Inherited thrombophilia and preeclampsia: is the evidence beginning to congeal?

As part of the routine clinical practice of maternal-fetal medicine, I am often asked to consult with both patients and physicians following an adverse pregnancy outcome. Frequently, an exceptionally thorough evaluation has been initiated by the referring physician, and the patient arrives with a small mountain of paperwork that requires review. Almost invariably, a comprehensive search for inherited thrombophilia is included in the laboratory results.

Despite the ubiquity of the thrombophilia workup, it is not at all clear that inherited thrombophilias are associated with adverse pregnancy outcome. Many studies have evaluated the putative association, but the results are frequently contradictory and the absolute risk small, when any is found at all.1

Preeclampsia, manifested by new-onset hypertension and proteinuria during pregnancy, complicates 3-5% of deliveries and appears to have future health consequences.2, 3 A potential association between inherited thrombophilia and preeclampsia has been the subject of significant scientific scrutiny. A PubMed search of the terms thrombophilia and preeclampsia returns 530 results, with more than 350 of these publications designated as original research. Unfortunately, many of these studies are retrospective in nature and are underpowered to detect an association, especially with regard to a specific thrombophilia. To detect an odds ratio greater than 2 with 80% power in a population with a prevalence of 5% for the thrombophilia in question, a sample size of at least 1000 is necessary.4 Regardless of design, few studies meet these criteria, and even fewer have narrow confidence intervals.4

In this issue of the Journal, Kahn et al5 report the results of the Montreal Preeclampsia Study, a nested case-control study within a cohort of more than 5000 pregnant women who were participants in a large multicenter study of causal pathways of preterm birth. The authors examined the association between the Factor V Leiden mutation, the prothrombin G20210A mutation, and the MTHFR C677T polymorphism and preeclampsia. Additionally, homocysteine and folate levels were measured and placentae underwent pathological evaluation as part of a substudy to assess the reliability of microscopy and histopathologic features consistent with presumed placental underperfusion, such as infarction, decidual vasculopathy, and syncytial knotting.

In their report, Kahn et al did indeed find pathologic suggestion of placental underperfusion (as defined by the authors) more frequently in preeclamptic subjects than controls (63.2% vs 46.2%; P < .001). However, Factor V Leiden, prothrombin G20210A, and MTHFR C677T (homozygous) mutations were not more frequent in women with preeclampsia than in control women (cases 14.3%, controls 20.6%; adjusted odds ratio [OR], 0.6; 95% confidence interval [CI], 0.3-1.3).

Paradoxically, homocysteine levels were lower in cases than in controls (mean [SD], 3.4 [0.9] vs 3.7 [0.9] μmol/L; P = .002), and cases were less likely than controls to have homocysteine levels in the top quartile of values (17% vs 25%; P = .05; adjusted OR, 0.8; 95% CI, 0.4-1.4), whereas folate levels were similar in cases and controls (mean [SD], 39.6 [32.4] vs 41.1 [36.5] nmol/L; P = .69). Reassuringly, obesity, diabetes, and previous preeclampsia were predictive of preeclampsia in the index pregnancy, as would be expected. Overall, though, these results contradict the findings of several other prominent studies cited by the authors.

Among the strengths of this study are its initial prospective design and comparatively large sample size. The pathologic results are especially intriguing and support the concept of placental syndromes.6 However, even a study of this size is underpowered to interrogate the effect of any specific thrombophilia. The authors also did not do functional assays to search for the presence of other thrombophilias, such as antithrombin, protein C, and protein S deficiencies, so their work is limited to a few genetic mutations.

Despite the negative findings, these results should not be interpreted to mean that all thrombophilias lack clinical relevance. Most thrombophilias markedly have an impact on the risk of maternal venous thromboembolism, especially during prolonged immobility or after cesarean section.7 Likewise, antiphospholipid syndrome has been associated with complications, such as recurrent pregnancy loss, fetal growth restriction, and stillbirth.8

Although the work of Kahn et al is unlikely to be the final word in the debate about whether thrombophilia is a cause of preeclampsia, this study does lend significant support to those who contend that if an association exists, it is likely to be a weak effect without major clinical implications. Nevertheless, even if one continues to subscribe to the belief that preeclampsia and thrombophilia are linked, it is a quantum leap further to assume that anticoagulation of any type will have an impact on the risk of preeclampsia. The evidence for that therapeutic approach is largely lacking, and as has been asserted by 1 group of experts,4 anticoagulation of patients with an inherited thrombophilia to prevent preeclampsia should not be attempted outside of a clinical trial.