-foot skin reaction, fatigue, diarrhea, mucositis, dys-phonia, weight loss, infection, hypertension (impaired an-giogenesis). Hepatotoxicity (sometimes fatal), increasedincidence of hemorrhage, myocardial ischemia, infarction,reversible post leukoencephalopathy syndrome, gastro-intestinal perforation and fistula are rare and should beclosely monitored.

The bioavailability of regorafenib is 69% after oraladministration in tablet form. Absorption is dependenton the fat content of a meal (better with low-fat diet). It

is metabolized by CYP3A4 and UGT1A9 to two activemetabolites (M-2 & M-5) and is excreted primarily bythe feces [33].

Regorafenib may interact strongly with CYP3A4 in-ducers (rifampin, phenytoin, and carbamazepine),CYP3A4 inhibitors (ketoconazole, voriconazole, grape-fruit juice, clarithromycin, itraconazole) and substratesof UGT1A1 (irinotecan) [34].

EGFR (epidermal growth factor receptor) inhibitorsEGFR is a transmembrane cell surface glycoproteinand a member of the human epithelial growth factorreceptor (HER) family of tyrosine kinases. Binding ofligands to the EGFR leads to homodimerization or het-erodimerization of HER family members and results inautophosphorylation of intracellular tyrosine residues,initiating several key cell-signaling pathways includ-ing RAS/RAF/MAPK axis, involved in cell prolifera-tion and PI3K/AKT pathway, involved in cell survivaland motility. Binding to EGFR enhances processescritical to tumor growth and progression such as an-giogenesis, apoptosis inhibition, tumor invasivenessand metastatic spread [35]. RAS is one of the mostimportant molecules in the signaling pathway down-stream of the EGFR. Three human RAS genes havebeen identified: HRAS, KRAS and NRAS. Since theactivation of the EGFR leads to activation of KRAS, itwas hypothesized that mutations in signaling pathwaydownstream of the EGFR may result in receptor inde-pendent pathway activation that renders the tumorsunresponsive to EGFR blockage at the cell surface.Mutations in KRAS codons 12 and 13 (exon 2) arepresent in 35-45% of CRC with high concordancebetween primary and metastatic lesions and have beenestablished as predictors of lack of response to anti-EGFR treatment. Consequently, KRAS exon 2 muta-tion testing is widely implemented for selection ofpatients with wild-type mCRC for anti-EGFR treat-ment. A moderate response rate of 40-60%, also inKRAS exon wild-type tumors, has motivated assess-ment of the predictive role of additional alterations,primarily within the RAS-RAF-MAPK and PI3K-AKT-mTOR pathways. Additional mutations in KRASas well as NRAS (effector binding domains identicalto KRAS and present in 10% of mutations) have beensuggested to infer resistance to anti-EGFR therapiesand the European label for panitumumab treatmentwas recently modified to require testing also forKRAS mutations outside of exon 2 and for NRASmutations. The added treatment predictive value ofmutation in BRAF (principal effector of KRAS andpresent in 8% of mutations), PIK3CA and PTEN is moreuncertain [36]. According to the most current data, theBRAF biomarker is deemed to be a negative predictivebiomarker as well as a prognostic biomarker.

The BRAF and KRAS status must be determined tobe wild-type to produce a positive response to anti-EGFR therapies.

Figure 1 shows the important role of the BRAF &KRAS status in the response of anti-EGFR drugs.

Testing is based on lesions with greater than 50% tumor.This is a tumor percentage that can be easily obtained froman untreated primary lesion or a well-defined metastasiswithout treatment or extensive inflammatory infiltrate. It isestimated that 20% of patients present with metastatic dis-ease, so mutation testing based on a primary lesion is notalways feasible. Although most studies have found thatmetastases maintain the mutation profile of the primarylesion in more than 90% of cases, 54-60% discordancemay arise for technical reasons or, rarely, because of meta-stases from separate primaries. For this reason, it is prefer-able to test the metastases rather than the primary tumor, ifpossible, because it’s the former that is being treated by theEGFR targeted therapy.

Two monoclonal antibodies that block EGFR signal-ing and prevent endogenous receptor activating ligandsare currently approved for the treatment of mCRC: cetu-ximab and panitumumab. The efficacy of these drugs islimited to patients with wild-type (non-mutated) KRAS(KRAS WT) (cetuximab) and non-mutated RAS (pani-tumumab) tumors.

A multitude of recent clinical trials have indicatedthat expanded RAS testing is more effective than tradi-tional KRAS testing for patients with CRC. As a resultof these findings, the number of patients with CRC whoare eligible for EGFR inhibitors has been reduced toapproximately 40% [37].

Cetuximab is a recombinant human/mouse chimericIgG1 monoclonal antibody that binds specifically to theextracellular domain of the EGFR on both normal andtumor cells, and competitively inhibits the binding ofEGF and other ligands, such as transforming growth fac-tor-α. The recommended dose of cetuximab, in combi-nation with irinotecan or as monotherapy, is 400 mg/m2

as an initial loading dose administered as a 120-minuteIV infusion. The recommended weekly maintenance doseis 250 mg/m2 infused over 60 minutes. WT KRAS statusis required for initiation of treatment with cetuximab butfurther studies also show that WT RAS status (mutations atexons 2, 3 and 4 of KRAS, NRAS and BRAF) is mandatoryprior to initiation of treatment with cetuximab using a vali-dated method of analysis, newly applied in Lebanon.

The addition of cetuximab improved the resection rateof initially unresectable, liver-limited, KRAS WT mCRC

FIGURE 1KRAS (Kirsten rat sarcoma viral oncogene homolog) / BRAF (V-Raf murine sarcoma viral oncogene homolog B1): mutations and anti-EGFR

therapy in mCRC. Anti-EGFR drugs block receptor signals thus preventing downstream events such as KRAS or BRAF mediated effects.

