Due to the degree of participation (79% response rate) and sampling (26% of large and 9% of small hospitals), we believe that the results of this survey accurately reflect the state of coagulation laboratory practices as viewed by the respondents in 2001. This response rate may be compared to those for written surveys obtained for other more specific (and shorter) surveys: 51% for a survey of activated partial thromboplastin (aPTT) reporting in Canadian medical laboratories,¹ 77%⁵ and 70%¹¹ for a survey of INR reporting in Canada, and 85% for a survey of prothrombin time (PT) monitoring of anticoagulation therapy in Massachusetts.¹²

Performance of Coagulation Testing

Of those responding, 98% of the large hospitals and 97% of the small hospitals stated that they performed coagulation testing.

Test Requisition and Specimen Management

Use of test requisition forms and information items requested on them. Fifty-six percent stated that they used test requisition forms. Responses of large and small hospital laboratories were as follows:

It is conceivable that some participants negatively responding to this question may have done so because they ordered coagulation tests electronically without using a paper-based requisition form. There was a wide variation in the proportion of respondents requesting specific information items on test requisition forms:

These findings indicate that a minority of both large and small hospital respondents noted their policy against use of indwelling catheters for specimen collection. It has been recommended that, due to the presence of anticoagulants at such collection sites, specimens used for monitoring heparin therapy should be collected from a different extremity than the one used for heparin infusion.¹⁶

Practices Relating to Prothrombin Time (PT) Assay

Performance of PT assay. Of the respondents that provided valid responses, 100% noted performing PT assay.

Anticoagulant concentration. The respondents used the following sodium citrate concentrations:

The respondents reported exclusive use of 3.2% sodium citrate as follows:

Based on the WHO recommendations and the NCCLS guidelines, 3.2% citrate is the anticoagulant of choice for coagulation testing.¹⁴ In a recent study,¹⁸ the recommendation to use 3.2%, as opposed to 3.8%, sodium citrate was supported by noting that the concentration of sodium citrate had a significant effect on PT and aPTT assay results. In another study,¹⁹ underfilling of specimen tubes containing 3.8% sodium citrate prolonged PT and especially aPTT compared to 3.2% sodium citrate. When responsive PT and aPTT reagents are used, concentration of sodium citrate has a significant effect on assay results, with 19% of patients receiving intravenous heparin therapy having a greater than 7-second difference when aPTT results were compared.²⁰ In a survey of aPTT reporting in Canadian medical laboratories,¹ 46% of laboratories were still using 3.8% citrate. In this survey, 27% of the respondents noted that they used 3.8% citrate (5% of these respondents noted that they used both 3.2% and 3.8% citrate), potentially resulting in falsely elevated PT or aPTT results.

Reporting of PT results. Virtually all (99.8%) respondents used international normalized ratio (INR) to report PT; 97% also reported PT in seconds and/or as therapeutic PT ratio:

Reporting PT results in seconds may lead clinicians to inappropriately compare results between institutions.⁵ Reliance on therapeutic PT ratio has been documented to cause errors in anticoagulant therapy.^5,21 Reporting of PT results in INR only is, therefore, the preferred method. Reporting of PT results in INR has increased over the years from 557% in 1992^11,12 to 98100% in 1996^5,10 and 99.8% in the current survey. Reporting of PT results in INR only in Canada increased from 15% of all licensed medical laboratories in 1992 to 36% of these laboratories in 1996. This survey shows that 3% of the U.S. hospital laboratories reported PT results in INR only. Reporting PT results in both seconds and INR increased steadily over the years from 36% in 1992 (Canada) to 60% in 1996 (Canada) and 80% in 2001 (U.S.). Finally, the practice of reporting PT results in seconds only decreased from 36% in 1992 to <1% in 19962001.

Ninety-seven percent of the respondents provided PT results in seconds. This may be compared to the rates of 99% reported in 1992,¹² 95% reported in 1996,¹⁰ 72% in a 1992 Canadian survey reported in 1995,¹¹ and 60% in a follow-up 1996 Canadian survey reported in 1998.⁵

Sixteen percent provided therapeutic PT ratio. This rate may be compared to a rate of 43% in a survey of anticoagulant therapy monitoring reported in 1996,¹⁰ and rates of 13% and 2.5% in the 1992¹¹ and 1996⁵ surveys of Canadian medical laboratories, respectively. The rate of 43% in the former survey is less generalizable since it involved only 58 U.S. health centers, which were all academic institutions. Furthermore, those conducting the survey of these centers themselves have noted that wide variation existed in the reporting of coagulation tests (seconds and INR) and of patient therapeutic status.¹⁰

