| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Letters to the Editor |
1 Department of Haematology, and Transfusion Medicine, Royal North Shore Hospital, Sydney, 2065, Australi
2 Northern Blood Research Centre, University of Sydney, RNSH campus, Sydney, Australia
3 Pacific Laboratory, Medicine Services, Royal North Shore Hospital, Sydney, Australia
aAddress correspondence to this author at: PaLMS Haematology, Royal North Shore Hospital, St. Leonards, NSW, 2065, Australia. Fax 612-9926-6066; e-mail maboud{at}nsccahs.health.nsw.gov.au.
To the Editor:
The recommended (1) step-wise approach to the study of lupus anticoagulant (LA) begins with a prolonged clotting time in a phospholipid-dependent clotting assay (screen), a noncorrection in a mix with pooled normal plasma (PNP), and then a correction in the presence of increased phospholipid (confirm). Laboratories differ in their application of these recommendations. Jacobsen et al.(2) integrate screen, mix, and confirm into a single assay and analyze the data as a lupus ratio (LR). Tripodi et al.(3) propose no prior mix with PNP even for patients on oral anticoagulant (OAC) and analyze their data as a percentage correction. We wished to compare these 2 approaches by using the common(4) dilute Russell viper venom test (dRVVT) as the clotting assay.
Samples were collected from 500 consecutive patients (291 women, median age 56 years) for whom LA tests were requested during a 5-month period. Repeat tests (n = 19) were excluded on patients as were 9 samples that contained unfractionated heparin. Of the remaining 491 patients, 85 were on OAC with international normalized ratios ranging from 1.7 to 3.5.
Plasma from 70 healthy volunteers (30 men and 40 women, median age 48 years) were used to establish the cutoff values. We used PNP from a pool of 30 similar healthy volunteers.
Blood from controls and patients was collected and stored as described (5). The LR of Jacobsen et al.(2) is the ratio of 2 clotting times for 1:1 mixes of patient’s plasma and PNP, one using a dilute phospholipid (screen) assay and the other using a phospholipid-rich reagent (confirm) assay. This ratio is normalized by dividing by the corresponding ratio for PNP performed in the same assay. The percentage correction based on the study of Tripodi et al.(3) is the difference between the test plasma screen normalized by PNP screen and test plasma confirm also normalized by PNP confirm. This difference is expressed as a percentage of the normalized screen clotting time.
Clotting times for the same batch of PNP provided data for between-assay imprecision for over 20 assays (screen, 0.91%; confirm, 0.88%). The between-assay imprecision over 10 assays for LA-positive patient plasma was 1.4% for screen and 1.3% for confirm. The within-assay imprecision for PNP was 0.34% for screen and 0.27% for confirm (n = 12), and for the LA-positive patient plasma was 0.61% for screen and 0.59% for confirm.
For the dRVVT-based assays we used the La Screen and La Confirm reagents from Life Therapeutics. All tests were performed on an STA-R coagulation analyzer from Diagnostica Stago. We used the 
2 test to compare the LA positive pick-up rate in the LR and percentage correction assays. We constructed 2 x 2 tables to compare results derived from the LR and from the percentage correction.
Both LR and percentage correction for 70 individual normal individuals were normally distributed. The cutoff values (2 SD above the mean) for 70 normal individuals were LR = 1.08 and percentage correction = 13.5%.
For all patients, including the subset on OAC, the overall agreement of the 2 methods (LR and percentage correction) was 462/491 = 94%. The increased positive test rate for percentage correction (96/491 = 19.5%) compared with the LR (73/491 = 14.9%) was statistically significant at P = 0.05 (Table 1
).
|
We subdivided the patient population into those on OAC (n = 85) and those not on OAC (n = 406). For patients on OAC, the concordance between the 2 methods was 74/85 = 87%. The increased positive test rate for percentage correction (23/85 = 27%) and LR (14/85 = 16.9%) did not differ significantly.
Samples from 26 patients (16 women, ages 19–76 years, median age 47 years) were LA positive by percentage correction and LA negative by LR (Table 1
). The incidence of thrombotic disease (17 venous thromboembolic and 2 arterial thromboembolic) was high (73%) in this group. Six of these patients, all of whom were diagnosed with venous thromboembolic disease, were clearly positive with a percentage correction >3 SD (>18.1%) above the mean.
It has been reported (6)(7) that 1:1 mixes of test plasma and PNP may dilute a weak, yet potentially pathogenic LA, yielding a false-negative result. The high incidence of thrombotic disease in the 26 patients (9 of whom were on OAC) poses the question of whether LA activity was diluted by the 1:1 mix in these patients. The increased sensitivity of the percentage correction to the presence of a potential LA compared with the LR may well be of clinical value, warranting further studies to clearly characterize the presence of an antiphospholipid antibody in these patients.
References
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
