Accepte

Reprint208, 300

0039-60

� 2010

doi:10.1

The value of various definitionsof intrathoracic goiter for predictingintra-operative and postoperativecomplicationsAntonio Rıos, PhD, Jose M. Rodrıguez, PhD, Marıa D. Balsalobre, MD, Francisco J. Tebar, MD, andPascual Parrilla, PhD, El Palmar, Murcia, Spain

Background. Intrathoracic goiter (IG) is a pathologic and clinical entity defined by criteria that variesfrom one series to the next. The objective of this study was to determine the most useful definition of IG forpredicting intra-operative and postoperative complications.Methods. The study included 201 patients treated for multinodular goiter who met the following criteria:(1) they had no previous thyroid surgery; (2) they had undergone total thyroidectomy; and (3) they werediagnosed with IG according to 1 of the following definitions: (1) clinical; (2) Hsu’s; (3) Kocher’s;(4) Torre’s; (5) Eschapase’s; (6) Lahey’s; (7) Lindskog’s; (8) Crile’s; (9) Katlic’s; and (10) subcarinal.Three variables were evaluated: (1) intra-operative complications; (2) need for a sternotomy; and(3) postoperative complications.Results. During orotracheal intubation, there were difficulties in 25 cases, all of which were detectedusing the 6 least restrictive definitions of IG (these range from the clinical definition to Lahey’sdefinition. In 6 (3%) cases, it was necessary to carry out a sternotomy. The thoracic approach could bepredicted using Katlic’s definition. None of the definitions of IG was useful for predicting postoperativecomplications.Conclusion. Most definitions of IG can be ignored because they are not clinically relevant. The 6least restrictive definitions overlap in their utility to predict intubation difficult during the anestheticprocess; consequently, the clinical definition should be used because it is the simplest to calculate. Katlic’sdefinition is the most useful for predicting a possible sternotomy for extirpating goiter. (Surgery2010;147:233-8.)

From the Servicio de Cirugıa General y del Aparato Digestivo I, Servicio de Endocrinologia y Nutricion,Hospital Universitario Virgen de la Arrixaca, El Palmar, Murcia, Spain

INTRATHORACIC GOITER (IG), described anatomicallyby Haller in 1749, generally refers to multinodulargoiter (MG).1,2 One of the main problems withcomparing studies about IG is that the criteriaused to describe this clinical and pathologic entityis inconsistent. Consequently, the published IG in-cidence rates range between 0.2% and 45% of allgoiters, depending on the definition used.1-14 Hav-ing many definitions for the same clinical entitymakes it difficult to compare series and to drawconclusions that are applicable overall and, there-fore, they are of little practical value.

d for publication June 18, 2009.

requests: Antonio Rıos, PhD, Avenida de la Libertad no07 Casillas, Murcia, Spain. E-mail: arzrios@teleline.es.

60/$ - see front matter

Mosby, Inc. All rights reserved.

016/j.surg.2009.06.018

The clinical utility of most definitions of IG isvery debatable. Until now, evaluation has not takenplace to determine: (1) differences in the defini-tions for predicting anesthetic, intra-operative, andpostoperative complications; (2) the progressionof these complications; and (3) surgical ap-proaches.12-15

The objective of this study is to determine themost useful IG definition in medical practiceaccording to its benefit for predicting the devel-opment of: (1) postoperative complications,(2) intra-operative complications, and (3) theneed for sternotomy in the operative approach.