WT KRAS/ WT BRAF

MUT

KRAS / MUT

BRAF

NUCLEUS

When the status of KRAS and/or BRAF aremutated (MT), the cell cycle progression

continues: cell proliferation, survival, invasion & metastases

CELL CYCLE PROGRESSION

EGFR signaling pathways

EGF

EGFR

Response to drug following the status ofKRAS and/or BRAF non mutated (WT):cell proliferation, survival, invasion &metastases are blocked

Monoclonal antibodies

(cetuximab or panitumumab)

KRAS

BRAF

MEK

ERK

Nuclear cellular abnormally increasedEffect of monoclonal anti-EGFR antibodies

when compared with the corresponding chemotherapyregimen alone. The efficacy of cetuximab was derivedfrom two randomized controlled trials. The CRYSTALstudy is a phase III, multicenter, open-label randomizedcontrolled trial, which compared cetuximab in combina-tion with FOLFIRI with FOLFIRI alone. The OPUSstudy trial is a phase II, multicenter, open-label random-ized controlled trial, which compared cetuximab in com-bination with FOLFOX4 with FOLFOX4 alone [38-39].The results of the CRYSTAL trial for the KRAS WT sub-group showed a statistically significant increase in PFSwith cetuximab in combination with FOLFIRI (9.9 monthsvs 8.7 months). The OPUS trial also showed a statisticallysignificant increase in PFS with a median PFS of 7.7months for cetuximab in combination with FOLFOXcompared with 7.2 months for FOLFOX alone. The RRwas also improved: 60.7% in the cetuximab + FOLFOXvs 37.0% FOLFOX alone. The OPUS trial included30.5% patients with additional RAS mutations; those inthe cetuximab + FOLFOX arm had inferior survival, PFSand objective RR than those assigned to receive FOLFOXalone. Safety evaluations showed no new findings attrib-utable to cetuximab when comparing WT and mutatedRAS populations: the inferior outcome was due to lack ofefficacy in combination with the known toxicity profile.As a result, some evidence of WT RAS status and ofKRAS (and NRAS) is required before initiating treatmentwith cetuximab and that this drug is contraindicated incase of mCRC with mutant RAS or unknown RAS status.The most serious adverse reactions observed in clinicaltrials of cetuximab, alone or in combination with irino-tecan, were infusion reactions, dermatologic toxicity(acneiform rash), interstitial lung disease, fever, sepsis,renal dysfunction, pulmonary embolism, dehydration anddiarrhea.

Panitumumab is a fully humanized IgG2 monoclonalantibody that binds to the extracellular domain of theEGFR of both tumor and normal cells with high affinity.By competing with endogenous ligand binding, panitu-mumab inhibits receptor phosphorylation and activationof EGFR associated cell signaling [40]. This also resultsin down regulation of EGF receptors through receptorinternalization, induction of apoptosis, autophagy andinhibition of angiogenesis. Panitumumab was approvedfor use by the FDA in September 2006 but registered inLebanon since 2014. It is administered intravenously ata recommended dosing schedule of 6 mg/kg once every2 weeks. The elimination half-life of panitumumab isapproximately 7 days, longer than cetuximab where thehalf-life is 4 days, providing the rationale for the weeklyschedule of cetuximab versus the alternate weekly admin-istration schedule for panitumumab. It also has dual clear-ance mechanisms. Specific clearance of panitumumab in-volves binding to EGFR, internalization into the cell anddegradation in lysosomes.

The evidence to support the use of panitumumab in thefirst-line treatment setting comes from the PRIME study(phase III trial) and two phase II trials [41-44]. The

PRIME trial randomized untreated patients with mCRCto FOLFOX chemotherapy and panitumumab versusFOLFOX chemotherapy. The results of the study areshown in table III. Patients with WT KRAS mCRC whoreported grades 2 to 4 skin toxicities (rash, acneiform der-matitis, pruritus, dry skin, skin fissures or erythema), diar-rhea and electrolytes deficiencies (hypokalemia) hadlonger PFS and OS as well as ORR. The development ofskin toxicity is an early clinical indicator of the panitu-mumab regimen’s efficacy. This updated analysis confirmsthe value of adding panitumumab to FOLFOX in the first-line treatment of patients whose tumors are determinedto be WT KRAS. Because crossover to an EGFR-targetedmonoclonal antibody after progression on FOLFOX waspermitted, potential survival differences may be attenuatedrelated to panitumumab. PRIME also confirms thatpatients whose tumors are KRAS-mutated should not betreated with panitumumab. The recently completed PEAKstudy (also panitumumab in first-line treatment but phaseII trial) is the first head-to-head analysis of FOLFOX com-bined with either panitumumab or bevacizumab in KRASWT mCRC (Table III).

The use of panitumumab in second-line treatment ofKRAS WT mCRC comes from two large phase III rando-mized trials, and two single-arm phase II trials (Table III).

To summarize, the addition of a biological agent to achemotherapy backbone containing oxaliplatin or irino-tecan may increase the number of patients potentiallyeligible for resection and improve outcomes.

COMPARATIVE DATABETWEEN mCRC TARGETED THERAPIES

The ASPECCT trial is an open-labeled, randomized, phaseIII trial designed to compare the effects of panitumumaband cetuximab for monotherapy of chemorefractory pa-tients with KRAS WT tumors. The ORR was 22% for thepanitumumab arm compared to 19.8% for the cetuximabarm. The PFS was 4.1 months in patients treated with pani-tumumab versus 4.4 months in patients treated with cetuxi-mab. The study met its primary endpoint, demonstratingthat panitumumab was non-inferior to cetuximab for OS(10.4 months vs 10 months). Therefore, based on theseresults and those from other previous trials, there is notherapeutic preference in clinical practice for using cetuxi-mab versus panitumumab either as monotherapy or incombination with chemotherapy in naïve or chemorefrac-tory mCRC.

Treatment selection is an art more than an algorithm.The first step is to take into account the goal of therapyand the toxicity profile that the patient is willing to livewith.

The use of anti-EGFR antibody treatment in combi-nation with chemotherapy in earlier lines of treatment isa less common and more complex issue. For RAS WTpatients the optimal targeted agent in first-line therapy re-mains undefined with the addition of either bevacizumabor an anti-EGFR antibody supported by multiple phase III


trials. Several factors may lead the clinician to prefer ananti-EGFR antibody to bevacizumab. In patients with acontraindication to bevacizumab, the addition of anti-EGFR antibody to chemotherapy improves PFS and ORRcompared with chemotherapy alone.

