ANNALS O F CLINICAL AND LABORATORY SCIEN CE, Vol. 23, No. 3 Copyright 1993, Institute for Clinical Science, Inc. Reemergence of the International Normalized Ratio for the Standardization of Prothrombin Time* EDWARD E. MORSE, M.D., SOPHIA PANEK, M.S., PATRICIA PISCIOTTO, M.D., THOMAS KIRALY, M.S., ROBERT BONA, M.D.,t KATHLEEN VOSE, MT(ASCP),t and FREDERICK RICKLES, M.D. The University o f Connecticut Health Center, Department o f Laboratory Medicine, Farmington, CT 06030 and fst. Francis Hospital and Medical Center, Hematology and Laboratory Medicine, Hartford, CT 06105 ABSTRACT A survey of physicians demonstrated that half had knowledge of the International N orm alized Ratio (INR) but none used the value for m onitoring their patients because it was not available from the Coagulation Laboratory. The Laboratory then provided the INR value at a physician s request. A six month review of prothrombin time (PT) results showed that only the physicians from the Cardiology Clinic and the Hematology Clinic employed the INR for monitoring their patients. General Medical and Surgical, Vascular, and O rthopedic Clinics continued to use the PT in seconds. This dichotomy allowed the unique opportunity to compare variability of PT in patients followed by INR and those followed by PT in seconds. Inpatients on daily m onitoring were used as the standard for close control. During a six month period, laboratory reports from all patients having regular PTs and/or INRs recorded were analyzed for mean level of PT maintained, variability betw een individual PTs in any given patient, and instances when the PT changed = 5 seconds (sec) or increased to S20 sec. Physicians intended to keep the PTs between 16 and 19 sec (INR 2.0 to 3.0). Results show ed statistically significantly lower values of PT, less variation in values of PT and a smaller fraction of patients with changes in PT of =5 sec in the group followed by INR. This group was comparable to the inpatient group b u t significantly different from the outpatient group followed by PT in sec. These results suggest that either the INR allows more stable application of coumadin doses, or physicians using the INR educate their patients in better compliance in diet, medications, and other factors affecting coumadin or vitamin K metabolism. Use of the INR does appear to decrease the variability of the PT in patients so followed. The clinical implications are as yet to be determined. * Send reprint requests to: Edward E. Morse, M.D., The University of Connecticut Health Center, Department of Laboratory Medicine, Room C2067A, MC #2225,263 Farmington Avenue, Farmington, CT 06030. 184 0091-7370/93/0500-0184 $00.90 Institute for Clinical Science, Inc.
REEM ERGENCE OF THE INR FOR THE STANDARDIZATION OF PROTHROMBIN TIME* 185 Introduction T he International N orm alized Ratio (INR) is a concept first developed in 1972 and w idely applied in the U nited Kingdom, Europe, and Canada in an attempt to standardize the various preparations of throm boplastins used for determ ining the prothrombin time (PT) in different countries. T he history of this developm ent has been recently review ed.1 In brief, during the 1970s, a cooperative study, organized under the auspices of the International C om m ittee on Standardization in Hematology/International Council on Thrombosis and Hemostasis (IC SH/IC T H) to standardize throm boplastin, developed a m ethod for d eterm ining an International Sensitivity Index (ISI) for each throm boplastin used. The ISI is the activity of any given thromboplastin relative to the World Health Organization (WHO) standard thromboplastin made from hum an brain extract. Results w ere published in 1979,2 and the international community was encouraged to use the INR system. In North America, comm ercial m anufacturers of throm boplastin had developed m ethods for producing rapid and reproducible throm boplastins extracted from rabbit brain, bovine lung, or human placentas, and the pathologists enjoyed the comfort of rapid turnaround time for clinicians and small coefficients of variation (CVs) for proficiency surveys of the College of American Pathologists (CAP). T he latte r system of standardization involved collection of normal and patient plasmas which were more stable in storage and available in larger volumes in contrast to the standardized throm boplastins of the European system.3,6 W hen it became clear that the rapid, b u t less sensitive throm boplastins produced from rabbit brain (ISI of 2.1 to 2.6) w ere more often associated w ith b leed in g e p iso d e s in p a tie n ts re c e iv in g coumarin derivatives as com pared with their European counterparts who were monitored by the more sensitive (ISI 1.1 to 1.6) but variable human brain thromb o p la s tin s,1,5,7 p h y sician s from th e United States reacted by decreasing the doses of coumarin. They aimed for ratios of PT in the range of 1.3 to 1.8 instead of the 2.0 to 2.5 target recom mended when human brain thromboplastins were used. Since no universally accepted hum an brain throm boplastin was readily available in the U nited States and the practical problem was solved, the INR failed to gain a foothold. The Ad Hoc committee of the ICSH/ ICTH continued to improve upon the INR using an increased num ber of standardized normal and coum arin-treated patient plasm as for standardization of throm boplastins.4 Furtherm ore, international travel in creased, not only for patients w ith an expanding variety of throm botic and vascular diseases, but also for sales persons of a burgeoning num ber of manufacturers of coagulation analyzers and reagents. By 1985, the INR was reemerging in Europe and later in America1,7 until in 1991 even the CAP was moved to produce a Q Probe to determine the reasons for the lack of use of the INR in the United States. Stimulated by this Q Probe, the current authors determ ined among physicians treating patients w ith coum arin derivatives in our facility the level of interest in the INR, and our group undertook to compare the results of patients followed by INR with those followed by the traditional methods (seconds or ratio of PT to normal m ean value). Methods The PT was determ ined by automated technique using optical density m easurem ent (Coag-A-Mate).* The throm boplas * Organon Teknika, 100 Akzo Avenue, Durham, NC 27704.