PATIENTS AND METHODS

Selection criteria. A retrospective review of pa-tients diagnosed and operated on for MG wascarried out in our department (the EndocrineSurgery Unit at the Virgen de La Arrixaca Univer-sity Hospital) between January 1980 and December2001. All the patients who met the following

SURGERY 233

Table I. Distribution of patients with intrathoracicgoiter according to the different definitions

Definition No. of patients (N = 201)

Clinical 201Hsu7 201Kocher16 199Torre17 199Eschapase18 191Lahey19 182Lindskog20 98Crile21 98Katlic3 52Subcarinal22 12

SurgeryFebruary 2010

234 Rıos et al

criteria were included in this study: (1) patientshaving goiter with an intrathoracic componentand who met the criteria of at least 1 of thedefinitions of IG. The definitions were as follows:(1) the clinical definition: a thyroid gland that, onneck examination without being in hyperexten-sion, has a portion that remains permanentlyretrosternal; (2) Hsu’s7 definition: a thyroid gland(clinically or radiologically) below the sternal ma-nubrium; (3) Kocher’s16 definition: a thyroid glandin which some portion remains permanently retro-sternal; (4) Torre’s17 definition: goiter with its lowerposition permanently remaining below the sternalnotch with the neck in hyperextension; (5) Escha-pase’s18 definition: goiter totally or partially locatedin the mediastinum that, in operating position, hasits edge at least 3 cm below the sternal manu-brium; (6) Lahey’s19 definition: goiter which needsexeresis to be performed in the upper mediasti-num; (7) Lindskog’s20 definition: thyroid growthup to the level of the fourth thoracic vertebraeon x-ray examination; (8) Crile’s21 definition: thy-roid growth up to the aortic arc; (9) Katlic’s3 defini-tion: goiter in which at least 50% is retrosternal;and (10) the subcarinal22 definition of goiter: goi-ter with thyroid growth that reaches the carina tra-cheae. The other 2 criteria for study inclusion were(2) patients who had received total thyroidectomysurgical treatment and (3) those without any previ-ous thyroid surgery.

Study sample. Of the 221 patients that met theselection criteria, 20 were excluded from the studybecause of a lack of data that prevented comple-tion of the study protocol; without these data, thepatients could not be classified according to thedefinitions. Table I presents the distribution of pa-tients according to each definition. The clinicaldefinition was clearly the least restrictive (201 pa-tients met this criteria), and the subcarinal defini-tion22 was the most restrictive (only 12 cases metthis criteria). A peculiarity of the definitions wasthat some were included within others. Thus, forexample, an IG according to Lindskog’s defini-tion20 was also defined by the 6 previous defini-tions (clinical, Hsu’s, Kocher’s, etc).

Given that some definitions overlapped andcould be used to classify the same number ofpatients in the study, they were combined into 3main groups: (1) the clinical definition andHsu’s definitions7; (2) Kocher’s16 and Torre’s17

definitions; and (3) Lindskog’s20 and Crile’s21 def-initions. Consequently, in the final analysis, 7 pa-tient groups were studied rather than 10: the 3groups just mentioned plus Eschapase’s,18 La-hey’s,19 Katlic’s,3 and the subcarinal definitions.22

Therapeutic diagnostic protocol. In all patients,thyroid hormones were determined (T4 and TSH)and a simple posterior-anterior chest x-ray wastaken. When a mediastinal mass was detected, alateral projection was carried out. Computed to-mography (CT) was indicated in cases that met atleast 1 of the following criteria: (1) the MG had alarge intrathoracic component (the CT was used toidentify connections with neighboring structures);(2) a mediastinal tumoration was found, and therewere diagnostic doubts about the origin of thethyroid (the CT clarified the diagnosis and identi-fied connections with the neighboring structures);and (3) the patients had serious compressive symp-toms (respiratory insufficiency, dysphagia, etc).

All the patients were considered to be at riskfrom orotracheal intubation, and so they wereassisted by experienced anesthetists. The caliberof the orotracheal tube was 8--8.5 mm for malesand 7--7.5 mm for females, although the anesthe-tist could choose a narrower caliber if deemednecessary depending upon the tracheal caliberestablished preoperatively by the CT technicians.Intubation was considered to be either: (1) normalor easy, when it was achieved on the first attempt us-ing the tube selected by the anesthetist after pa-tient assessment; or (2) difficult, when it requiredthe use of smaller tubes and a guiding wire orthe use of a fiberoptic bronchoscopy. Fiberopticbronchoscopy was directly indicated for patient in-tubation according to the anesthetist’s criteriawhen, by using the various imaging techniques, itwas found that: (1) the trachea had a very narrowcaliber, or (2) there was severe deviation of the tra-chea. The chosen surgical approach was cervical.