The consistent effects on response and survival fromRAS mutations outside of KRAS exon 2 clearly suggeststhat mutation testing for ant-EGFR prediction shouldmove from selective KRAS testing to broader RAS testing.Recently, data from the FIRE-3 study [45-46], which isthe first head-to-head comparison between two biologicalagents for the treatment of mCRC, have been designed tocompare FOLFIRI + cetuximab with FOLFIRI + bevaci-zumab in the first-line setting for patients with KRASexon 2 WT mCRC. The primary endpoint of this study, theORR, was not met (62% for FOLFIRI-cetuximab vs 58%for FOLFIRI-bevacizumab). No difference in PFS wasobserved (10.0 months vs 10.3 respectively), however, asignificant difference in OS was observed in favor ofFOLFIRI-cetuximab (28.7 months vs 25.0 respectively).In a preliminary analysis, the effects of mutations in KRASand BRAF on ORR, PFS, and OS were evaluated. Patientswith WT RAS had a median OS of 33.1 months in theFOLFIRI-cetuximab group vs 25.6 months with FOLFIRI-bevacizumab. However, in patients with mutant RAS, therewas no difference between the two regimens. Therefore, acomplete determination of RAS status may be necessary

before prescribing EGFR inhibitors in first-line treatmentof mCRC, and the prescription of cetuximab should berestricted to patients with WT RAS (and WT KRAS). Thusfar, the factors that contribute to the difference in OS areunclear.

Figure 2 summarizes the therapeutic strategies in thetreatment of metastatic colorectal cancer.

CONCLUSION and PERSPECTIVES

The outlook for the adjuvant treatment of CRC resultsout of the progress in the treatment of mCRC for the lastfive years. Such progress has been marked by the adventof combination therapies and the very recent emergenceof biotherapy improving the effectiveness of chemother-apy. With the discovery of selective targeted drugs, a newera for the treatment of mCRC has been opened. Addingthese drugs, such as bevacizumab and/or cetuximab, tostandard chemotherapy resulted in improved PFS andOS in patients with mCRC, either as first or second-lineof therapy.

Considering the substantial cost of targeted therapiesfor mCRC, their cost-effectiveness has also to be mea-sured within the Lebanese society’s perspective. Thistype of evaluation is currently ongoing based on datafrom two Lebanese university hospitals [8].


TABLE IIIPROGRESSION-FREE SURVIVAL and OVERALL SURVIVAL in FIRST & SECOND TRIALS of PANITUMUMAB [46]

TRIALS KRAS status Treatment regimen PFS Hazard Median OS Hazard ORR(control) (months) ratio (month) ratio

First linePRIME(N = 1183) KRAS WT FOLFOX4 + P 9.6 0.8* 23.9 0.83 55%

(FOLFOX4) 8.0 19.7 48%KRAS MT FOLFOX4 + P 7.3 1.29* 15.5 1.24 40%

(FOLFOX4) 8.8 19.3 40%PEAK(N = 285) KRAS WT FOLFOX + P 10.9 0.87 Not reached 0.72 58%

FOLFOX + Bev 10.1 25.4 54%Second line

(N = 1186) KRAS WT FOLFIRI + P 5.9 0.73* 14.0 0.85 35%Peeters et al. FOLFIRI 3.9 12.5 0.94 10%

KRAS MT FOLFIRI + P 5.0 0.85 11.8 13%FOLFIRI 4.9 11.1 14%

PICCOLO(N = 460) KRAS WT IrPan NR 0.78* 10.4 1.01 34%*

Irinotecan NR 10.9 12%SPIRITT KRAS WT FOLFIRI + P 7.7 1.01 18.0 1.06 32%(N = 182) FOLFIRI + Bev 9.2 21.4 19%

CI: 95% confidence interval PFS: progression-free survival OS: overall survival ORR: objective response rate N: number of patientsKRAS: Kirsten rat sarcoma gene FOLFOX: infusional 5-Fluorouracil/folinic acid/oxaliplatin P: panitumumab MT: mutated WT: wild-type (nonmutated) Bev: bevacizumab FOLFIRI: infusional 5-Fluorouracil/folinic acid/irinotecan IrPan: irinotecan/panitumumab NR: not reported * p < 0.05

REFERENCES

1. Siegel R, Ma J, Zhaohui Z, Jennel A. Cancer Statistics,2014. CA Cancer J Clin 2014; 64 (2): 104-17.

2. American Cancer Society. Cancer facts and figures, 2014.3. www.moph.gov.lb/Publications/cancer statistics 2012. 4. Raskov H, Pommergard HC, Burtcharth J et al. Colo-

rectal carcinogenesis, updates and perspectives. World JGastroenterol 2014 Dec 28; 20 (48): 18151-64.

5. Venook A, Curley S. Management of potentially resect-able colorectal cancer liver metastases. Jan 2015 Uptodate.

6. Bismuth H, Adam R, Vibert E. L’apport de la chimiothé-rapie dans le traitement des métastases hépatiques colorec-tales. Académie nationale de Chirurgie 2008; 7 (2) : 20-5.

7. Temraz S, Mukherji D, Shamseddine A. Sequencing of treat-ment in metastatic colorectal cancer: Where to fit the target?World J Gastroenterol 2014 Feb 28; 20 (8): 1993-2004.

8. Henaine AM, Chahine G, Salame P et al. Management ofmCRC in Lebanese hospitals and estimation of directcost: a retrospective cohort study. Oral communicationpresented at the Lebanese Order of Pharmacists 22nd

Pharmaceutical Congress 24-26 October 2014, Hilton

Habtoor Grand Hotel, Sin El Fil, Beirut, Lebanon.9. Tourmigand C, Andre T, Achille E et al. FOLFIRI fol-

lowed by FOLFOX6 or the reverse sequence in advancedCRC: a randomized GERGOR study. J Clin Oncol 2004;22: 229-37.

10. Masi G, Vasile E, Loupakis F et al. Randomized trial oftwo induction chemotherapy regimens in metastatic colo-rectal cancer: An updated analysis. J Natl Cancer Inst2011; 103 (1): 21-30.

11. Falcone A, Ricci S, Brunetti I et al. Phase III trial of infu-sional fluorouracil, leucovorin, oxaliplatine & irinotecan(FOLFOXIRI) compared with infusional fluorouracil,leucovorin & irinotecan (FOLFIRI) as 1st line treatmentfor mCRC: the Gruppo Oncologico Nord Ovest (GONO).J Clin Onc 2007; 25 (13): 1670-6.

12. Falcone A, Cremolini C, Masi G et al. FOLFOXIRI/bevacizumab vs FOLFIRI/bevacizumab as 1st line treat-ment for mCRC: results of the phase III TRIBE trial byGONO group. 2013 ASCO Annual meeting.