The following table shows how PT results were reported in our survey compared to the results of the 1992 and 1996 Canadian surveys:

Reference interval for PT assay. Ninety-two percent of the respondents conducted in-house evaluations to establish reference intervals for PT assay. Responses from large and small hospitals were significantly different:

The respondents not conducting in-house evaluations to establish reference intervals for PT did so as follows:

Most respondents (46% of the large hospitals and 74% of the small hospitals, P < 0.001) used less than 40 subjects to establish their PT reference intervals. To establish a reference interval, the NCCLS has recommended a minimum of 120 subjects for each reference population or subclass as the smallest number allowing determination of a 90% confidence interval around reference limits.²² In this survey, 5% of the respondents noted using at least 120 subjects to establish their reference ranges for PT assay.

Methods and reagents used to assay for PT. The respondents used the following methods to assay for PT:

Among thromboplastins listed, Innovin is the only recombinant protein (human tissue factor or factor VIIa).²³ Results obtained with recombinant thromboplastin on an optical analyzer even after prolonged storage of the plasma samples at room temperature were considered suitable for oral anticoagulation control.²⁴ Advantages of recombinant reagents such as Innovin and Recombiplastin are purity and consistency of reagents. Innovin and Recombiplastin have been shown to yield different INR values when compared to traditional reagents purified from tissue extracts.^23,25,26 Of all the respondents, 20% used Innovin. Of the large hospital respondents, 23% reported using Innovin compared to 16% of the small hospital laboratory respondents (P = 0.019). Although use of Recombiplastin as the other major recombinant thromboplastin was not assessed, larger facilities may use this and other recombinant reagents in preference to traditional reagents prepared from tissue extracts with their inherent variability from lot to lot or from reagent to reagent.

Sensitivity of PT assay to heparin

>Determining sensitivity of PT assay to heparin. Seventeen percent of the respondents determined the sensitivity of their PT assays to heparin.

Selecting a PT-thromoboplastin reagent insensitive to heparin in the therapeutic range. Fifty percent of the respondents selected a PT-thromboplastin reagent that was insensitive to heparin in the heparin therapeutic range. Significantly different proportions of the large and small hospitals selected an insensitive PT-thromboplastin reagent:

According to consensus guidelines developed at the 1997 CAP conference, laboratories should determine the sensitivity of their PT assay to heparin and, where possible, select a thromboplastin that is insensitive to heparin in the therapeutic range.¹⁴

International sensitivity index (ISI) of thromboplastin lot currently used. The ISI of the respondents current thromboplastin lot was 0.892.63 (average, 1.60; median, 1.81). The large hospital respondents reported an average ISI of 1.52 (median, 1.56), while the small hospitals reported an average of 1.70 (median, 1.89).

Due to increased variability in INR resulting from ISI values deviating significantly from unity, various professional organizations have recommended using thromboplastin reagents with ISI values closer to 1. The CAP recommends thromboplastins with a manual ISI between 0.9 and 1.7 and toward the lower end of this scale.¹⁴ The ISI of the respondents current reagent lots ranged between 0.89 and 2.63. Forty-four percent reported ISI values of < 1.70. Of the large hospital respondents, 50% reported ISI values of < 1.70 compared to 36% of the small hospital respondents (P = 0.001).

Two published articles have reported ISI values of survey participants. A 1992 report involving 88 acute care hospitals in Massachusetts reported ISI values of 1.892.74.¹⁰ The second involved a survey of all licensed Canadian medical laboratories in 1996 and 1992.⁵ This latter report noted that average ISI value had decreased from 2.07 in 1992 to 1.63 in 1996.¹¹ Of the respondents to the 1996 survey of Canadian medical laboratories, 35% had reported ISI values of < 1.2 as recommended by the American College of Chest Physicians.²⁷ In this survey, 34% of the respondents used an ISI of < 1.20. Forty-two percent of the large hospital respondents reported ISI values of < 1.20 compared to 24% of the small hospital respondents (P < 0.001).

In a survey of 58 academic institution in the U.S., 79% of the hospitals did not confirm accuracy of the ISI for their own analyzers.¹⁰ Coagulation instruments may affect ISI, which can differ from the assigned value by a clinically significant degree.^14,28 We did not capture in this survey the proportion of the respondents confirming accuracy of the ISI value for their own analyzers.

Practices Relating to Activated Partial Thromboplastin Time (aPTT) Assay

Performance of aPTT assay. Ninety-nine percent of the respondents performed aPTT assay.