186 MORSE, PANEK, PISCIOTTO, KIRALY, BONA, VOSE, AND RICKLES tin used for this study was a rabbit brain preparation (Simplastin)* with an ISI 2.2. Normal values were determ ined on citrate t anticoagulated plasm a sam ples from 30 m ale and fem ale laboratory workers. The INR was determ ined as PT of the patient divided by mean normal control PT and the result raised to the power of the ISI. A chart was prepared for the INR values at each 0.5 second interval for quick conversion of m easured PT in seconds to INR if the physician ordered it. Follow ing an announcem ent to all clinics about the availability of the INR, results of patients having PTs m easured at regular intervals (daily inpatient), weekly, or m onthly o u tp atien ts, w ere analyzed. Patients who had at least three PTs determ ined after the INR was applied for monitoring were compared with patients who continued to be m onitored by seconds. Physicians were also surveyed for determ ination of any complications. TABLE I Relationship of Prothrombin Time to the International Normalized Ratio at the University of Connecticut Health Center* PT In Seconds INR PT In Seconds INR 9.0 0.6 17.5 2.5 9.5 0.65 18.0 2.7 10.0 0.7 18.5 2.8 10.5 0.8 19.0 3.0 11.0 0.9 19.5 3.2 11.5 1.0 20.0 3.4 12.0 1.1 20.5 3.6 12.5 1.2 21.0 3.8 13.0 1.3 21.5 4.0 13.5 1.4 22.0 4.2 14.0 1.5 22.5 4.4 14.5 1.7 23.0 4.6 15.0 1.8 23.5 4.8 15.5 1.9 24.0 5.0 16.0 2.1 30.0 8.2 16.5 2.2 40.0 15.5 17.0 2.4 Using rabbit brain thromboplastin with an International Sensitivity Index (ISI) of 2.2. PT = prothrombin time. INR = international normalized ratio. Results The questionnaire associated with the CAP Q Probe revealed that of the 10 physicians sampled, half knew of the INR, but none calculated their own values or used the IN R for oral anticoagulant monitoring. Four of ten who answered the questions thought that there was no reluctance on the part of physicians, nor on the part of laboratory staff to im plem ent use of the INR. All of the p h y sician s u se d th e PT in seconds because it was available from the laboratory. All 10 physicians surveyed in d i cated relatively conservative lim its of prolongation of PT used for m onitoring t Becton Dickinson, 383 Hillen Road, Towson, MD 21204. th e ir patients, starting as low as 14 seconds with most using 19 seconds as the highest value and only two using upper values of 20 or 22 seconds. Laboratory staff indicated their opinion that technical difficulty m ight be encountered in calculating the INR until an analytical instrum ent that produced results reported as INR was available. Im m ediate implem entation was accomplished by posting table I. Testing normal subjects revealed a mean value for the PT of 11.5 ± 0.75 seconds (95 percent confidence interval 10 to 13 seconds). L aboratory reports w ere exam ined from all patients who had had at least three PTs ordered and who had values of PT greater than 13.0 seconds. Forty-nine inpatients w ere analyzed as a closely m onitored group. Fifty-nine outpatients
REEM ERGENCE O F THE INR FOR THE STANDARDIZATION OF PROTHROMBIN TIME 187 follow ed by IN R in C ardiology and Hematology Clinics were compared with 59 outpatients followed by seconds in G eneral M edical and Surgical Clinics (table II). There was a small but statistically significant difference in mean prothrombin time, in the variation betw een individual PTs, and in the proportion of PTs showing a variation of ^ 5 sec betw een the groups. In each of these determinations, the patients followed by INR showed values comparable to the inpatient group of patients. In contrast, the group of outpatients followed by the traditional method (PT in sec) showed small but statistically significantly greater values than either the inpatient group or the INR group in these param eters. Both outpatient groups show ed a statistically greater proportion of patients with PT 20 sec than the closely m onitored inpatient group (table II). Discussion Relatively little resistance to the routine use of the INR in monitoring oral anticoagulation was encountered once physicians ordering the test understood the advantages of its use and the technologists in the laboratory understood the ease of conversion of the PT in seconds to INR. D u rin g th e first six m onths of follow-up, it was clear that the two clinics involved in studies of oral anticoagulant (low dose coumadin) in cardiology and throm bosis tre a tm e n t in hem atology were eager to obtain and use INR data. The general M edical and Surgical C linics, and the Vascular, and O rthopedic Clinics continued to use PT in seconds for monitoring. This dichotomy allowed the examination of levels of PT and of fluctuations (As) betw een values of PT in each set of patients. The inpatient group, which was m onitored on a daily basis with PT in seconds, showed the smallest variation in PT and the lowest proportion of PT g 20 sec. In addition to daily monitoring and coumadin adjustments, inpatients are likely to receive a more uniform diet and have b e tte r co m p lian ce w ith m e d ic a tio n intake since these are provided by a schedule by nurses. Therefore, fluctuation in PT, both in frequency and in magnitude, are likely to be minimized. Taking this group as a standard for comparison, the group followed by INR in the two clinics where a small group of physicians was involved in studies of long-term coum adin effects show ed alm ost identical values for m ean PT, APT, and proportion of PT ^ 5 sec to those of the standard inpatient group. The proportion of PT = 20 sec was statis TABLE II Comparison of Inpatients with Two Groups of Outpatients Treated with Coumarin N m s x PT± SD APT± SD PT>20Sec #A > 5ec IP (Sec) 49 (306) 16.1 ±2.9 1.5 ± 1.60 19/306 6.0% 13/257 5.3% OP (Sec) 59 (273) 17.2 ± 3.6a 2.93 ± 3.08a 33/273b 12.1% 31/214= 14.5% OP (INR) 59 (308) 16.2 ± 3.0a 1.86 ± 1.75a 34/308d 11.4% 13/249c 5.1% a p < 0.01 NP t Test bp <0.025 X2 Test c p < 0.005 Y? Test d p NSX2Test p values refer to differences with the group above the symbol.