13. Gruenberger T, Bridgewater JA, Chau I et al. Random-ized phase II study of bevacizumab with mFOLFOX6 orFOLFOXIRI in patients with initially unresectable liver


FIGURE 2Chemotherapy for advanced or metastatic disease applied in the Lebanese hospitals [8].

FOLFIRI: infusional 5-Fluorouracil/folinic acid/irinotecan FOLFOX: infusional 5-Fluorouracil/folinic acid/oxaliplatin KRAS: Kirsten rat sarcoma geneNRAS: RAS oncogene WT: wild type or non-mutated

Chemotherapy for metastatic colorectal cancer in a patient appropriate for intensive therapy

FOLFOX or XELOX or FOLFOX + Bevacizumab or

XELOX + Bevacizumab

Regorafenib or clinical trial or best supportive care

ORInitial therapy

Therapy after1st progression

Therapy after 2nd progression

Therapy after 3rd progression

Cetuximab or Panitumumab (KRAS/NRAS WTgene only) + Irinotecan or Regorafenib

Regorafenib (if not given previously) or clinical trial or best supportive care

FOLFOX + Cetuximab orPanitumumab (KRAS/NRAS WT

gene only)

FOLFIRI or FOLFIRI + Bevacizumab or FOLFIRI + ziv-Aflibercept or Irinotecan or

Irinotecan + Bevacizumab or FOLFIRI + Cetuximab or Panitumumab

(KRAS/NRAS WT gene only) or Irinotecan + ziv-Aflibercept or Irinotecan + Cetuximab orPanitumumab (KRAS/NRAS WT gene only)

FOLFIRI or FOLFIRI + Bevacizumab or FOLFIRI + ziv-Aflibercept or Irinotecan or Irinotecan +

Bevacizumab or Irinotecan + ziv-Aflibercept

metastasis from CRC: resectability & safety in OLIVIA.Ann Oncol 2015; 26 (4): 702-8.

14. Miura K, Kinouchi M, Ishida K et al. 5FU metabolism incancer and orally administrable 5FU drugs. Cancers2010; 2: 1717-30.

15. Twelve C, Gollins S, Grieve R, Samuel L. A randomizedcross-over trial comparing patient preference for oralcapecitabine and 5-FULV regimens in patients withadvanced colorectal cancer. Ann Oncol 2006; 17: 239-45.

16. Grothey A, Deschler B, Kroening H et al. Phase III studyof bolus 5-FU/LV (Mayo) vs weekly high-dose 24h 5-FU infusion/LV + Oxaliplatin (OXA) (FUFOX) in advancedcolorectal cancer. Proc Am Soc Clin Oncol 2002; 21: 129.

17. Loumigand C, Cervantes A, Figer A et al. OPTIMOX 1,a randomized study of FOLFOX-4 or FOLFOX-7 withoxaliplatine in a stop & go fashion in advanced CRC. AGERGOR study. J Clin Oncol 2006; 24: 394-400.

18. Cecchin E, Innocenti F, D’Andrea M et al. Predictive role ofthe UGT1A1, UGT1A7, and UGT1A9 genetic variants andtheir haplotypes on the outcome of metastatic colorectalcancer patients treated with fluorouracil, leucovorin, andirinotecan. J Clin Oncol May 20 2009; 27 (15): 2457-65.

19. Innocenti F, Undevia SD, Iyer L et al. Genetic variants inthe UDP-glucuronosyltransferase 1A1 gene predict therisk of severe neutropenia of irinotecan. J Clin Oncol Apr15 2004; 22 (8): 1382-8.

20. Douillard JY, Cunningham D, Roth AD et al. Irinotecancombined with fluorouracil compared with fluorouracilalone as first-line treatment for metastatic colorectal can-cer: a multicenter randomized trial. Lancet 2000; 355:1041-7.

21. Saltz LB, Cox JV, Blanke C et al. Irinotecan plus fluoro-uracil and leucovorin for metastatic colorectal cancer.Irinotecan Study Group. N Engl J Med 2000; 343: 905-14.

22. Koehler BC, Jager D, Shulze BH et al. Targeting cellulardeath signaling in CRC: current strategies and future per-spectives. World of Gastroenterol 2014; 20 (8): 1923-34.

23. Bergers G, Hanahan D. Modes of resistance to antiangio-genic therapy. Nat Rev Cancer 2008; 8: 592-603.

24. Staton CA, Chetwood ASA, Cameron LC et al. Theangiogenic switch occurs at the adenoma stage of theadenoma-carcinoma sequence in colorectal cancer. Gut2007; 56 (10): 1426-32.

25. Bendell JC, Bekaii-Saab TS, Cohn AL et al. Treatmentpatterns and clinical outcomes in patients with metastat-ic colorectal cancer initially treated with FOLFOX-Bevacizumab or FOLFIRI–Bevacizumab: Results fromARIES, a bevacizumab observational cohort study. TheOncologist 2012; 17: 1486-95.

26. Stathopoulos GP, Batziou C, Trafalis D et al. Treatmentof colorectal cancer with and without Bevacizumab: aphase III study. Oncology 2010; 78: 376-81.

27. Giordano G, Febbraro A, Venditti M et al. Targetingangiogenesis and tumor microenvironment in metastaticcolorectal cancer: Role of aflibercept. GastroenterologyResearch and Practice, 2014.

28. Patel A, Sun W. Ziv-aflibercept in metastatic colorectalcancer. Biologics 2014; 8: 13-25.

29. Ciombor KK, Berlin J, Chan E. Aflibercept. Clin CancerRes 2013 April 15; 19 (8): 1920-5.

30. Wilhelm SM, Dumas J, Adnane L et al. Regorafenib (BAY73-4506): a new oral multikinase inhibitor of angiogenic,stromal and oncogenic receptor tyrosine kinases withpotent preclinical antitumor activity. Int J Cancer 2011Jul 1; 129 (1): 245-55.

31. Mross K, Frost A, Steinbild S et al. A phase I dose-esca-lation study of regorafenib (BAY 73-4506), an inhibitorof oncogenic, angiogenic, and stromal kinases, in patientswith advanced solid tumors. Clin Cancer Res 2012 May1; 18 (9): 2658-67.

32. Schultheis B, Folprecht G, Kuhlmann J et al. Regora-fenib in combination with FOLFOX or FOLFIRI as 1rstor 2nd line treatment of CRC: results of a multicenterphase Ib study. Ann Oncol 2013; 24 (6): 1560-7.

