Pregnancy and Antiphospholipid Syndrome

Authors

  • Michael D. Lockshin

    Corresponding author
    • Department of Medicine and Obstetrics-Gynecology, Weill-Cornell Medical College, Hospital for Special Surgery New York, New York, NY, USA
    Search for more papers by this author

Correspondence Michael D. Lockshin, Hospital for Special Surgery, 535 E 70th Street, New York, NY 10021, USA E-mail: LockshinM@HSS.edu

Abstract

APS causes adverse pregnancy outcome, especially in younger patients with SLE or prior thromboses. LAC is the best predictor of adverse outcome. Prior conclusions on the efficacy of heparin for prevention of poor outcome may be suspect. New therapies are under evaluation in animal models but how to introduce them into the clinic is unclear.

The antiphospholipid syndrome (APS) is the occurrence of a clinical event (thrombosis and/or recurrent pregnancy loss) in a patient who carries a high-titer diagnostic antiphospholipid (aPL) autoantibody for at least 12 weeks.[1] APL is a general term that encompasses IgG or IgM anticardiolipin (aCL), IgG or IgM antibody to beta-2-glycoprotein I (aβ2GP1), and/or lupus anticoagulant (LAC). Formal definitions exist for all the antibodies.[2, 3] Low-titer antibody, IgA antibody, false-positive tests for syphilis, and antibodies to phospholipids other than cardiolipin do not fulfill consensus definitions, primarily because they either have not passed rigorous standardization tests nor have they been systematically correlated with clinical events.[4]

Both the clinical and the serological criteria must be present to diagnose APS. International consensus panels have defined the details of the criteria.[1] APS is often associated with a systemic autoimmune illness, systemic lupus erythematosus (SLE) being most frequent, but about half of patients have no recognized other illness. Asymptomatic persons with antiphospholipid antibody are frequently seen. Absent symptoms, they do not have APS; in general, these persons do not need treatment.

Patients with APS may have features not included among the diagnostic criteria.[5] Such features may include thrombocytopenia, livedo reticularis, a renal lesion known as thrombotic microangiopathy, non-stroke central nervous system disease (hyperintense non-enhancing white matter lesions on brain MRI; cognitive dysfunction), cardiac valve vegetations, and intracardiac thrombi. Most of these are late findings, but they occasionally occur in pregnant patients.

It is important to understand that the serologic tests do not closely correlate with one another.[6] A patient may have high-titer aCL or aβ2GP1 or LAC with or without the other autoantibodies. Some authors argue that ‘triple positivity’, i.e. all three antibodies positive, carries the worst prognosis,[7, 8] while others argue that LAC alone worsens prognosis independent of aCL or aβ2GP1 titer.[9] Most experts concur that coexisting predisposing factors, such as smoking or diabetes or orthopedic surgery (for thromboses) or hypertension, diabetes, or renal insufficiency (for pregnancy) worsen prognosis in individual patients.

A literature dating from the 1990s states that treatment for a pregnant woman with APS should consist of low-dose heparin if she has not had prior thrombosis and anticoagulant dose heparin if she has had prior thrombosis.[10-12] Most recommendations also include the use of aspirin in doses of 50–325 mg/day. No studies distinguish between unfractionated and low molecular weight heparin; early studies were carried out with unfractionated, but most American obstetricians now use low molecular weight heparin because of its ease of use and greater safety profile and despite its higher cost. Studies on which recommendations to use heparin are based on a rule which did not stratify patients by clinical or serological risk characteristics; for outcome, they mostly used fetal survival whether or not the pregnancy was premature or complicated by pre-eclampsia. Controlled treatment trials are few.

Our own recently published study[9] offers some insight into this literature and suggests a need to rethink some conclusions.

Predictors of PRegnancy Outcome: BioMarkers In Antiphospholipid Syndrome and Systemic Lupus Erythematosus (PROMISSE) is an NIH, Alliance for Lupus Research and Mary Kirkland Center-funded multicenter observational study on pregnancies of patients with APS, SLE, or both, compared with normal controls (Jane E. Salmon, Principal Investigator). Normal subjects had no diagnosable disease, negative APS tests, and at least one prior successful pregnancy without recurrent pregnancy loss. Our publication described 144 patients with aPL (with or without APS), all of whom had tests confirmed in core laboratories.

