intl.elsevierhealth.com/journals/thre

Thrombosis Research (2004) 114, 161--165

Regular Article

Patient preferences for capillary vs. venous INRdetermination in an anticoagulation clinic:a randomized controlled trial

Karen Woods, James D. Douketis*, Terri Schnurr, Krystyna Kinnon,Peter Powers, Mark A. Crowther

Department of Medicine, McMaster University and St. Joseph’s Healthcare, Room F-541,50 Charlton Avenue East, Hamilton, ON, Canada L8N 4A6

Received 19 February 2004; received in revised form 27 May 2004; accepted 27 May 2004Available online 14 July 2004

0049-3848/$ - see front matter n 200doi:10.1016/j.thromres.2004.05.013

* Corresponding author. Tel.: +1-90521-6068.

E-mail address: jdouket@mcmaster

KEYWORDSInternational

normalized ratio;Patient preference;Warfarin

ABSTRACT

Background: Patients who are receiving warfarin therapy require frequent bloodtesting to monitor the intensity of anticoagulation. Although previous studies suggestthat capillary blood monitoring of the international normalize ratio (INR) is rapid andreliable, patient preferences for the method of blood drawing have not beeninvestigated. Methods: We performed a randomized controlled trial of patientsattending an anticoagulation clinic in which patients were randomly allocated toundergo capillary or venous INR monitoring. Patient satisfaction with the outpatientvisit, pain associated with blood drawing, and time spent in the clinic were assessedfor each patient. Results: Sixty patients were studied. Using a 10-point visualanalogue scale to quantify patient satisfaction (0---very satisfied; 10---very dissat-isfied), patients expressed a strong preference for capillary INR monitoring overvenous INR monitoring (1.64 vs. 4.45; P< 0.001). Using a 10-point visual analoguescale to quantify pain with blood sampling (0---no pain; 10---very painful), patientswho underwent capillary INR testing had less pain than venous INR testing (0.83 vs.2.23; PV 0.004). Patients spent, on average, 33 fewer minutes in the clinic withcapillary INR testing than venous INR testing ( P< 0.001). Discussion: Our findingssupport the routine use of capillary blood testing, using a portable monitor, for themanagement of patients in outpatient anticoagulation clinics.n 2004 Elsevier Ltd. All rights reserved.

Introduction

4 Elsevier Ltd. All rights reserv

5-521-6178; fax: +1-905-

.ca (J.D. Douketis).

Warfarin is a widely used and effective anticoagu-lant in patients with chronic atrial fibrillation,venous thromboembolism, coronary artery disease,

ed.

K. Woods et al.162

or a prosthetic heart valve [1]. Its effectiveness isdependent upon maintaining the international nor-malized ratio (INR) within a narrow therapeuticrange, as an INR > 4.5 is associated with an in-creased risk of bleeding and an INR < 2.0 is associ-ated with an increased risk of thromboembolism[2--4]. Consequently, frequent INR testing is re-quired, usually every 2--4 weeks [5], to maintainthe INR within a therapeutic range of 2.0--3.5,depending on the indication, and to avoid extremesof over- and under-anticoagulation [6].

The conventional approach to monitoring theINR involves sampling of venous blood and labora-tory analysis to determine the INR [7]. Recentinterest has focused on the use of point-of-carecapillary blood systems that measure the INR with adrop of whole blood obtained by finger prick [8].The ease of use of capillary blood INR monitoring,coupled with evidence that point-of-care devicesprovide reliable measures of anticoagulant effect,has the potential to allow widespread use in clin-ical practice. However, capillary INR monitoring isnot widely used for several reasons, which includea lack of physician familiarity, increased cost pertest compared to laboratory-based methods, andresistance to changing an established method ofINR monitoring.

Conventional, laboratory-based, INR testing inan anticoagulation clinic setting or by a communitylaboratory requires a long turnaround time of 1--2hours that includes obtaining, delivery, and pro-cessing of the venous blood sample. Consequently,INR testing becomes inconvenient for patients andmight result in decreased compliance with sched-uled INR testing, with the potential for prolongedperiods with sub- or supra-therapeutic INR levels.

As the first step in assessing capillary bloodmonitoring in an anticoagulation clinic setting, weperformed a randomized controlled trial to deter-mine patient satisfaction with conventional venousblood INR monitoring system compared to a point-of-care capillary blood INR monitoring system.

