Recurrent Fetal Loss and Antiphospholipid Antibodies: Clinical and Therapeutic Aspects

Recurrent fetal losses indicate screening for antiphospholipid antibodies, especially after the third consecutive fetal loss, or when they occur after 12 weeks gestation or when the mother presents with thrombosis or other ailments of antiphospholipid syndrome. Fetal loss may be caused by thromboses of placental vasculature. There is no agreement concerning the mechanism of thromboses: protein C pathway and/or annexin V are the best candidates. When fetal loss occurs early during gestation, murine models suggest that antiphospholipid antibodies can also act on trophoblasts by inhibiting syncytia formation. Among the high risk patients with more than two fetal losses, an association of aspirin and heparin given early during gestation is successful in 70–80% of cases.

pathway; annexin V; antithrombotic therapy ecurrent fetal loss (FL) affects 1-5% of women of childbearing age. 1,z Chromosomal abnormalities account for 60% of recurrent FL. Immunological mechanisms may account for the remaining 40%, and in particular, the antiphospholipid syndrome (APLS) appears to be implicated in 7-25% of recurrent FL. z In 1980, Soulier and Boffa 3 suggested a relationship between recurrent FL and lupus anticoagulant (LA). The APLS was defined 7 years later 4 based on a combination of clinical and biological abnormalities: on the one hand, venous or arterial thrombosis and/or recurrent FL; on the other, recurrent evidence of LA and/or anticardiolipin antibodies (ACL). Objections may be raised to this definition, 5 one of the major problems being the interlaboratory variation in ACL assays and the absence of consensus concerning the normal upper limit of ACL. 6 In our laboratory, we consider the normal upper limit to be 20 GPL units for IgG ACL and 20 MPL units for IgM ACL.
Antiphospholipid-associated obstetrical disorders are not limited to FL. They also include infertility, intrauterine growth retardation, fetal distress, prematurity, 7 and in the mother, preeclampsia 8 and chorea gravidarum. 9 We shall limit ourselves to FL, a term including abortion and fetal death, the limit between these being considered by Branch 1 to be at 20 weeks gestation.

WHEN TO SCREEN FOR ANTIPHOSPHOLIPID ANTIBODIES (APL)?
It is generally considered that screening for APL is unnecessary in asymptomatic gravida 0 or nulliparous women. 11-3 Indeed, LA and ACL positivity are infrequent in these women, being present in less than 4%, and the relationship to the outcome of pregnancy is unclear. However, the findings of Lynch et al. 14   Screening for APL also seems unjustified after a first FL. The case-control study conducted by Infante-Rivard et al. is with 331 patients and 993 controls showed 5.1% LA positivity in patients vs. 3.8% in controls (relative risk:l.42) and 1.2% IgG ACL positivity in patients vs. 1.5% in controls.
On the other hand, screening becomes cost effective after two, or better, three FL. 13'6-z The problem is that the percent APL positivityin these patients varies widely in different studies. Table   shows an LA variation from 4.5% to 48%, and the percent IgG ACL positivity varies from 6.8% to 35%. There is generally a significant difference between patient and control groups, except in the study by Petri et al. z In this study, the difference between a group of 44 patients with recurrent FL and a group of 40 controls was not significant; however, a number of questions can be raised about this study: the sensitivity of the method used for LA testing (using the dilute Russell's viper venom time) is probably too low, and the study population is not large enough to detect a significant difference (9% LA positivity in patients vs. 0% in controls, and 11% ACL positivity in patients vs. 2.5% in controls); moreover, in this study, more than 50% of patients had very early miscarriages.
Apart from the number of FL, other criteria should be taken into account, especially the date of the FL. Within a pregnancy, the first 12 weeks of gestation (the preembryonic and embryonic periods) should be distinguished from the following weeks which are, strictly speaking, the fetal period. According to Branch et al, zz whose experience of FL is quite extensive; FL occurs within the first 12 weeks of gestation in nearly half of cases, and in the other half, mostly between 16 and 20 weeks gestation. However, in several large studies, especially in a number of recent randomized therapeu-tic trials, pretreatment FL occurred mostly during the first trimester, z3-zs In Kutteh's z4 study, 80% of pretreatment FL occurred before 12 weeks gestation. In any case, the occurrence of FL after 12 weeks gestation rules out a syndrome of recurrent early pregnancy loss, or intrauterine embryonic death, ,z6 which always occurs before 12 weeks gestation and is not associated with detectable antibodies in the mother, and might be due to antitrophoblast natural killer (NK) cell toxicity, z7 The inevitable occurrence of FL in all subsequent pregnancies in women who have experienced one FL is also a strong argument for testing for APL. Lubbe  The occurrence of thrombosis in the mother during, or above all, immediately after pregnancy 31-33 is also an argument for screening for APL, and besides, two of three patients in Soulier and Boffa's 3 original description had had a history of thrombosis. However, the association of FL with thrombosis can also be due to a congenital coagulation disorder, especially resistance to activated protein C (APCR). In Brenner et al., 4 a Leiden mutation of factor V was present in 19 of 39 patients. The problem becomes more difficult in that acquired APCR without a Leiden mutation was present in 9 of 39 patients, and as is well known, APL can induce APCR. 3s,36 Apart from thrombosis, the other factors of the APL are also very suggestive (  According to other authors, 46 ACL is the best predictive factor for FL, but this finding should be qualified. The risk increases as ACL levels increase. 47,48 Persistence of ACL positivity is particularly important; 49 transitory ACL positivity has practically no diagnostic value, s In SLE patients, the relative risk of FL is 10 in ACL-positive patients compared with ACL-negative patients) s Aoki et al. sl demonstrate that among APLS, the incidence of antibodies to anionic beta 2 GPldependent phospholipids appears to be higher in patients who have had FL. However, when testing patients with APLS at risk for thrombosis, anti-beta 2 GP1 appears to be less sensitive than LA. Commercially available kits for global screening of APL also appear to have low sensitivity. In a personal study of 25 patients with ALPS and ACL, the sensitivity of screening with BMD kits was only 76%, and this test did not detect two of ten patients with recurrent FL syndrome. In the event of negative findings with the standard laboratory tests, in the context of highly suggestive clinical findings, screening for antibodies to anti-beta 2 GPl, or to neutral phospholipid such as phosphatidylethanolamine, is justified, sz In practice, evaluation of LA and ACL initially and at 8 weeks should be recommended in patients with a suspicion of APLS. In clinically suggestive cases with negative laboratory findings, it is justified to also test for false-positive VDRL, anti-beta 2 GPl, and antiphosphatidylethanolamine.