Variables analyzed. The variables analyzedwere as follows: (1) epidemiologic variables(age, sex, and residence in an endemic goiterarea); (2) disease-related variables (family ante-cedents of thyroid pathology, evolution of thedisease over time, and clinical features);

Table II. Intra-operative data of intrathoracic multinodular goiter according to the definition used

Clinicaland Hsu(n = 201)

Kocher andTorre

(n = 199)Eschapase(n = 191)

Lahey(n = 182)

Lindskogand Crile(n = 98)

Katlic(n = 52)

Subcarinal(n = 12)

Control(n = 218)

Difficulty with intubation,n (%)

25 (13)* 25 (13)* 24 (13)* 24 (13)* 13 (15)* 6 (12)* 2 (17)* 3 (1)

Utilization of a fiberopticbronchoscopy, n (%)

7 (4)* 7 (4)* 7 (4)* 7 (4)* 3 (3)* 1 (2) 0 0

Sternotomy, n (%) 6 (3)* 6 (3)* 6 (3)* 6 (3)* 4 (4)* 4 (8)* 3 (25)*y 0Operative time, min 121 ± 40* 121 ± 41* 122 ± 40* 122 ± 41* 122 ± 43* 130 ± 51* 132 ± 68* 107 ± 35

*P < .05 compared to the control group.yP < .05 compared to the clinical definition and Hsu’s definition.

SurgeryVolume 147, Number 2

Rıos et al 235

(3) intrathoracic definition of the goiter;(4) surgical data (difficulty with orotracheal intu-bation for the anesthesia, the need to perform asternotomy, the duration of surgery); (5) anato-mopathologic data (associated carcinoma andgoiter weight); and (6) surgical morbidity andmortality (hypoparathyroidism, recurrent nerveinjuries, general complications, local complica-tions, and mortality).

With respect to complications, hypoparathyroid-ism in the immediate postoperative period wasconsidered to occur when calcium levels werebelow 7.5 mg/dL or when they were below 8.5mg/dL and the patient presented with symptomsdue to hypocalcemia.23 Hypoparathyroidism wasconsidered to be definitive when the serum cal-cium level was less than 8.5 mgrs/dl without cal-cium supplements 1 year after surgery. Arecurrent nerve injury was considered if therewas alteration in the tone, timber, or intensity ofthe voice after surgery, with laryngoscopic confir-mation of vocal cord paralysis; the diagnosis wasconsidered definitive if the symptoms persistedfor more than 12 months. Superior laryngeal nervelesion was considered if the patient presented witha normal voice, but after speaking, experienced aloss of tone and timber.

Control group. The control group consisted of218 patients with MG who had undergone surgeryduring the same period as the study group. Thesepatients fulfilled the following criteria: (1) theydid not comply with the any of the IG definitions;(2) they did not have a history of previous thyroidsurgery; and (3) they had undergone a totalthyroidectomy.

Statistical analysis. For the statistical analysis,the v2 test was used, complemented by an analysisof remainders and the Student t test. The Fisherexact test was used if needed. If it was necessaryto apply a nonparametric test, the Wilcoxon testwas applied. The analyses were carried out using

SPSS statistical software (version 11.0; SPSS, Chi-cago, IL). Differences were considered significantat P < .05.

