EVALUATION AND COMPARISON OF MIXING STUDY RESULTS USING PERCENT CORRECTION FORMULAE (CHANG VS. ROSNER INDICES) TO OPTIMIZE AND STANDARDIZE THE REPORTING OF MIXING STUDY RESULTS. F. ALI, M.D., S. BHAGAVATTI, M.D., J. C. MATTSON, M.D.* CLINICAL PATHOLOGY

The Antiphospholipid syndrome is an acquired immune-mediated thrombotic disorder. Its diagnosis requires demonstration of antiphospholipid antibodies (APA). Some APA act in vitro as an "anticoagulant" called lupus-anticoagulant (LA) prolonging results of clotting assays. This prolongation cannot be corrected by mixing studies in which normal plasma is added to the patient samples. Such mixing studies are commonly used in the coagulation laboratory to screen for inhibitors such as the LA. Deciding how much shortening of a clotting time constitutes a "correction" has proven to be a major problem in standardizing this assay. Two different formulae to standardize the calculation of percent correction have been introduced in the literature (Chang’s formula and Rosner index).

The purpose of this study was to examine the performance characteristics of these two indices and to determine which one more accurately reflects the presence of a "lupus anticoagulant". In a retrospective analysis of cases from 2001-2005, the result of each index was compared to the results determined by subjective interpretation (i.e. the original reported presence or absence of an inhibitor). Results of other tests for lupus anticoagulant and patient’s history of thrombotic events or fetal loss were reviewed and correlated with mixing study results.

 

Chang’s formula:               PP aPTT – 1:1 Mix aPTT x 100                                                  Positive = < 58% Borderline= 58%-70%

                                               PP aPTT – CNP APTT

Rosner Index:                    1:1 Mix aPTT – CNP aPTT x 100                                                     Positive = >15%

                                                          PP aPTT

 

53 patients (20 males and 33 females) with original reports of positive mixing studies for prolonged aPTT were found in this 5-year period. A random sampling of 44 mixing studies originally reported to be negative in the same period were also collected for comparison.

Chang’s

Chang’s

Positive

Borderline

Positive n= 53

 

30(57%)

 

15(28.3%)

 

37(70%)

Negative n= 44

 

6(14%)

 

9(20%)

 

3(7%)

 

The Chang’s formula showed sensitivity of 85% and specificity of 66%, while Rosner index showed sensitivity of 70% and specificity of 93%. The Chang’s formula has better sensitivity, while the Rosner index has better specificity for detection of lupus inhibitor. Our data suggests that while neither index is perfect, the use of two indices together may help to standardize reporting of positive, negative and equivocal results.