33. Sujan R, Ashwani K Singal. Regorafenib: an evidence-based review of its potential in patients with advancedliver disease. Core Evid 2014; 9: 81-7.

34. DeWit M, Boers-Doets CB, Saettini A et al. Preventionand management of adverse events related to regora-fenib. Support Care Cancer 2014; 22 (3): 837-46.

35. Jean GW, Shah SR. Epidermal growth factor receptormonoclonal antibodies for the treatment of metastaticcolorectal cancer. Pharmacotherapy 2008 Jun; 28 (6):742-54.

36. Therkildsen C, Henrichsen-Schnack T, Bergmann TKet al. The predictive value of KRAS, NRAS, BRAF,PIK3CA & PTEN for anti-EGFR treatment in mCRC: asystematic review and meta-analysis. Acta Oncol 2014;53: 852-64.

37. Bendell J, Cannon S, Fakih M. Optimizing treatment withEGFR inhibitors in CRC. http://www.onclive.com/peer-exchange/refractory-crc/Optimizing-Treatment-With-EGFR-Inhibitors-in-CRC.

38. Tejpar S, Lenz HJ, Köhne CH et al. Effect of KRAS andNRAS mutations on treatment outcomes in patients withmetastatic colorectal cancer (mCRC) treated first-linewith cetuximab plus FOLFOX4: New results from theOPUS study. J Clin Oncol 2014; 32 (Suppl 3).

39. Van Custem E, Lenz HJ, Köhne CH et al. Fluorouracil,Leucovorin & Irinotecan + Cetuximab treatment & RASmutations in CRC. J Clin Oncol 2015; 33 (7): 692-700.

40. Hocking CM, Price TJ. Panitumumab in the managementof patients with KRAS wild-type metastatic colorectalcancer. Therap Adv Gastroenterol 2014 Jan; 7 (1): 20-37.

41. Douillard JY, Oliner KS, Siena S et al. Panitumumab-FOLFOX4 treatment and RAS mutations in colorectalcancer. N Engl J Med 2013; 369: 1023-34.

42. Patterson S, Peeters M, Siena S et al. Comprehensiveanalysis of KRAS and NRAS mutations as predictivebiomarkers for single agent panitumumab response in a randomized, phase III metastatic colorectal cancer studyJ Clin Oncol 2013; 31 (Suppl.): Abstract 36/7.

43. Seymour MT, Brown SR, Middleton G et al. Panitumu-mab and irinotecan versus irinotecan alone for patientswith KRAS wild-type, fluorouracil-resistant advancedcolorectal cancer (PICCOLO): a prospectively stratifiedrandomised trial. Lancet Oncol 2013 Jul; 14 (8): 749-59.

44. Van Cutsem E, Peeters M, Siena S et al. Open-label phaseIII trial of panitumumab plus best supportive care com-pared with best supportive care alone in patients withchemotherapy-refractory metastatic colorectal cancer. JClin Oncol 2007 May 1; 25 (13): 1658-64.

45. HeinemannV, von WekersthalL, Decker T et al. FOLFIRIplus cetuximab versus FOLFIRI plus bevacizumab asfirst-line treatment for patients with metastatic colorectalcancer (FIRE-3): a randomised, open-label, phase 3 trial.The Lancet Oncology 2014 Sep; 15 (10): 1065-75.

46. Hocking CM, Price TJ. Panitumumab in the managementof patients with KRAS wild type mCRC. Therap AdvGastroenterol 2014; 7 (1): 20-37.


INTRODUCTION

Intracranial hemorrhage due to arteriovenous malforma-tion or intracranial aneurysms is a rare but severe com-plication of pregnancy with maternal and fetal mortalityof 20% and 33% respectively [1]. Cerebral aneurysmshave a variable clinical presentation during pregnancy[2]. Unruptured ones can be asymptomatic or presentas hyperemesis gravidarium [3-4]. Whereas rupture cancause headaches, consciousness alterations and some-times the difference between eclampsia and cerebral hem-

orrhage due to aneurismal rupture can prove to be diffi-cult [5-7]. Aneurismal subarachnoid hemorrhage (SAH)during pregnancy occurs more often in primiparae andin the third trimester of pregnancy [6,8-10]. The confir-mation of the diagnosis is made by computed tomogra-phy (CT) or lumbar puncture and cerebral angiography.Aneurismal management should be performed in anemergency but fetal prognosis should be considered. Areview of the literature yielded on Pubmed displayedthree case reports [11-13] or series of fifteen cases or less[1,5,8,14-16]. Whether to deliver the patient first, or treatthe aneurysm first, is still controversial, but an emer-gency cesarean section followed by aneurismal treat-ment appears to be a widely accepted strategy in preg-nant women with cerebral aneurysmal complications[5,8,16].

We report here the case of a female patient with rup-tured cerebral aneurysm at 36 weeks of gestation whichpresented to our department.


CCAASS CCLLIINNIIQQUUEE// CCAASSEE RREEPPOORRTTCEREBRAL ANEURISMAL RUPTURE WITH SUBARACHNOID HEMORRHAGE DURING PREGNANCY: A CASE REPORThttp://www.lebanesemedicaljournal.org/articles/63-4/case1.pdf

David ATALLAH1, Fersan MANSOUR1, Elie SAMAHA2, Nadine EL KASSIS1, Joseph NASSIF3

Atallah D, Mansour F, Samaha E, El Kassis N, Nassif J.Cerebral aneurismal rupture with subarachnoid hemorrhageduring pregnancy : a case report. J Med Liban 2015 ; 63 (4) :228-231.

Atallah D, Mansour F, Samaha E, El Kassis N, Nassif J.Rupture d’anévrisme cérébral avec hémorragie méningée pen-dant la grossesse : étude d’un cas. J Med Liban 2015 ; 63 (4) :228-231.

1 Department of Obstetrics and Gynecology, Hôtel-Dieu deFrance University Hospital, St Joseph University (HDF-USJ),Beirut, Lebanon.

2 Department of Neurosurgery, HDF-USJ.3 Department of Obstetrics and Gynecology, American Uni-versity of Beirut Medical Center, Beirut, Lebanon.