RESULTS

Intra-operative management. Orotracheal intubation:Using a combination of the clinical definition andHsu’s7 definition, there was difficulty in orotra-cheal intubation during anesthesia in 25 (13%)of the 201 patients in the study group comparedto only 4 (2%) of the 218 patients in the controlgroup (P < .0001). Significant differences werefound between the control group and all theIG groups, although no differences were foundaccording to the definition used. As the definitionbecame more restrictive, the difficulty for intubationpersisted between 13% and 17% of patients, as shownin Table II. In addition, approximately 2--4%of patients in the study group required the use of afiberoptic bronchoscopy for intubation comparedto no patients in the control group (P < .05).

The need for a sternotomy: In 6 patients, a thoracicapproach was needed to gain access to the thyroidgland and to be able to extirpate it. All of theseprocedures were carried out using a midsternot-omy. The need for a thoracic approach is reflectedin the same way by all the definitions. If we exclude2 patients from the first years of the series duringwhich a sternotomy was performed due to techni-cal problems, we find that no sternotomies werecarried out on any of the patients with a definitionthat was less restrictive than Katlic’s3 definition.The 2 excluded patients were treated at the begin-ning of the 1980s and were operated on by inexpe-rienced surgeons. After assessing the 2 cases, webelieve that sternotomy was a consequence of aninappropriate operative technique that led tobleeding of the lower thyroid pole, making sternot-omy necessary in both cases. A thoracic approachwas not needed in any of the cases in the controlgroup (Table II).

Table III. Morbidity and mortality in surgery for intrathoracic multinodular goiter according to thedefinition used

Clinicaland Hsu(n = 201)

Kocherand Torre(n = 199)

Eschapase(n = 191)

Lahey(n = 182)

Lindskogand Crile(n = 98)

Katlic(n = 52)

Subcarinal(n = 12)

Control(n = 218)

Global complications, n (%) 62 (31) 61 (31) 59 (31) 56 (31) 35 (36) 16 (31) 5 (42) 68 (31)Transitory hypoparathyroidism, n (%) 33 (16) 32 (16) 31 (16) 30 (16) 17 (17) 7 (14) 2 (17) 38 (17)Transitory recurrent lesion, n (%) 29 (14) 29 (15) 27 (14) 25 (14)* 19 (19) 8 (15) 5 (42)y 47 (11)Superior laryngeal nerve lesion, n (%) 3 (2) 3 (2) 3 (2) 3 (2) 2 (2) 1 (2) 0 5 (2)Permanent hypoparathyroidism,

n (%)1 (0.5) 1 (0.5) 1 (0.5) 1 (0.5) 0 0 0 2 (1)

Permanent recurrent lesion, n (%) 3 (1.5) 3 (1.5) 2 (1) 2 (1.1) 2 (2) 1 (1.9) 0 2 (1)Tracheotomy, n (%) 1 (0.5) 1 (0.5) 1 (0.5) 1 (0.5) 1 (1) 0 0 2 (1)Local complications, n (%) 4 (2) 4 (2) 4 (2) 3 (2) 2 (2) 0 0 3 (1.4)

Hemorrhage, n (%) 2 (1) 2 (1) 2 (1) 1 (1) 1 (1) 0 0 2 (1)Infection, n (%) 2 (1) 2 (1) 2 (1) 2 (1) 1 (1) 0 0 1 (0.5)Seroma, n (%) 0 0 0 0 0 0 0 0

Systemic complications, n (%) 1 (1) 1 (1) 1 (1) 1 (1) 1 (1) 1 (2) 0 2 (1)

*P < .05 compared to the control group.yP < .05 compared to the clinical definition and Hsu’s definition.

SurgeryFebruary 2010

236 Rıos et al

Morbidity and mortality. Overall morbidityranged between 31% and 42%, and there wereno significant differences between definitions or incomparison to the control group (31%)(Table III). There were no differences in the ratesof permanent complications according to the dif-ferent definitions (range, 0.2--1.5%). In addition,there were 2 cervical hemorrhages, 1 of whichrequired an urgent tracheotomy (Table III).