Patients and methods

Clinical setting

The Anticoagulation Clinic at St. Joseph’s Health-care manages the anticoagulant control of approx-imately 560 patients per year who are receivingwarfarin therapy. The clinic oversees patients whoare at a high risk for thromboembolism, bleedingcomplications, or who have a history of difficultanticoagulation control. At the start of each visit,

patients have a venous blood sample collected forINR determination, which is transported by a pneu-matic tube system to the hospital core laboratoryfor processing using standardized laboratory tech-niques. The INR value is reported to the clinic witha computerized laboratory information system. Aphysician or a nurse-clinician reviews the INR resultand provides recommendations about warfarin dos-ing and the date of the next INR test. Patients areprovided with counseling about thromboembolicand bleeding risks, and additional patient concernsare addressed.

Patients

Consecutive patients who were receiving warfarintherapy, with a target INR of 2.0--3.5, were eligiblefor this study. Patients were excluded if they hadone or more of the following characteristics: (1)< 18 years of age; (2) previous participation in thisstudy; (3) required blood testing other than the INRtest; or (4) did not provide written informed con-sent. The Research Ethics Board of St. Joseph’sHealthcare approved the study protocol.

Intervention

Patients were randomly allocated to undergo cap-illary or venous INR testing by opening sealed,sequentially numbered, opaque envelopes. Therandomization sequence was generated using arandom number generator and blocked in groupsof 2, 4, or 6. In patients who were allocated toundergo capillary INR testing, a point-of-care de-vice (CoaguChek-S, Roche Diagnostics, Laval, PQ)was used. This system was operated by one of twoexperienced nurse clinicians. For this test, a singledrop of blood was drawn from the finger tip usingsingle use, sterile lancets and commercially avail-able test strips. In patients who were allocated toundergo venous INR testing, blood was drawn fromthe antecubital vein using standard venipuncturetechniques by one of two trained technologists.Subsequently, the blood was processed through thehospital laboratory. Irrespective of the type ofblood monitoring system used, the clinic staffplaced the INR result on the patient’s chart assoon as the results were available. The INR resultwas then reviewed by a clinic physician or nursewho made specific dosing recommendation to thepatient.

Outcomes

The primary study outcomes were comparisons ofpatient satisfaction with their clinic visit and the

Table 1 Baseline characteristics of the studypopulation

Characteristics INR testing allocation

Capillary(n = 30)

Venous(n = 30)

Mean age (years) 64.4 59.5Female (number) 17 15Reason for anticoagulanttherapyChronic atrial fibrillation 7 4Venous thromboembolism 20 22Mechanical heart valve 3 4English as first language 23 20

Table 2 Study outcomes

CapillaryINR testingn = 30(95% CI)

Venous INRtestingn = 30(95% CI)

P-value

Satisfactionwith clinicvisit, ymeanVAS score(0--10)

1.64(0.65--2.63)

4.45(3.48--5.42)

< 0.001

Pain due tobloodsampling,zmean VASscore (0--10)

0.83(0.36--1.30)

2.23(1.45--3.01)

0.004

Time in clinic, 42 75 < 0.001

Patient INR preference 163

amount of discomfort associated with the bloodcollection method. The outcome was quantifiedusing a visual analogue scale (VAS). For the assess-ment of clinic visit satisfaction, the range of theVAS scoring was from 0 (very satisfied) to 10 (verydissatisfied). For the assessment of pain with theblood collection method, the range of the VASscoring system was from 0 (no pain) to 10 (verypainful). The VAS was assessed at the end of theclinic visit. The duration of each clinic visit was alsoascertained.

Outcomes were measured by providing patientswith a questionnaire at the beginning of their clinicvisit. The questionnaire was divided into two sec-tions. The first section documented patient char-acteristics, registration time, time blood sampleswere taken, INR result, and the time of dischargefrom the clinic. The second section contained thevisual analogue scales. Patients were instructed tocomplete their questionnaires during the course oftheir clinic visit and the completed questionnaireswere collected when patients left the clinic. Studystaff did not assist patients in completing anyportion of the questionnaire.