MECHANISM OF FL
Additional evidence of the relationship between APL and FL is shown with experimental models, s These models lead us to the investigation of the precise mechanism of FL.
1. Pathological findings show numerous placental infarcts in these patients, 4,41,4 suggesting that FL might be caused by thrombosis of the placental vasculature. There is no agreement as to the mechanism causing these thromboses. A fibrinolytic disorder, inhibition of prostacyclin formation by vascular walls, s4 and a disorder of the protein C pathway ss have successively been mentioned, and, more recently, an interaction between APL and annexin V. s6 The two latter hypotheses are the most convincingly documented.
As far as the protein C pathway is concerned, the presence of an acquired deficit of free protein S appears to be established in both APLS secondary The first question to be addressed is at what point treatment should be initiated, i.e., starting with which pregnancy, and at what time in the pregnancy. If the APLS is diagnosed because of FL, a complete workup, particularly immunologi-cal, should be carried out certainly by the third pregnancy, and APL positivity confirmed. In practice, this workup is often initiated after the second FL. Therefore, the patient will initially be treated at her third or fourth pregnancy. It seems logical to start treatment as soon as pregnancy is confirmed, since it has been shown that APL may inhibit the development of trophoblasts, thus inducing very early FL.
If APL positivity has been established before the first pregnancy, for instance in a SLE patient presenting with thrombosis, or even in an asymptomatic woman being operated or before she gets married, and if the patient is given no drug when she gets pregnant, in principle it would be possible to give her no treatment since such a patient is known to have a in 2 chance of a full-term pregnancy despite APL positivity. However, routine administration of aspirin (50-150 mg daily) starting from this first pregnancy seems advisable, since such therapy has been shown to be harmless for both the fetus and the mother. 64'6s What methods are available? In the 1980s, immunomodulant therapy was favored, but it is currently supplemented with antithrombotic therapy.

Antithrombotic Drugs
Several studies, some open, some randomized, 66'67 have demonstrated that full-term pregnancies could be achieved with aspirin alone, even in patients with a long history of FL (Table 3). In a study by Silver et al., 67   Success rate was 42% in the group on aspirin and 71% in the group on aspirin plus heparin. The risk of osteoporosis associated with heparin therapy given to patients during their entire pregnancy should be mentioned, [69][70][71] and it is worth noting that Rai et al. 2s showed a 5.4% decrease of lumbar bone density postnatally. Lastly, Italian researchers 72 treated 23 pregnancies in 22 patients with high dose omega 3 fatty acids alone (18 capsules daily, whereas the recommended dosage is 6 daily). The success rate was dramatic: 22 live births in 23 pregnancies (96%). It is unclear whether the antiaggregant action of fish oils alone was involved, or if another mechanism should be considered.