Correspondence : David Atallah, MD.e-mail: [email protected]

RÉSUMÉ • Mise au point : L’hémorragie intracrânienne dueà une malformation artério-veineuse ou à un anévrismeintracrânien est une complication rare et grave de la gros-sesse. Les mortalités maternelle et fœtale sont de l’ordre de20% et 33% respectivement. Accoucher la patiente en pre-mier lieu ou bien traiter l’anévrisme est un sujet de débat.Pourtant, une césarienne d'urgence suivie d'un traitementd'anévrisme semble être une stratégie conventionnelle pourles femmes enceintes présentant des complications d’ané-vrisme cérébral. Étude de cas: Une femme, G3P2A0, âgéede 38 ans, se présente à 36 semaines d’aménorrhée à notreservice avec une hémorragie sous-arachnoïdienne bilaté-rale accompagnée de saignement au niveau du quatrièmeventricule et d’hydrocéphalie. Elle avait un anévrisme cé-rébral de l’artère communicante postérieure gauche surl’artériographie. Une césarienne a été réalisée le premierjour de son admission. Ceci a été immédiatement suivi parune dérivation ventriculaire externe avec clippage de l’ané-vrisme au niveau de l’artère communicante postérieuregauche. La mère et son nouveau-né sont sortis de l’hôpitalen bon état de santé, sauf pour l’aphasie de Broca chez lamère. Conclusion : En l’absence de recommandations,nous insistons sur le rôle des soins combinés des neuro-chirurgiens et obstétriciens, au cas par cas, en fonction del’âge gestationnel, l’état neurologique de la mère et de l’ex-périence des professionnels de santé.

ABSTRACT • Background: Intracranial hemorrhage dueto arteriovenous malformation or intracranial aneurysm is arare but severe complication of pregnancy with maternaland fetal mortality of 20% and 33% respectively. Whether todeliver the patient first, or to treat the aneurysm first is stillcontroversial, but an emergency cesarean section followedby aneurismal treatment appears to be a widely acceptedstrategy in pregnant women with cerebral aneurysmal com-plications. Case : A 38-year-old patient, G3P2A0, presentedat 36 gestational weeks with a diffuse bilateral subarachnoidhemorrhage with fourth ventricle bleeding and hydrocepha-lus. She had a cerebral aneurysm of the left posterior com-municating artery on arteriography. A cesarean section wasperformed on the first day of admission, and an externalventricular derivation with clipping of the aneurysm on theleft posterior communicating artery were done immediatelyafter the cesarean section. Mother and newborn were dis-charged from hospital in a good health status except Broca’saphasia in the mother. Conclusion: In the absence of cate-gorical recommendations, we stress the role of combinedcare by both neurosurgeons and obstetricians, on a case tocase basis according to gestational age, mother neurologi-cal status and experience of caregivers.

Keywords : cerebral aneurysm, pregnancy, subarachnoidhemorrhage

D. ATALLAH et al. – Cerebral aneurysm rupture in pregnancy Lebanese Medical Journal 2015 • Volume 63 (4) 229

CASE PRESENTATION

We report a case of a patient, 38 years old, G3P2A0 (twonormal deliveries), pregnant at 36 gestational weekswithout any pregnancy complication. She has no signifi-cant medical or surgical histories.

She presented with acute onset headache and vomiting.An alteration of her consciousness developed few hourslater (Glascow score = 8). She was intubated and sedated.Her laboratory tests showed hematocrit of 34.0%, hemo-globin 11.2 g/dL, serum sodium level of 141.0 mmol/L,potassium 3.5 mmol/L and a blood type of A+. A cerebralCT-scan showed a diffuse bilateral SAH with fourth ven-tricle bleeding and hydrocephalus. An arteriography, withprotection of the abdomen, showed a cerebral aneurysmof the left posterior communicating artery with arterio-graphic signs of hydrocephalus (Figure 1).

A cesarean section was performed on the first day ofadmission (clear amniotic fluid), and an external ventric-ular derivation with clipping of the aneurysm on the leftposterior communicating artery were done immediatelyafter the cesarean section.

The male newborn had an Apgar score of 2 over 10 atone minute of life. He was intubated and transferred to theneonatalogy department for neonatal respiratory distress.He weighed 2475 g, had a cranial perimeter of 34 cmand 48 cm of length. He was treated with antibiotics(cefotaxime and ampicilline) for ten days. A chest X-rayshowed a bilateral bronchogram. He was extubated on thethird day of life. Thereafter, a chest X-ray was done on theninth day of life showing a superior left lobe atelectasia.As a result, the neonate left the department on the twelfthday of life in very good health status.

When it comes to the mother, she was extubated thenext day of the interventions with normal neurologic sta-tus except for a Broca’s aphasia which ameliorated dur-ing her hospital stay. On the fourth postoperative day, acerebral CT-scan showed a shrinking of 30% of the ven-

tricles’ size with regression of the SAH. Also was noteda left fronto-parietal hygroma measuring 5 mm. Theexternal ventricular derivation was ablated. This CT-scanning was repeated on the fourteenth postoperativeday with a complete regression of the SAH. A majora-tion of 50% of the hygroma’s size was noted withoutmass effect on the median line. A hypodense left fronto-parietal lesion with disappearing of the interface be-tween the white and gray matter was also noted, suggest-ing ischemia. Normal ventricular size was found. Thena transcranial Doppler was done showing a severevasospasm. Prior to her discharge, the patient had a cere-bral magnetic resonance imaging (MRI) showing a re-cent left superficial sylvian ischemia with a little chron-ic subdural hematoma in the left fronto-temporo-parietalregion. The ventricles were slightly enlarged.

DISCUSSION

Since the first case published by Feldman et al. in 1955[17], the treatment of cerebral aneurysms in pregnancy isstill debated. This problem is critical because over 50%of ruptured arterial aneurysms in women under the ageof 40 are pregnancy-related [18].

While the obstetrical management of women knownto have a cerebral vascular malformation is controversial[14], cerebral aneurysms are treated whether they rup-ture or not. The risk of bleeding from cerebral arteriove-nous malformations is not significantly increased duringpregnancy. However, the risk of hemorrhage is slightlyincreased at the end of pregnancy, but unchanged duringlabor and delivery, in women with arterial aneurysms[14]. This difference can be due to the hemodynamic andendocrine changes in the growth and rupture of aneu-rysms [6,19]. In fact, on one hand, cardiac output in-creases by 60% at the end of the second trimester asblood volume and blood pressure reached their maxi-mum values at term [20]. On the other hand, pregnancyhormones (estrogen, progesterone and human chorionicgonadotropin) predispose to intracranial aneurysms for-mation, enlargement and rupture [21].

