Unrealistic Estimation of Population Prevalence of Antiphospholipid Antibody–Related Pregnancy Loss: Comment on the Article by Andreoli et al

Authors


We read with interest the article by Andreoli et al ([1]) published recently in Arthritis Care & Research, particularly with regard to the prevalence of antiphospholipid antibodies (aPL) in women with pregnancy morbidity. We agree with their conclusion that it is difficult to determine the frequency of a clinically significant aPL profile because of the lack of consistent reproducible laboratory results and the heterogeneous nature of the study populations. Of the 47 studies included in their assessment of aPL-associated pregnancy morbidity (including pregnancy loss; intrauterine growth restriction; hemolysis, elevated liver enzymes, and low platelets syndrome; and preeclampsia/eclampsia, as listed in Supplementary Table 1 of the original article), only 9 studies confirmed the presence of aPL, 9 included patients with low levels of anticardiolipin antibodies, 5 did not report how aPL positivity was defined, and 15 did not include lupus anticoagulant in their aPL assessment. It would appear to be a Herculean task to derive meaningful prevalence data from such a discrepant group of studies for aPL-associated pregnancy loss among the general population.

We are concerned, therefore, that the authors attempted to do just that, using a median from 19 heterogeneous studies over the past 25 years. This resulted in an estimate of 50,000 potential aPL-related pregnancy losses per year in the US, a figure that appears to be remarkably inflated, especially given the patient accrual difficulties experienced by many, if not all, studies of aPL-related pregnancy loss. The Predictors of Pregnancy Outcome: Biomarkers in Antiphospholipid Antibody Syndrome and Systemic Lupus Erythematosus study, for example, prospectively evaluated aPL-positive pregnancies involving 7 centers around the US and Canada; after 6 years, a total of 144 aPL-positive patients had been enrolled (24 per year) ([2]). Cervera et al used the national registries of 20 university centers in 15 European countries over a 5-year period to accrue 1,000 patients with antiphospholipid syndrome (APS), of whom 250 had pregnancy morbidity (50 per year) ([3]). A recent study evaluating treatment efficacy recruited aPL-positive patients from 14 centers in the UK, Europe, and Mexico ([4]). After 5 years, 232 eligible patients were enrolled, of whom 83 (35%) had obstetric APS (17 per year). Cuadrado et al specifically identified low accrual as a major drawback in their ability to complete their study as planned within their proposed timeframe.

If there were 50,000 patients per year in the US with aPL-related pregnancy loss, as hypothesized (albeit cautiously), then these low study accrual data would be completely inexplicable. Therefore, we suggest a more reasonable explanation, that aPL-associated pregnancy loss is actually an uncommon, if not rare, event only seen after the hormonal, anatomic, and genetic etiologies of pregnancy loss have been excluded and aPL levels have been appropriately measured and interpreted following established guidelines.

Ancillary