Histologic study. The weight of the extirpatedpiece was greater when the definition was morerestrictive (166 g vs 219 g, using the clinical andsubcarinal definitions, respectively). In 12 (6%) ofthe 201 cases, there was an associated thyroid carci-noma. All cases were well-differentiated papillarycarcinomas except 1 that was follicular. Of these,11 were microcarcinomas. Of the papillary cases,1 had capsular invasion and 0 had vascular orlymph node capsular invasion, although 4 weremultifocal. In contrast, the patient with follicularcarcinoma presented with vascular and capsular in-vasion and lymph node metastasis. Currently, allpatients are disease-free. The presence of associ-ated carcinomas is less frequent in IG comparedto non-IG, as shown in Table IV; however, no differ-ences have been found according to the differentintrathoracic definitions applied.

DISCUSSION

The thyroid gland is limited by rigid structuresexcept at its lower edge, which is therefore a pointof least resistance. In addition, negative intratho-racic pressure, descending traction produced bydeglutition, or the weight of the thyroid (gravity

mechanism) favor its intrathoracic location.15,24

However, although it is usually located below thesternal notch, there is no consensual definitionwhen discussing IG.25,26 There are many defini-tions of IG, some of them are fairly recent,22 sug-gesting that no particular definition is ideal perse. This lack of a definition means that, when dis-cussing IG, each author presents entities that can-not be compared, resulting in differences in theincidence of IG. Thus, if we observe our data, theincidence of intrathoracic MG in our series couldvary from 48% if we use the least restrictive defini-tion to 2.8% if we use the most restrictive one. It isnoteworthy that we have not found any studies thatcompare IG definitions in the revised bibliography.

Based on our results, we believe that we canproperly evaluate any IG by using only 2 defini-tions. The first should be the clinical definitionbecause it is easy to carry out, it does not requirecomplementary examinations, and it is as sensitiveas the definitions of Hsu,7 Kocher,16 Torre,17

Eschapase,18 Lahey,19 Lindskog,20 and Crile21 fordetermining the presence of compressive clinicalfeatures, difficulty with orotracheal intubation,and the development of postoperative complica-tions. This similarity is because a goiter that ex-tends beyond the sternal manubrium can alreadypotentially produce the same complications asany other goiter that penetrates further into thethorax.

Additionally, it should be remembered thatgenerally compressive clinical features occur dueto fitting in and compression at the entry point tothe mediastinum as the goiter is compressed

Table IV. Histologic data of intrathoracic multinodular goiter according to the definition used

Clinicaland Hsu(n = 201)

Kocherand Torre(n = 199)

Eschapase(n = 191)

Lahey(n = 182)

Lindskogand Crile(n = 98)

Katlic(n = 52)

Subcarinal(n = 12)

Control(n = 218)

Weight of theextirpated goiter

166 ± 132* 167 ± 133* 171 ± 134* 173 ± 135* 183 ± 98* 199 ± 90*a 219 ± 127*y 65 ± 43

Associated carcinoma,n (%)

12 (6)* 11 (6)* 11 (6)* 10 (6)* 6 (6) 1 (2)* 1 (8%) 29 (13%)

*P < .05 compared to the control group.yP < .05 compared to the clinical definition and Hsu’s definition.

SurgeryVolume 147, Number 2

Rıos et al 237

between the manubrium and the vertebral col-umn.14,15,18 The quantity of intrathoracic tissuedoes not affect goiter compression. This featurealso explains why the clinical definition is thebest of the various definitions for detecting poten-tial cases where there is going to be difficulty withorotracheal intubation.27 According to the clinicaldefinition, there was difficulty in intubation duringanesthesia in 13% (25/201) of IG cases, making itnecessary to use fiberoptic bronchoscopy in 7(4%) of the 201 cases. Therefore, it was necessaryto have the operating room ready in case an urgentintubation was needed. It should be noted that, al-though indications for fiberoptic bronchoscopyare well accepted, there is variability from oneanesthetist to another in practice. In our series,there was a small group of anesthetists, and sothe criteria for its use were quite homogenous.