Lynch JC et al. report that surgical management of theaneurysms is associated with lower maternal and fetalmortality than the conservative endovascular treatment;they treated some aneurysms before delivery [1].

According to Mosiewicz et al., ruptured intracranialaneurysms should be treated similarly in pregnant andnon-pregnant patients [6-7,19]. When the aneurysm issuccessfully clipped, the pregnancy can be allowed toprogress to term. In these cases, vaginal delivery is pre-ferred by most clinicians. Moreover, unruptured aneu-rysms should be treated if they are symptomatic orenlarging. Other aneurysms should be treated on an indi-vidual basis [19]. Cesarean section would be indicated inseveral circumstances: 1) when the clinical state of themother is severe (coma, brain stem damage); 2) if theaneurysm is diagnosed during labor, or 3) if the periodbetween the neurosurgical treatment of aneurysm and

FIGURE 1. Arteriography showing cerebral aneurysmof the left posterior communicating artery.

230 Lebanese Medical Journal 2015 • Volume 63 (4) D. ATALLAH et al. – Cerebral aneurysm rupture in pregnancy

labor is less than eight days [6]. In all other cases, a vag-inal delivery is preferable under epidural anestheticwhich should be given if medical induction is carriedout, and where instrumental delivery is being carried outsystematically, unless radical treatment is being per-formed. Our patient was admitted unconscious in our de-partment, which justifies the C-section followed by theclipping of the aneurysm.

Recent trends are the endovascular treatment with suc-cessful maternal and fetal outcome [11-12,14]. Shahabi etal. performed this treatment after emergency cesarean de-livery in a patient at 38 weeks of pregnancy [12]. Itshould also be noted that recurrence of an aneurysm fol-lowing coiling can occur in 25% to 30% of cases with anaverage time of 12 months [22]. Moreover, surgical clip-ping was shown to be superior to endovascular coiling interms of total occlusion rate, in an international subarach-noid aneurysm trial [23-24].

Neurosurgical considerations generally take prece-dence over obstetric ones to avoid the two great risks thatfollow, namely recurrence of hemorrhage and ischemia.

Van Buul BJ et al. reported the effect of general anes-thesia in surgical repair for intracranial aneurysm inpregnancy on fetal heart rate (FHR). There was a com-plete disappearance of FHR variability without decelera-tions or bradycardia. At 41-gestational week, a healthygirl of 4015 g was born [13].

Subarachnoid hemorrhage in pregnancy can be easilymissed due to the rarity of the condition and similarity ofsymptoms with eclampsia. In fact, the most common dif-ferential diagnoses of intracranial hemorrhage duringpregnancy include aneurysm, arteriovenous malforma-tion and preeclampsia [25].

First, patients with aneurysmal subarachnoid hemor-rhage present with a sudden onset of severe headache,with or without alteration of consciousness, and menin-geal irritation. Nausea and vomiting are more frequentafter the rupture. Hematoma or vasospasm are also re-sponsible for focal neurologic deficits [25]. As for pre-eclamptic patients, they present with diastolic hyperten-sion, proteinuria, pathological edema and acceleratedweight gain [25]. Preeclampsia becomes eclampsia whena grand mal seizure occurs with or without thrombocy-topenia, signs of hemolysis or disseminated intravascu-lar coagulation. Seizure may be followed by a temporarycoma. The latter can be prolonged in presence of impor-tant cerebral edema or intracranial hemorrhage [25]. Assuch, a high index of suspicion of subarachnoid hemor-rhage is important for early diagnosis and prevention ofmorbidity and mortality.

Lastly, in the absence of categorical recommenda-tions, we stress the role of combined care by neurosur-geons, obstetricians, anesthesiologists and interventionalradiologists, on a case to case basis according to the ges-tational age, the mother’s neurological status and the ex-perience of caregivers.

The authors indicate that written informed consent was

obtained from the patient for publication of her case reportand the accompanying figure. This paper has not been pre-sented in a national medical society and no financial supportwas attributed to its writing.

REFERENCES

1. Lynch JC, Andrade R, Pereira C. Intracranial hemorrhageduring pregnancy and puerperium: experience with fif-teen cases. Arq Neuropsiquiatr 2002; 60 (2-A): 264-8.

2. Gupta SN, Kechli AM, Kanamalla US. Intracranial hem-orrhage in term newborns: management and outcomes.Pediatr Neurol 2009; 40 (1): 1-12.

3. Georgantopoulou C, Jha R. Intracranial aneurysm inpregnancy presenting as hyperemesis gravidarum. JObstet Gynaecol 2003; 23 (1): 74-5.

4. You SH, Kong DS, Kim JS et al. Characteristic featuresof unruptured intracranial aneurysms: predictive risk fac-tors for aneurysm rupture. J Neurol Neurosurg Psychiatry2010; 81 (5): 479-84.

5. Roman H, Descargues G, Lopes M et al. Subarachnoidhemorrhage due to cerebral aneurysmal rupture during preg-nancy. Acta Obstet Gynecol Scand 2004; 83 (4): 330-4.

6. Mosiewicz A, Jakiel G, Janusz W, Markiewicz P. Treat-ment of intracranial aneurysms during pregnancy. GinekolPol 2001; 72 (2): 86-92.

7. Maurizio G, Roberto A, Valeria P et al. Aneurysmal sub-arachnoid haemorrhage in pregnancy: a case series.Transl Med UniSa 2012; 2: 59-63.

8. Kriplani A, Relan S, Misra NK, Mehta VS, Takkar D.Ruptured intracranial aneurysm complicating pregnancy.Int J Gynaecol Obstet 1995; 48 (2): 201-6.

9. Kataoka H, Miyoshi T, Neki R, Yoshimatsu J, Ishibashi-Ueda H, Iihara K. Subarachnoid hemorrhage from intra-cranial aneurysms during pregnancy and the puerperium.Neurol Med Chir (Tokyo) 2013; 53 (8): 549-54.

10. Tarnaris A, Haliasos N, Watkins LD. Endovascular treat-ment of ruptured intracranial aneurysms during pregnan-cy: is this the best way forward? Case report and reviewof the literature. Clin Neurol Neurosurg 2012; 114 (6):703-6.

11. Piotin M, de Souza Filho CB, Kothimbakam R, Moret J.Endovascular treatment of acutely ruptured intracranialaneurysms in pregnancy. Am J Obstet Gynecol 2001; 185(5): 1261-2.

