Clinical Significance of Antiphospholipid Protein Antibodies. Receiver Operating Characteristics Plot Analysis

JACEK MUSIAL, JAKUB SWADZBA, AGNIESZKA MOTYL, and TERESA IWANIEC

ABSTRACT.

Objective. To investigate the utility of receiver operating characteristic (ROC) analysis in determining the strength of association between various antiphospholipid and anti-protein cofactor antibodies (aPA) and thrombosis, pregnancy morbidity, and thrombocytopenia.

Methods. Clinical and laboratory variables were retrospectively studied in 204 patients: 160 with systemic lupus erythematosus (SLE), 22 with lupus-like syndrome (SLE-LS), and 22 with primary antiphospholipid syndrome (APS). Laboratory evaluation included detection of lupus anticoagulant (LAC) and measurement of IgG and IgM anticardiolipin (aCL), antiphosphatidylserine (aPS), antiphosphatidylinositol (aPI), anti-ß2 glycoprotein I (aß₂GPI), and antiprothrombin (aPT) antibodies. ROC plot analysis was used to determine the clinical accuracy of aPA tests, and calculate cut-off values which best associate with clinical symptoms typical for APS.

Results. The LAC was associated with a history of thrombosis [odds ratio (OR): 3.04; 95% confidence interval (CI): 1.5-6.2] and even more strongly with recurrent fetal loss (OR: 8.7; 95%CI: 2.8-26.7). ROC plot analysis revealed that the most accurate test for thrombosis was aCL IgG (ROC-derived cutoff value > 17.2 GPL; OR: 3.69; 95% CI: 1.8-7.4), for recurrent fetal loss, aPI IgG [> 22.1 theoretical units (TU); OR: 6.21; 95%CI: 2.1-18.5], closely followed by aCL IgG and aß₂GPI IgG, and for thrombocytopenia aPS IgM (> 6.7 TU; OR: 1.9; 95%CI: 1.04-3.4). Among 182 autoimmune patients (SLE + SLE-LS), 6.6% presented clinical symptoms of APS without classic aPA (LAC and/or aCL), but with elevated levels of antibodies against other phospholipids, mainly aPI IgM. Conclusion. A laboratory that evaluates APS patients should establish its own threshold values for aPA tests. We suggest that ROC plot analysis is a valuable tool in establishing cutoff values. LAC and aCL determinations seem sufficient for the majority of laboratories. However, in specialized centers other tests should be available to detect those patients with clinical symptoms for APS but who are positive for antiphospholipid antibodies other than aCL and the LAC. (J Rheumatol 2003;30:723-30)

Key Indexing Terms:

ROC PLOTS
ANTIPHOSPHOLIPID SYNDROME
ANTIPHOSPHOLIPID ANTIBODIES
LUPUS ANTICOAGULANT

---

From the Department of Medicine, Jagiellonian University School of Medicine, Kraków, Poland.

Supported by a grant from the Polish Committee for Scientific Research No 6PO5B 13620.

J. Musial MD, PhD, Professor; J. Swadzba MD, PhD; A. Motyl MD, PhD; T. Iwaniec MS.

Address reprint requests to Prof. J. Musial, Department of Medicine, Jagiellonian University School of Medicine, Skawinska 8, 31-066 Kraków, Poland. E-mail: mmmusia@cyf-kr.edu.pl

Submitted March 4, 2002; revision accepted September 4, 2002.

---