Statistical analysis

Patient satisfaction was measured by the mean VAS,and the scores were compared between the labo-ratory INR and capillary INR testing groups using thet-test. The secondary outcome measures were alsocompared between the two groups using the t-test.

expressed inminutes(range)

(29--55) (69--81)

INR level(range)

2.5(2.1--2.9)

2.4(2.2--2.6)

0.69

CI = confidence interval.y VAS score: 0 = very satisfied; 10 = very dissatisfied.z VAS score: 0 = no pain; 10 = very painful.

Results

Patient characteristics

Between October 28, 2002 and December 9,2002, 103 patients were screened for the study.

Forty-three patients were excluded (13 were non-consenting, 10 required extra blood testing, and20 had previously participated). Sixty patientswere randomized to undergo INR determinationby venous blood or capillary blood testing. Thepatient characteristics are presented in Table 1.The only difference between patient groups was atrend to patients allocated to capillary testingbeing older.

Patient satisfaction with capillary and labo-ratory INR testing

Patients expressed a strong preference for a clinicvisit associated with capillary INR testing. Themean VAS score on a scale of 0--10 in the capillarygroup was 1.64 (95% CI: 0.65--2.64) and 4.45 (95%CI: 3.48--5.42) in the venous group (P < 0.001).Patients also reported less pain with capillary thanvenous INR testing. The mean VAS score on a scaleof 0--10 in the capillary group was 0.83 (95% CI:0.36--1.30) and 2.23 (95% CI: 1.46--3.01) in thevenous group (P = 0.004). INR testing was signifi-cantly faster with capillary testing as the turn-around time was 75 min in the venous INR groupcompared to 42 min in the capillary INR group(P< 0.001). The study outcomes are presented inTable 2.

The study was not designed to monitor theaccuracy of capillary INR testing. However, there

K. Woods et al.164

was no significant difference in the mean INR in thecapillary INR group and the venous INR group (2.5vs. 2.4, P = 0.69). This finding supports findingsfrom previous studies that have documented thereliability of the point-of-care portable coagulom-eter used in this study (CoaguChek) against a ref-erence standard [9--11].

Discussion

Our study found that in patients who attended ahospital-based anticoagulation clinic, INR testingwith a point-of-care capillary blood method wasassociated with greater patient satisfaction withthe clinic visit and less pain than INR testing with aconventional venous blood sampling method. Capil-lary testing also significantly shortened the clinicvisit time, an observation that likely contributed, inpart, to the greater satisfaction with capillary INRtesting.

Our results are likely to be valid. The study wasa prospective, randomized controlled trial andenrolled unselected patients from an anticoagula-tion clinic. To reduce the possibility of bias inallocation, the randomization process occurredimmediately following patients registering for theclinic and consent being obtained. Sequential,unselected patients were approached for consent.A nurse clinician obtained consent from allpatients and was also responsible for allocatingpatients to the method of INR determination; thiswas accomplished by opening opaque numberedenvelopes after consent had been obtained. Weobtained 100% compliance with the study outcomeassessment.

To our knowledge, no other study has addressedpatient preferences in the method of venous sam-pling for the INR determination within an anti-coagulation clinic or other clinical setting. Point-of-care devices to determine the INR would,ideally, be placed in a community laboratory orpharmacy setting to allow easier access for patientswith the potential to optimize compliance. How-ever, such devices may also be useful in a special-ized anticoagulation clinic, which, as with ourclinic, manages more complicated, high-riskpatients. In this setting, the turnaround timewould be greatly reduced with the potential toincrease patient compliance with INR testing andallow a greater volume of patients to be assessed.Furthermore, randomized trials have demonstratedsuperiority of INR control when undertaken in ananticoagulation clinic than in a community-basedsetting [12--14].

There are potential limitations of our study thatshould be addressed. First, the study was under-taken in a specialized anticoagulation clinic set-ting, with the potential that our findings may notbe comparable to other clinical settings. However,the principle that capillary testing is less painfulthan venous testing should apply in any clinicalsetting in which INR monitoring is carried out.Second, four patients allocated to capillary testingwere unable to provide an adequate sample; al-though included in the capillary group for thepurposes of analysis, these patients had satisfac-tion and pain scores that were similar to, or lessfavorable than, those in the venous group. Thus,these patients reduced the beneficial affect foundwith capillary testing and thereby likely underesti-mated the benefit of capillary sampling. Third, oursample size was small and limited to patients froma single center. However, our results were statisti-cally significant and our patients were typical ofthose taking warfarin, suggesting that our resultsare broadly applicable.