Therapeutic range. The respondents noted having an aPTT therapeutic range for heparin as follows:

Specimens for aPTT assay are stable for up to 8 hours at room temperature except for patients receiving unfractionated heparin therapy.³⁴ Heparinized samples, when stored uncentrifuged at room temperature, demonstrate a clinically significant shortening of aPTT and individual samples demonstrate a more than 50% decrease in ex vivo heparin levels at 4 hours. According to an approved NCCLS guideline,³⁵ samples can be assayed up to 4 hours after phlebotomy if centrifuged within 1 hour of collection.

Ninety-six percent of all respondents stated that specimens were assayed within 4 hours after phlebotomy, and 88% noted that specimens were centrifuged within 1 hour of collection. Also, 22% of the respondents noted that specimens were kept at 4 C prior to testing, and 82% of the respondents stated that specimens were kept at room temperature before testing. In the 19981999 survey of Canadian medical laboratories,¹ 90% of responding laboratories reported analyzing specimens within 4 hours of specimen collection. Of the participants responding to 1 or both questions relating to aPTT assay within 4 hours after collection and tube centrifugation within 1 hour of specimen collection, 99% adhered to at least 1 practice and 74% adhered to both. In this survey, of the large hospital respondents, 69% assayed for aPTT within 4 hours of phlebotomy and centrifuged specimens within 1 hour of collection compared to 79% of the small hospital respondents (P = 0.005).

Practices Relating to Assays for von Willebrand Disease (vWD)

Performance of von Willebrand factor antigen (vWF Ag) assay. Six percent of the respondents performed vWF Ag assay. Responses from large and small hospitals were significantly different:

Reporting of ABO specific reference interval for vWF Ag assay. Nineteen percent of the respondents that performed vWF Ag assay, reported an ABO specific reference interval for this assay.

Significantly higher vWF Ag levels have been found in individuals homozygous for the Se allele than in those heterozygous for this allele,³⁶ and lower levels of factor VIII and vWF Ag have been reported in individuals with blood type O compared to individuals with other ABO blood types.³⁷ Data clearly show significant linkage between the ABO locus and vWF Ag, with levels of vWF Ag exhibiting significant differences between O heterozygotes and non-OO homozygotes. However, use of ABO adjusted ranges for vWF Ag levels might not be essential for the diagnosis of this disorder; bleeding symptoms may depend on vWF Ag regardless of the ABO type.³⁸

Methodology used for vWF Ag assay. The respondents used the following methodologies to assay for vWF Ag:

The electroimmunodiffusion (EID) method, also known as Laurel rocket immunoassay is associated with the greatest variability in vWF Ag test results, and both the EID and LIA methods show poorer sensitivity at low vWF Ag levels compared to ELISA methods.⁴ Forty percent of the respondents used LIA or EID methods.

Performance of assay for von Willebrand factor (Ristocetin cofactor) activity. Seven percent of the respondents performed assay for Ristocetin cofactor activity. Responses from the large and small hospital respondents were significantly different:

Methodology used for assay of Ristocetin cofactor activity. The respondents used the following methodologies to assay for Ristocetin cofactor activity:

Collagen binding assays for vWF activity (Ristocetin cofactor activity) have been shown to perform better than other assays in terms of their ability to detect functional von Willebrand factor (vWF) discordance, i.e., Type 2 vWD.⁴ However, these assays are less sensitive than vWF Ag assay in terms of their ability to measure overall level of vWF since they detect only highly adhesive vWF. For assays of vWF activity, the greatest variability in results and the poorest sensitivity to low vWF Ag levels was obtained using the platelet agglutination/aggregation methods; however, these methods showed better performance in identifying Type 2 vWD.⁴ In our survey, 5% performed ELISA to assay for vWF activity and 2% used a collagen binding assay. The remaining 93% used platelet function tests based on platelet agglutination/aggregation as functional tests for vWF activity. Of the respondents noting that they performed 1 or more vWF assays (antigen, activity or multimers assays), 38% performed both vWF Ag assay and a vWF activity assay, 25% performed only vWF activity assays, 15% performed only vWF Ag assays, 15% performed all 3 vWF assays, and 6% provided results for vWF multimers.