12. Shahabi S, Tecco L, Jani J et al. Management of a rup-tured basilar artery aneurysm during pregnancy. ActaChir Belg 2001; 101 (4): 193-5.

13. van Buul BJ, Nijhuis JG, Slappendel R, Lerou JG,Bakker-Niezen SH. General anesthesia for surgicalrepair of intracranial aneurysm in pregnancy: effects onfetal heart rate. Am J Perinatol 1993; 10 (2): 183-6.

14. Kizilkilic O, Albayram S, Adaletli I et al. Endovasculartreatment of ruptured intracranial aneurysms duringpregnancy: report of three cases. Arch Gynecol Obstet2003; 268 (4): 325-8.

15. Sayegh I, Clement HJ, Gaucherand P, Rudigoz RC.Cerebral vascular malformations and pregnancy: obstet-rical and anesthetic management. J Gynecol Obstet BiolReprod (Paris) 2002; 31 (4): 379-86.

16. Jaeger K, Ruschulte H, Muhlhaus K, Tatagiba M. Com-bined emergency Caesarean section and intracerebralaneurysm clipping. Anaesthesia 2000; 55 (11): 1138-40.

D. ATALLAH et al. – Cerebral aneurysm rupture in pregnancy Lebanese Medical Journal 2015 • Volume 63 (4) 231

17. Feldman RL, Gross SW, Wimpfheimer S. Rupturedintracranial aneurysm during pregnancy: diagnosis andtreatment. Am J Obstet Gynecol 1955; 70 (2): 289-95.

18. Barrett JM, Van Hooydonk JE, Boehm FH. Pregnancy-related rupture of arterial aneurysms. Obstet GynecolSurv. 1982; 37 (9): 557-66.

19. Stoodley MA, Macdonald RL, Weir BK. Pregnancy andintracranial aneurysms. Neurosurg Clin N Am 1998; 9(3): 549-56.

20. Pedersen H, Finster M. Anesthetic risk in the pregnancysurgical patient 1979; 51: 439-51.

21. Nelson LA. Ruptured cerebral aneurysm in the pregnantpatient. Int Anesthesiol Clin 2005; 43: 81-97.

22. Plowman RS, Clarke A, Clarke M, Byrne JV. Sixteen-year single-surgeon experience with coil embolizationfor ruptured intracranial aneurysms: recurrence rates andincidence of late rebleeding. Clinical article. J Neurosurg2011; 114 (3): 863-74.

23. Molyneux AJ, Kerr RS, Yu LM et al. International Sub-arachnoid Aneurysm Trial (ISAT) Collaborative Group.International subarachnoid aneurysm trial (ISAT) of neu-rosurgical clipping versus endovascular coiling in 2143patients with ruptured intracranial aneurysms: a random-ised comparison of effects on survival, dependency,seizures, rebleeding, subgroups, and aneurysm occlu-sion. Lancet 2005; 366 (9488): 809-17.

24. Molyneux AJ, Kerr RS, Birks J et al. ISAT Collabo-rators. Risk of recurrent subarachnoid haemorrhage,death, or dependence and standardised mortality ratiosafter clipping or coiling of an intracranial aneurysm inthe International Subarachnoid Aneurysm Trial (ISAT):long-term follow-up. Lancet Neurol 2009; 8 (5): 427-33.

25. D’Haese J, Christiaens F, D’Haens J, Camu F. Combinedcesarean section and clipping of a ruptured cerebralaneurysm: a case report. J Neurosurg Anesthesiol 1997;9 (4): 341-5.

IINN MMEEMMOORRIIAAMMhttp://www.lebanesemedicaljournal.org/articles/63-4/inmemoriam1.pdf

Nous ne trouvons pas les mots pour décrire notre douleur quand nous avons appris le décèsde Joseph Moubarak le 20 septembre 2015 à l’âge de 84 ans après une cruelle lutte contreplusieurs tares métaboliques, à laquelle il a fait face avec courage et sérénité.

Avec lui, c’est un des grands noms de la médecine générale libanaise qui disparaît couronnantune carrière médicale tout à fait exceptionnelle.

Né à Rechmaya le 28 mai 1931, Libanais de pure souche, Joseph franchit brillamment toutesles étapes de la carrière médicale universitaire à la faculté de médecine de l’Université Saint-Joseph où il acquit son titre de docteur en médecine en 1957. II approfondit ses connaissancesentre Bordeaux et Paris pour s’installer à Beyrouth en 1960, entre son cabinet privé à Achafriehet l’Hôpital Libanais, centre hospitalier universitaire à Geitaoui. Il s’est marié en 1972.

II exerçait sa profession en pleine conscience, spécialement dans sa région, Aley. Il yexcellait par ses qualités remarquables de clinicien et d’homme de science. II est resté tout aulong de sa carrière attaché à une certaine conception de la médecine, tout en étant très large-ment ouvert aux progrès auxquels il a contribué.

II a toujours fait l’unanimité autour de lui. Nous pleurons tous cet homme merveilleux auquelj’étais lié par une amitié profonde depuis 1980 (étant interne à Geitaoui), forgée par de nombreusesrencontres régulières, entourés d’amis communs.

La disparition de Joseph Moubarak est une grande perte. Que son épouse Jeanine et ses enfants,Jocelyne et Jean-Paul, ainsi que toute sa famille, reçoivent nos sentiments de sympathie attristée.

Dr Georges SALIBAPrésident du Comité de formation

médicale continue de l’OML

Joseph Chucri Moubarak1931 - 2015

VOLUME 63 No 4 Octobre-DécembreOctober-December2015

JM

LJO

UR

NA

LM

ED

ICA

LL

IBA

NA

IS

Vo

l. 6

3 (

4)

- p

p 1

71

-23

2O

cto

bre

-Dé

ce

mb

re

20

15

Octo

be

r-D

ece

mb

er

LE

BA

NE

SE

ME

DIC

AL

JO

UR

NA

LL

MJ

ISSN 0023-9852 Indexed by • Index Medicus • Embase (Excerpta Medica)

• Cab Abstracts and Global Health - CABI Publishing

VOLUME 63 No October-December2015oml.org.lb/Temp/Files/10ac7191-cd0e-465b-8a4d-cb4fae8118cd.pdf ·...

Documents

Transcript of VOLUME 63 No October-December2015oml.org.lb/Temp/Files/10ac7191-cd0e-465b-8a4d-cb4fae8118cd.pdf ·...