Our study has potential clinical implications.First, a probable reason why capillary monitoringhas not been widely adopted in clinical practicerelates to the relative high cost of the test strips.However, the fact that capillary testing is preferredby patients, is less painful, and shortens clinic visitson average by 33 min suggests that capillary testingmay be cost-effective if the value of the patient’stime is included in the analysis. Second, the INR is avery common laboratory test, and use of capillarytesting has the potential to reduce the need forlaboratory services, particularly if INR testing isneeded during evenings and weekends.

In summary, we found that patients prefer cap-illary blood monitoring to venous monitoring of theINR. The fact that capillary monitoring is also lesspainful, is more rapid, and produces similar INRresults suggests that capillary monitoring couldbecome the measurement tool of choice for mostpatients being followed through outpatient antico-agulant clinics.

Acknowledgements

Dr. Douketis holds a research scholarship fromthe Heart and Stroke Foundation of Canada. Dr.Crowther holds a research scholarship from theCanadian Institutes for Health Research. RocheDiagnostics Canada kindly provided the capillaryblood monitors and test strips. The company wasnot otherwise involved in any component of thestudy and did not influence the analysis or themanuscript.

Patient INR preference 165

References

[1] Ansell J, Hirsh J, Dalen J, Bussey H, Anderson D, Poller L,et al. Managing oral anticoagulant therapy. Chest 2001;119:22S--38S.

[2] Hylek EM, Skates SJ, Sheehan MA, et al. An analysis of thelowest effective intensity of prophylactic anticoagulationfor patients with nonrheumatic atrial fibrillation. N Engl JMed 1996;335:540--6.

[3] Cannegieter SC, Rosendaal FR, Wintzen AR, et al. Optimaloral anticoagulant therapy in patients with mechanicalheart valves. N Engl J Med 1995;333:11--7.

[4] Kearon C, Ginsberg J, Kovacs MJ, et al. Comparison of low-intensity warfarin therapy with conventional-intensity war-farin therapy for long-term prevention of recurrent venousthromboembolism. N Engl J Med 2003;349:631--9.

[5] Gage BF, Fihn SD, White RH. Management and dosing ofwarfarin therapy. Am J Med 2000;109:481--8.

[6] Hirsh J, Dalen J, Anderson DR, Poller L, Bussey H, Ansell J,et al. Oral anticoagulants: mechanism of action, clinicaleffectiveness, and optimal therapeutic range. Chest 2001;119(1 Suppl):8S--21S.

[7] Fairweather RB, Ansell J, van den Besselaar AM, Brandt JT,Bussey HI, Poller L, et al. College of American PathologistsConference XXXI on laboratory monitoring of anticoagulanttherapy: laboratory monitoring of oral anticoagulant ther-apy. Arch Pathol Lab Med 1998;122(9):768--81.

[8] Douketis JD. Patient self-monitoring of oral anticoagulanttherapy. Potential benefits and implications for clinicalpractice. Am J Cardiovasc Drugs 2001;1:245--51.

[9] Douketis JD, Lane A, Milne J, Ginsberg J. Accuracy of aportable international normalization ratio monitor in out-patients receiving long-term oral anticoagulant therapy:comparison with a laboratory reference standard using clin-ically relevant criteria for agreement. Thromb Res 1998;92:11--7.

[10] Cromheecke ME, Levi M, Colly LP, et al. Oral anticoagula-tion self-management and management by a specialist an-ticoagulation clinic: a randomised cross-over comparison.Lancet 2000;356:97--102.

[11] Sawicki PT. A structured teaching and self-managementprogram for patients receiving oral anticoagulation. JAMA1999;281:145--50.

[12] Chiquette E, Amato MG, Bussey HI. Comparison of an anti-coagulation clinic with usual medical care: anticoagulationcontrol, patient outcomes, and heath care costs. Arch In-tern Med 1998;158:1641--7.

[13] Wilt VM, Gums JG, Ahmed OI, Moore LM. Outcome analysisof a pharmacist-managed anticoagulation service. Pharma-cotherapy 1995;15:732--9.

[14] Wilson SJ, Wells PS, Kovacs MJ, et al. Comparing the qualityof oral anticoagulant management by anticoagulation clin-ics and by family physicians: a randomized controlled trial.CMAJ 2003;169:293--8.