Immunomodulant Drugs
This treatment aims to reduce APL levels. Steroids and high dose intravenous immunoglobulins (IVIg) have been used successively, and alsom anecdotallymplasma exchanges. In fact, these therapies have practically always been combined with antithrombotic drugs, so that their specific efficacy cannot be readily assessed (Table 4).
Lubbe et al. 63 in 1983 and Branch et al. z9 in 1985 must be credited with demonstrating that in small (fewer than 10) groups of patients at very high risk of FL, administration of steroids and aspirin resulted in live births in most cases. Subsequent findings are conflicting. According to Silveira et al. 73 this combination is very efficacious. According to Lockshin et al., 74 it is not. In Lockshin et al.'s 74 study, 14 FL occurred in 21 pregnancies, but it should be noted that the dosage of steroids was not standardized and it was tapered rapidly during the pregnancy. In a recent randomized study by Laskin et al., 75 the combination of steroids and aspirin was no better than placebo, but in this study APL-positive patients were mixed with patients testing positive for antinuclear antibodies only. If only APL-positive patients are taken into account, the success rate is 60% in the group receiving steroids and aspirin and 52% in the placebo group.
However, it should be noted that the success rate in the placebo group is considerable, while it is admitted that less than 10% full-term pregnancies are achieved in high risk patients, z8,z9 In a personal non-randomized series including 27 LA-positive patients who had experienced several FL, 8 patients were given steroids alone and 19 patients steroids plus aspirin. The success rate was 3 of 8 (37.5%) in the group on steroids alone, and 10 of 19 (53%) in the group on steroids plus aspirin. The numbers were too small to reach significance. In this study, LA negativation was a good predictor of favorable outcome (68% live births in the subset of those who became LA-negative patients). least 12 reports in the literature, 76-8z but in all cases aspirin and/or heparin and/or steroids.were combined. We have used IVIg in three cases after failure of standard steroid plus aspirin treatment, successfully in all three. In one case, the outcome was satisfactory in spite of increasing ACL and beta 2 GP1 antibodies increasing again at the end of pregnancy.
Two reports of the efficacy of plasma exchanges have been published. 8z,83 Combination Therapies (Antithrombotic and/or Immunomodulant Drugs) At least two randomized studies have compared steroids and antithrombotic drugs (Table 5). In the study by Silver et al. 67 mentioned above, the outcome was dramatic in the group on aspirin alone (N 22) and in the group on aspirin plus steroids (N 12) since no FL occurred in either group. However, patients could be included after a single FL and 5 of 17 initially randomized patients in the aspirin plus steroids group were not evaluated. A randomized study by Cowchock et al. 84 compared heparin with steroids in a small group of 20 patients who also received aspirin; success rate was 75% in each group. In both studies, 67,84 premature deliveries were more frequent in patients who received steroids.
Therapeutic Strategies In women with a history of at least three FL, a minimal immunological workup (antinuclear antibodies, IgG ACL, LA) seems advisable. In practice, this is often done after only two FL. The initial intervention may include aspirin alone, 50-150 mg daily. After failure of this therapy, or if three FL have already occurred and if the patient is relatively old, aspirin and heparin may be prescribed from the beginning of pregnancy. We consider that the choice of a therapeutic dose of heparin administered in two subcutaneous injections is justified. Low molecular weight heparins are not licenced in France during pregnancy, but they have proved to be harmless in studies conducted in France and elsewhere. Laboratory follow-up during pregnancy includes at least a complete blood count (CBC), lactate dehydrogenase (LDH), haptoglobin and reactive protein, creatinine and uric acid, and transaminases (so that a HELLP syndrome is not missed). Fetal ultrasound scan with Doppler studies of the uterines and umbilical arteries are recommended from 22 weeks gestation; Doppler study of the uterine arteries is performed as early as 15 or 16 weeks gestation by some teams. Aspirin should be stopped at 35 or 36 weeks gestation, so that it has been withdrawn for at least 1 week when epidural anesthesia is performed. At that stage, aspirin should be replaced with two daily injections of low molecular weight heparin; if the patient was already on heparin, this treatment should be continued. Heparin should be interrupted for as short a period as possible at the time of delivery. These precautions are mandatory in patients with a history of thrombosis. Strokes have been observed in the immediate postpartum period in patients in whom all antithrombotic drugs had been withdrawn. 32 In the event of failure of combined aspirin plus heparin therapy, treatment with aspirin plus steroids 0.5 mg/kg/day, or even with aspirin plus steroids plus heparin, should be administered. Should 12. Harris