188 MORSE, PANEK, PISCIOTTO, KIRALY, BONA, VOSE, AND RICKLES tically significantly greater than the inpatient group, probably reflecting w ider variations seen in outpatients because of lo n g er (m onthly) in te rv als b e tw e e n determ ination of PT. On the other hand, the outpatients followed by a wider variety of physicians in general clinics using the traditional PT in seconds show ed small but statistically significantly higher PT, APT, and an increased proportion of PT ^ 5 sec compared to the group followed by INR. This suggests a variety of potential differences in practice in the control of the patients in the G eneral M edical and Surgical Clinics relating to dosage of coumadin, com pliance, frequency of follow -up, dietary habits, and other m edications. Use of the INR scale for monitoring may allow few er changes in dosage of coumadin. The physicians using the INR may better educate their patients to the complicating effects of diet (vitamin K content) and other drugs (antibiotics, alcohol); the effects of coumadin being of prim ary im portance to these patients. The patients in the General Clinics, on the other hand, may be different in educational and ethnic background, have different dietary habits, and their physicians may have other priorities (hypertension, diabetes, infections, renal failure, arthritis, to name a few) which modify the emphasis given to oral anticoagulants for what may be view ed as a secondary complication of the prim ary disease. Finally, the clinical significance of the small, b u t statistically significant reduction in variability shown in the group of outpatients followed by INR needs to be evaluated. The data in this study support some of the enthusiastic predictions that INR may allow better control and safer practice.5 There is a statistically significant difference in levels of PT and in APTs betw een the two groups of outpatients w hich m ight b e correlated w ith clinical complications. In an institution w here all the physicians use relatively conservative doses of coum arin derivatives and therapeutic PTs range from 14 to 20 sec (INR 1.5 to 3.4), there appears to be small but statistically significant lower values in PT and in variability of PT when patients are followed by INR. The values were comparable with the closely monitored group of inpatients who have daily measurem ents of PT and adjustment of dose. F u tu re work should include: a prospective random ized study w here coum adin dosages are actively adjusted according to the PT or INR, standardization of the frequency of obtaining PT in patients being followed, and specific d e lin e a tio n of c lin ic a l sig n ific a n ce (bleeding or recurrent thrombosis). The present findings support the obvious value of INR when the patient is having PTs perform ed at different laboratories and may have PTs m easured with m arkedly d iffe re n t th ro m b o p lastin s. There is a clear need for a mechanism to compare research studies from institution to institution and to provide published data which can be interpreted in context around the world. References 1. H i r s h, J.: Oral anticoagulant drugs. New Engl. J. Med. 324:1865-1875, 1991. 2. ICTH/ICSH: Prothrombin time standardization: Report of the expert panel on oral anticoagulant control. Thromb. Haemost. 42:1073-1114, 1979. 3. Ko e p k e, J. A., G il m e r, P. R., T r i p l e t t, D. A., and O SULLIVAN, M. B.: The prediction of prothrombin time system performance using secondary standards. Am. J. Clin. Pathol. 68:191-194, 1977. 4. L o e l i g e r, E. A.: ICSH/ICTH recommendations for reporting prothrombin time in oral anticoagulant control. Thromb. Haemost. 54: 155-156, 1985. 5. M e i e r, F. A. and S c h i f m a n, R. B.: Q Probe: The INR and monitoring of oral anticoagulation. College of American Pathologists, 1991. 6. M la LE, J. B. and K e n t, B. A.: Performance characteristics of reference thromboplastins. Am. J. Clin. Pathol. 60:453-457, 1973. 7. POLLER, L.: P rogress in standardization in an tic o a g u la n t c o n tr o l. H e m a to l. R ev. 2:225-241, 1987.