Sixty-nine percent of the respondents performed vWF Ag test. In a 1999 Australasian multi-laboratory survey of diagnostic practice and efficacy of laboratory tests for vWD, all 25 laboratories performed tests for vWF Ag, 60% of the respondents performed both vWF Ag assay and a vWF activity assay, and 40% performed all 3 vWF assays.⁴ In our 2001 U.S. survey, 46% performed a single vWF test (vWF activity, 25%; vWF Ag, 15%; vWF multimers, 6%) to evaluate vWD, while no respondents in the 1999 Australasian survey did so. Assaying for vWF Ag will not detect many qualitative defects; so, use of this assay alone will lead to many Type 2 vWD patients being missed by the laboratory.⁴ A survey conducted in Australasia pointed to considerable variation among laboratories in the tests and methods as well as the composite test panels used to diagnose vWD.³ What we have observed here implies a laboratory practice pattern not consonant with an adequate assessment of vWD.³⁹ Performance of vWF Ag assay by only 69% of the respondents and performance/use of only a single test by 46%, including 15% using vWF Ag alone, were somewhat surprising. Caution is indicated against diagnosis of vWD made based on single vWF assay results.⁴⁰ In our survey, 79% of the respondents assayed for vWF activity; and they did so either alone (25%), in combination with vWF Ag assay only (38%), and in combination with both vWF Ag and vWF multimers assays (15%). These results should be viewed in light of the fact that the questions on vWF Ag and vWF activity dealt with actual performance of these assays; some participants responding negatively to these questions may provide results for vWF Ag and vWF activity tests by sending their specimens to outside laboratories.

Provision of vWF multimers results. Two percent of the respondents provided results for vWF multimers. The responses from large and small hospitals were significantly different:

The respondents noted the following circumstances for the assay of vWF multimers:

This assay is used to confirm or subtype vWD that has been previously diagnosed.⁴⁰ It has been suggested that it would not be appropriate to assay for vWF multimers during initial vWD investigation process. Although 8 of the 11 the respondents assayed for vWF multimers in combination with assays for vWF Ag and vWF activity, 3 respondents assayed for vWF multimers without noting that they assayed for either vWF Ag or vWF activity. In 2 recent Australasian surveys of vWF testing laboratories, 1216% of the respondents assayed for vWF multimers.^3,4 This may be compared to 21% of the U.S. hospital laboratory respondents doing the same as reported in this survey. The result of this survey should be considered in light of the observation that the value of vWF multimers assay is diminishing in the face of better test systems and diagnostic processes.⁴⁰ Assays for vWF multimers are used to imply probable Type II A or Type II M vWD and not Type II B vWD.⁴⁰ Interestingly, 9 of the 11 respondents stated that they performed vWF multimers assays only when ordered by a clinician.

The following shows vWF testing patterns in this 2001 U.S. survey and a 1999 Australasian survey:

Practices Relating to Thrombosis/Hypercoagulability Workup

Protein S assays. Five percent of the respondents usually performed the assay for protein S activity (functional test) before the antigenic assay. The responses from large and small hospitals were significantly different:

If the results of the functional test were decreased, 17% performed antigenic assay to differentiate Type I deficiency from Type II while 20% performed free and total protein S antigen assay.

Performance of activated protein C (APC) resistance and factor V Leiden mutation assays. Six percent of the respondents performed activated protein C (APC) resistance assay. The responses from large and small hospitals were significantly different:

If after performing the APC resistance assay, results indicated resistance to APC, 61% obtained results for factor V Leiden mutation.

More than 11 million Americans are factor V Leiden carriers, and 143,000 are homozygous for this mutation.⁴¹ Homozygosity or heterozygosity for factor V Leiden in the absence of symptoms does not necessarily indicate that preventive treatment is required.⁴² Furthermore, there is no established intervention to reduce thrombotic risk. Thrombotic risk can be modified by various acquired and environmental conditions such as pregnancy, oral contraceptives, estrogen therapy, malignancy, stroke with extremity paresis, trauma, surgery or immobility.⁴³

Depending on the APC resistance functional assay used and the cut-off values for defining an abnormal result, factor V Leiden mutation may account for 8595% of patients with APC resistance.⁴³ Factor V Leiden mutation is believed to produce a relative risk of venous thrombosis of 7-fold in the heterozygous state and 80-fold in the homozygous state. Even in the homozygous state, however, this mutation does not appear to cause disease early in life, as seen with protein S and protein C homozygosity. Activated protein C resistance with normal factor V genotype is a risk factor for venous thrombosis.⁴⁴ However, there is no major effect of APC resistance on life expectancy.⁴⁵ Consequently, long-term anticoagulation in carriers of factor V Leiden, on the basis of the carrier state alone, is not indicated. In view of all these, it was interesting to note that 61% of the respondents obtained results for factor V Leiden mutation if results indicated resistance to APC.

Algorithm for Diagnosing a Lupus Anticoagulant (LA)

Offering an LA profile. Eighteen percent of the respondents offered an LA profile. The responses from large and small hospitals were significantly different:

Our obtained rate of 18% for offering an LA profile may be compared to the rate of 34% obtained in the 19981999 Canadian survey.¹

Practices leading to mixing studies. Participants noted the following circumstances for performing a mixing study when PT result was prolonged: