INR reporting of prothrombin time.

Here's an inside look at one lab system's switch to INR from PT and its effect on physician anticoagulation dosing.

THE INTERNATIONAL NORMALIZED RATIO (INR) is replacing the use of prothrombin time (PT) to express the level of anticoagulation in patients taking oral anticoagulants.|1^ The INR, unlike PT, takes into consideration the sensitivity of the thromboplastin reagent and has accepted therapeutic limits for different anticoagulation indications.|2^

The INR is calculated from the following formula:

INR = |(Patient PT/Normal PT).sup.ISI^

Normal PT is obtained by averaging the PTs of a group of healthy individuals not taking medications.

The International Sensitivity Index (ISI) is provided by the manufacturer of the thromboplastin and is lot specific and somewhat instrument dependent. The ISI of most thromboplastins used in the U.S. ranges from 2.0 to 3.0.|3^ The use of thromboplastins with ISIs exceeding 2.6 (low sensitivity thromboplastins) is discouraged, as they will yield relatively short PTs and inaccuracy in the laboratory diagnosis of the adequacy of the anticoagulated state. While thromboplastins with low ISIs (less than 1.5) will provide optimal definition of the anticoagulated state, they yield markedly prolonged PTs for anticoagulated patients. Furthermore, the usefulness of low ISI thromboplastins has not yet been demonstrated in patients with liver disease, congenital factor deficiencies, and disseminated intravascular coagulation.

On Dec. 15, 1992, we began reporting INR in our clinic system, a 370-physician, 17-site multispecialty group practice. While this transition required significant laboratory planning and coordination, it yielded a more effective anticoagulation monitoring system. This article describes the transition and its effects on patient INR and physician anti-coagulation dosing.

Table 1
Primary indications for anticoagulation

                     FP      IM1     IM2

Dysrhythmia          11       34       1
CAD                   3       13       0
DVT                   6       21       5
Stroke                5        9       5
Thrombophlebitis      2       10       0
TIA                   2        4       0
Prosthetic valve     10       11       7
Cancer                0       11       0
Other                 0        1       0
Total                39      114      18
Table 2
Average weekly coumarin doses, INRs, and intraindividual standard deviations
before and after INR reporting (includes valve replacements)

                   Dose (mg/w)       Av. INR       SD INR

FP, before            33.5            2.99          0.92
FP, after             29.4(*)         2.61(*)       0.86
IM1, before           30.1            2.90          0.91
IM1, after            28.2(*)         2.47(*)       0.70(*)
IM2, before           27.6            2.96          0.66
IM2, after            25.5            2.56          0.75
All, before           30.7            2.94          0.87
All, after            28.1(*)         2.52(*)       0.75(

* = p |is less than^ 0.01
@ = p |is less than^ 0.05

* Implementation of INR. Four of our 17 laboratories measure PTs, including our central laboratory, which does all of the special hematology, chemistry, and microbiology testing. The central laboratory measures PTs with an MLA 700, two of the sites use an MLA 750, and one other uses a fibrometer. The hematology section supervisor and a technical specialist supervise the central hematology laboratory and provide the technical support to all 16 off-site laboratories. A "team leader" supervises each site.

It took about 9 months to introduce INR reporting because of our lab's complexity and clinic structure and delays in receiving the first lot of more sensitive thromboplastin. Figure 1 shows our time line in converting to the INR system.

* Clinical effectiveness. We studied the changes in patient INR and dosing after INR reporting. Patients were studied if they had at least two PT values measured at our central lab between Aug. 15, 1992, and April 15, 1993. The equation mentioned earlier was used to calculate INR values for patient PTs reported before Dec. 15, 1992. A total of 1,122 patients (8,151 INR values) were studied, 510 females (average age, 63.7 years) and 612 mates (average age, 63.3 years).

Exponential smoothing was used to demonstrate the trends in the interindividual INR average and standard deviations. Figure 2 shows the exponentially smoothed interindividual INR averages and standard deviations over the period of the study.(4) Also shown are all of the individual INR values. The average INR drops from about 3.0 to 2.6, and the smoothed standard deviation drops from 1.4 to under 1.2.

Three groups of patients were studied to determine whether changes in INR were accompanied by changes in anticoagulant dosing: internal medicine at the central clinic (IM1), family practice at the central clinic (FP), and internal medicine at one of our site clinics (IM2). Table 1 summarizes the indications for anticoagulation for the three patient groups. Table 2 shows the weekly coumarin dose, the INR averages, and intraindividual standard deviations for INR values collected at least 1 month after the initiation of INR reporting. There are significant decreases in the weekly coumarin dose as well as intraindividual INR average and standard deviation.

Figure 3 shows the cumulative frequency distributions of the three different patient groups. Patients with heart valve replacements are separated, since their current INR guidelines are higher than those for any other indication. Cumulative frequency distribution curves allow the visual estimation of conformance to INR standards. To illustrate, if the range of acceptable INRs is between 1.5 and 3.0 for patients without heart valve replacements, then for the IM2 population, approximately 70% are within the therapeutic limits, with 17% of the INR observations under 1.5 and 12% above 3.0. For the IM2 population, very few observations exceed INRs of 4.0.

* Reduced doses/variation. Our positive experience with INR reporting has led to decreases in oral anticoagulant doses and reduced variation in inter- and intraindividual variation in anticoagulation. Also, the lab calls in far fewer critical PT values to the clinician. It is gratifying to know that several local hospitals now use a reporting style similar to ours. If the scope of the study were broadened to include clinical outcome and more patients, we believe the decreased coumarin dosage would be associated with a decreased risk of bleeding.

It is difficult to assess the acceptability of the 65% to 70% proportion of non-valve replacement patients within the therapeutic range (INR of 1.5 to 3.0). About one half of our patients are health maintenance organization patients, and we have a continual influx of new patients. The outlying data are partially due to the new patients, some of whom are being anticoagulated differently. While we have presented these results to our clinicians, we believe more rigorous adherence to the therapeutic range will require a formal group approach.

References

1. Brandt J. INRs increasingly used but not problem-free. Summing Up. Northfield, Ill: College of American Pathologists; Summer 1993: 1.

2. Hirsh J, Dalen JE, Deykin D, Poller L. Oral anticoagulants: Mechanisms of action, clinical effectiveness, and optimal therapeutic range. Chest. 1992; 102(suppl): 3125-3265.

3. Meier FA, Schifman RB. The INR and monitoring of oral anticoagulation: Data analysis and critique. College of American Pathologists. Q-Prove 91-10A. 1992.

4. Cembrowski GC, Carey RN. Laboratory Quality Management: QC and QA. Chicago, Ill: ASCP Press; 1989: 26-28.

Cembrowski is laboratory director, Anderson is hematology supervisor, Steeber is a specimen processor, and Burnett is an internal medicine consultant, all at the Park Nicollet Medical Center, Minneapolis, Minn. The authors would like to thank Drs. Mike McNeely and Malcolm Brigden of Island Medical Laboratories, Victoria, B.C., for allowing them to use their protocol for INR conversion, and Dr. Bob Coupland of Temple University Hospital, Philadelphia, Pa., for his constructive comments.

Figure 1

Time line for conversion to INR reporting

March '92

* Held initial meeting of hematology administration to decide INR implementation and select reading materials for participating team leaders.

* Sent letter to heads of oncology/hematology, cardiology, cardiovascular surgery, and internal medicine asking for input on desirable INR ranges.

* Drafted INR computer report.

April

* Met with vendor to learn about the different thromboplastins available. Decided to use thromboplastin with lower ISI of 2.0 (compared to 2.5 used in 1991-92).

* Met with coumarin sales representatives, who provided copies of introductory booklet on INR for medical department education.

* Met with team leaders to discuss computer report and selection of new thromboplastin; distributed reading materials; introduced plan to confirm current PT methodologies and reference ranges and to evaluate new thromboplastin.

May-Sept.

Confirmed that all sites were producing PTs with same means and reference ranges using old thromboplastin; awaited new thromboplastin. Taught laboratory staff about INR.

Sept.-Oct.

Compared results of plasma specimens analyzed with both old and new thromboplastins (50 specimens at central laboratory, 20 each at other sites.)

Determined linear relationship between old and new PT values so we could print PT values that would have been obtained if higher ISI thromboplastin were used.

Initiated clinician education consisting of clinician-led discussions of anticoagulation monitoring.

Oct.-Nov.

Performed normal range study (50 specimens at central lab; 10 each at other sites.)

Grouped and analyzed data to determined 5th and 95th percentile normal ranges.

Nov.

Sent newsletter to clinicians, notifying them of the change and that we would be printing both INR and new PT as well as PT values if old thromboplastin were used (calculated from linear relationship).

Dec. 15

Started to report INRs. Began clinical evaluation.

Spring '93

Developed patient education material.

June

Discontinued printing PT values using the higher ISI thromboplastin.