The second definition that should be used isKatlic’s definition,3 because it allows us to includenearly all patients who are going to require a ster-notomy. In our series, these patients comprised 8%of the entire study group and accounted for nearly100% of those patients who required a sternotomy.There were 2 cases of a sternotomy carried out onIGs identified using a less restrictive definitionthan that of Katlic.3 However, these 2 cases oc-curred in the first 2 years of the series, and therewere technical problems. Currently, the goiterwas extirpated through the neck without any diffi-culty and without the need for a sternotomy inmany similar cases.

We would like to emphasize that the sternotomyrate in experienced endocrine units is usuallycalculated according to Katlic’s definition,3 withpercentages between 7% and 15%.28 Some serieshave a higher rate, generally because they includethyroid carcinomas and relapsed goiter,29 whichare the 2 most frequent reasons for requiring athoracic approach during thyroid surgery. Clearly,there are many cases in which one surgeon will ‘‘re-quire’’ a sternotomy and a different surgeon will beable to remove the goiter through the cervical

incision. Therefore, it is important for these pa-tients to be operated on by experienced surgeons.

In relation to complications, there are practi-cally no differences with respect to the controlgroup or the use of different definitions. Thisfinding is because total thyroidectomy can becarried out in endocrine surgery units with nearlythe same guarantees as there are for surgery onnon--IG cases.30 It is worth noting, however, thatthere is a high percentage of temporary complica-tions (dysphonia and transitory hypoparathyroid-ism), but this finding is possibly because we haveused a definition that is too broad. If we had ruledout asymptomatic hypocalcemia as hypoparathy-roidism, the number of complications wouldhave been drastically reduced. Even so, we believethat it would have been an accurate reflection ofthe situation because, although these patientscould be discharged early, they required calciumsupplements for a period of time.

In conclusion, we believe that most classifica-tions of IG could be made redundant because theydo not provide any clinical relevance. The 6 leastrestrictive definitions (clinical, Hsu,7 Kocher,16

Torre,17 Eschapase,18 and Lahey19) overlap eachother regarding the prediction of intubation diffi-culty during the aesthetic process, which is why theclinical definition should be adopted because it isthe easiest to use. The most useful definition forpredicting a possible sternotomy for extirpatingthe goiter is Katlic’s3 definition. No other defini-tion predicts a greater risk of developing complica-tions derived from surgery.

REFERENCES

1. Singh B, Lucente FE, Shaha AR. Substernal goiter: a clinicalreview. Am J Otolaryngol 1994;15:409-16.

2. Hurley DL, Gharib H. Evaluation and management of mul-tinodular goiter. Otolaryngol Clin North Am 1996;29:527-40.

3. Katlic MR, Wang C, Grillo HC. Substernal goiter. AnnThorac Surg 1985;39:391-9.

4. Rodrıguez JM, Hernandez Q, Pinero A, Ortiz S, Soria T,Ramırez P, Parrilla P. Substernal goiter: clinical

SurgeryFebruary 2010

238 Rıos et al

experience of 72 cases. Ann Otol Rhinol Laryngol 1999;108:501-4.

5. Katlic MR, Grillo HM, Wang CA. Substernal goiter. Analysisof 80 patients from Massachusetts General Hospital. Am JSurg 1985;149:283-7.

6. Madjar S, Weissberg D. Retrosternal goiter. Chest 1995;108:78-82.

7. Wong CKM, Wheeler MH. Thyroid nodules: rational man-agement. World J Surg 2000;24:934-41.

8. Alvarez A, Cerezo F, Baamonde C, Salvatierra A, Lopez FJ.Tratamiento quirurgico del bocio cervicomediastınico (inSpanish). Arch Bronconeumol 1997;33:84-8.

9. Hall TS, Caslowitz P, Popper C, Smith GW. Substernal goiterversus intrathoracic aberrant thyroid: a critical difference.Ann Thorac Surg 1988;46:684-5.