It is noteworthy that 612 women had been referred as having either SLE or aPL; of these, 142 were confirmed as having aPL, as were two women from the ‘normal’ controls; 274 women had SLE without aPL. Only 132 of 279 thought to be aPL positive on the basis of outside commercial referral laboratories were confirmed in the core laboratory, and an additional 10 SLE patients thought to be negative in commercial laboratories were positive in the core laboratory, a testament to the continuing inaccuracy of commercial laboratory results in this area.

To avoid confounding with chromosomal, infectious, and other causes of fetal loss, we did not study pregnancies lost before 12 weeks. Hence, our data speak only to pregnancies lasting ≥ 12 weeks. Similarly, to obtain clear answers, we did not include twin or higher multiple pregnancies, and we excluded patients with severe nephritis or in severe lupus flare. We counted only one pregnancy per woman. Our data therefore apply to women who are relatively well at pregnancy onset.

We defined adverse pregnancy outcome as fetal death ≥ 12 weeks, neonatal death, preterm delivery <34 weeks due to gestational hypertension, pre-eclampsia, or placental insufficiency, or fetal size ≤5th percentile. We defined LAC by a confirmed positive test following an abnormal screening test using the dilute prothrombin, dilute Russell's viper venom, or dilute activated partial thromboplastin times (dPT, dRVVT, and dAPTT, respectively). We performed all three tests in all women at enrollment, at each trimester, and three months postpartum.

Thirty-five percent of women with LAC and low-titer (<40 GPL units) IgG aCL had adverse pregnancy outcome, mostly live infants, as did 43% with LAC and high-titer aCL (difference between low- and high-titer aCL not significant). By contrast, no women with low titer and only 8% of women with high-titer aCL who did not have LAC had adverse outcome, a rate not different from normal. If a woman had SLE and LAC, she had a 55% risk, and if she had LAC but did not have SLE, a 35% risk. Aβ2GP1 IgM and triple positivity added no information not given by LAC alone. Among the screening tests, dPT was sensitive but not specific (positive predictive value [PPV] 0.808, negative predictive value [NPV] 0.705, dRVVT most specific (PPV 0.560, NPV 0.910), and aPTT intermediate (PPV 0.630, NPV 0.829). Any positive LAC test gave a PPV of 0.926, and IgG aCL <40 gave an NPV of 0.883. Test at first pregnancy visit predicted outcome; subsequent tests during pregnancy did not change our conclusions.

Other features that, in multivariate analysis, adversely affected outcome were a diagnosis of SLE, prior thrombosis, and young maternal age. Surprisingly, parity and number of prior losses were not predictive. Surprisingly, also, when patients were stratified according to presence or absence of LAC, heparin (initiated at pregnancy confirmation), whether low molecular weight or unfractionated, did not influence outcome; aspirin (all but one patient received 81 mg/day) was associated with fewer adverse events. Neither hydroxychloroquine nor corticosteroid therapy influenced outcome, but few patients took either of these drugs.

The current theory about mechanisms of APS-associated pregnancy complication is that cell injury or activation leads to binding of β2GP1 to a receptor, whereupon β2GP1 undergoes a conformational change, leading to its recognition by aβ2GP1.[13] This recognition activates complement and triggers intracellular events leading to exteriorization of tissue factor and adhesion molecules, and activation of fluid phase inflammation and coagulation molecules and complement, resulting in local inflammation and thrombosis.

Several new treatments, based on steps in this mechanistic pathway, are under consideration for patients with APS; however, very few are being tested in pregnancy. Treatment options include anticoagulation alone (but in animal models, non-heparin-based anticoagulation is ineffective),[14] antiplatelet therapy, inhibition of complement (heparin is anticomplementary and in animal models probably works by inhibiting complement rather than by preventing placental thrombosis), B-cell inhibition, increasing aPL clearance, as with plasmapheresis or use of intravenous immunoglobulin, and blocking the binding of aβ2GP1 to its receptor. A peptide inhibitor of aβ2GP1 is under test in animal models.[15] Given the wide belief that heparin is useful in APS, designing treatment trials for pregnant women with aPL will be difficult.

Ancillary