10. Shahian DM, Rossi RL. Posterior mediastinal goiter. Chest1988;94:599-602.

11. Creswell LL, Wells SA Jr. Mediastinal masses originating inthe neck. Chest Surg Clin North Am 1992;2:23-9.

12. Michel LA, Bradpiece HA. Surgical management of subster-nal goitre. Br J Surg 1988;75:565-9.

13. Allo MD, Thompson NW. Rationale for the operative man-agement of substernal goiters. Surgery 1983;94:969-77.

14. Rıos A, Rodrıguez JM, Galindo PJ, Torres J, Canteras M, Bal-salobre MD, et al. Results of surgical treatment in multinod-ular goiter with an intrathoracic component. Surg Today2008;38:487-94.

15. Rıos A, Rodrıguez JM, Canteras M, Galindo PJ, Tebar FJ,Parrilla P. Surgical management of multinodular goiterwith compression symptoms. Arch Surg 2005;140:49-53.

16. Modlin IM. Surgical triumvirate of Theodor Kocher, HarveyCushing, and William Halsted. World J Surg 1998;22:103-13.

17. Torre G, Borgonovo G, Amato A, Arezzo A, Ansaldo G, DeNegri A, et al. Surgical management of substernal goiter:analysis of 237 patients. Am Surg 1995;61:826-31.

18. Dahan M, Gaillard J, Eschapase H. Surgical treatmentof goiters with intrathoracic development. In: Delarue NC,Eschapase H, editors. Thoracic surgery: Frontiers and

uncommon neoplasms. International trends in generalthoracic surgery. St Louis: Mosby; 1989.

19. Lahey FH. Diagnosis and management of intrathoracicgoiter. JAMA 1920;75:163-6.

20. Lindskog GE, Goldenberg IS. Differential diagnosis, pa-thology and treatment of substernal goiter. JAMA 1957;163:527-9.

21. Crile G. Intrathoracic goiter. Cleve Clin Q 1939;6:313-22.22. Sancho JJ, Kraimps JL, Sanchez-Blanco JM, Larrad A, Rodri-

guez JM, Gil P, et al. Increased mortality and morbidity asso-ciated with thyroidectomy for intrathoracic goiters reachingthe carina tracheae. Arch Surg 2006;141:82-5.

23. Reeve T, Thompson NW. Complications of thyroid surgery:how to avoid them, how to manage them, and observationson their possible effect on the whole patient. World J Surg2000;24:971-5.

24. Newman E, Shaha AR. Substernal goiter. J Surg Oncol 1995;60:207-12.

25. Hedayati N, McHenry CR. The clinical presentation and op-erative management of nodular and diffuse substernal thy-roid disease. Am Surg 2002;68:245-52.

26. Rıos A, Rodrıguez JM, Galindo PJ, Montoya MJ, Canteras M,Parrilla P. Surgical treatment of multinodular goiter inyoung patients. Endocrine 2005;27:245-52.

27. Rıos A, Rodrıguez JM, Galindo PJ, Balsalobre MD, Tebar FJ,Parrilla P. Manejo de la vıa aerea en la cirugıa del bociomultinodular con afeccion traqueal (in Spanish). Cir Esp2004;76:89-93.

28. Flati G, De Giacomo T, Porowska B, Flati D, Gaj F, TalaricoC, et al. Surgical management of substernal goitres. When issternotomy inevitable? Clin Ter 2005;156:191-5.

29. Monchik JM, Materazzi G. The necessity for a thoracicapproach in thyroid surgery. Arch Surg 2000;135:467-72.

30. Rıos A, Rodrıguez JM, Riquelme J, Soria T, Canteras M, Par-rilla P. Prospective study of postoperative complicationsafter total thyroidectomy for multinodular goiters bysurgeons with experience in endocrine surgery. Ann Surg2004;240:18-25.