Habitual abortions in 678 healthy patients: investigation and prevention

M.F. Reznikoff-Etievant1,4, V. Cayol1, G.M. Zou1, N. Abuaf2, A. Robert3, C. Johanet3 and J. Milliez1

1 Gynécologie–Obstétrique, Hôpital Saint Antoine, 184 rue du Faubourg Saint-Antoine, 75012 Paris, 2 Laboratoire d'Hématologie, Hôpital Rothschild, Paris, and 3 Laboratoire d'Hématologie, Hôpital Saint Antoine, Paris, France


    Abstract
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
The objective of this study of patients with habitual abortion (HA), was to determine their autoimmune profile and to try to prevent new abortions using low-dose aspirin for 7 months with prednisone in the first trimester only, or with low-dose aspirin alone. A total of 678 healthy patients with three or more HA were investigated for antiphospholipid antibodies, antinuclear and antithyroid antibodies. Among these patients, 277 pregnant women were treated, 214 were given prednisone and aspirin (161 autoantibody-negative and 53 autoantibody-positive women), and 63 autoantibody-negative women received aspirin alone. Autoantibodies were present in 33.9% of the patients, in 82.6% of them anticardiolipin antibodies were found to be isolated or associated with antiprothrombin, antithyroid, circulating anticoagulant, antinuclear or anti-ß2 glycoprotein 1 antibodies. In autoantibody-negative pregnant women treated by prednisone and aspirin or aspirin alone, the success rate of live births was 90.7% (146 out of 161) and 74.6% (47 out of 63) respectively (P < 0.01). In autoantibody-positive patients treated with prednisone and aspirin the success rate was 84.9% (45 out of 53) (not significant). Prednisone and aspirin seemed to be as efficient in autoantibody-negative or positive women but better than aspirin alone in autoantibody-negative women. A double-blind trial is in progress to confirm these results.

Key words: antiphospholipid/aspirin/autoantibodies/habitual abortions/prednisone


    Introduction
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
Habitual abortion (HA) may be associated with anatomic, genetic, or hormonal abnormalities, or autoimmune diseases, such as systemic lupus erythematosus or antiphospholipid syndrome. The frequency of autoantibody association with HA differs greatly from one study to the other, depending on the way patients are recruited, the tests performed and their threshold values (Stirrat, 1990Go; Gleicher, 1998Go). Although many studies have focused on the understanding of HA, about one third of them remain unexplained (Stephenson, 1996Go) and without well-established prevention.

Injection of the husband's leukocytes (Recurrent Miscarriage Immunotherapy Trialists Group, 1994Go), or i.v. {gamma}-globulin (German RSA/IVIG group 1994; Yamada et al., 1998Go) were used to prevent a new abortion in women with unexplained HA. Previously, it has been observed that 20% of autoantibody-free HA women became anticardiolipin (aCL)-positive in the next pregnancy (unpublished data) which could be the cause of the new abortion. On the basis of this observation, it was decided that glucorticoid and aspirin treatment should be used to try to prevent abortion.

The objective of this study was to determine the autoimmune profile of healthy women coming to the Department of Obstetrics and Gynecology for HA and before starting a double-blind control trial, to assess the possible prevention of a new abortion, in autoantibody-positive or -negative women, with low-dose aspirin until the end of the seventh month, without or with prednisone in the first trimester of pregnancy.


    Materials and methods
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
Patients
Only healthy patients were included: no autoimmune disease, systemic lupus erythematosus, arterial or venous thrombosis, chronic arthritis, Hashimoto's disease, high blood pressure, metabolic disease and no contraindication for prednisone or aspirin treatment. From February 1994 to September 1997 we selected 678 patients from our consultations who fulfilled the requirements listed in Table IGo. They all had three or more consecutive HA before 12 weeks, unexplained by uterine malformation, hormonal and infection causes or karyotype abnormalities of one or both partners. The mean age was 33.7 ± 5.5 years; the mean number of previous losses was 3.9 ± 1.5 per patient. There were no previous live births for 496 patients (73.2%) and 182 women (26.8%) had had one live birth before their HA; 36 women (5.3%) had experienced intrauterine fetal death and 31 women (4.5%) had extra-uterine pregnancies.


View this table:
[in this window]
[in a new window]
 
Table I. Baseline characteristics of the patients. Values in parentheses are percentages
 
The patients were instructed to test for ß-human chorionic gonadotrophin (ß-HCG), using a kit from BioMérieux (Marcy l'Etoile, France) on the first day of their missed period or two days before if possible (for instance by carrying out the test every cycle on the 26th day for a 28 day cycle), and to start taking prednisone and aspirin or aspirin alone. Among these patients, 312 became pregnant and 35 were excluded: 21 of them started their treatment too late (at 6–9 weeks), two had an extra-uterine pregnancy, one had a fetal loss due to cervix incompetence at 25 weeks, one had an induced abortion because the fetal karyotype was XYY, one because of a fetal death after amniocentesis, nine had prednisone before conception because the previous spontaneous abortions occurred at <6 weeks amenorrhoea. The results of the abortion prevention were analysed in the remaining 277 women, among whom the first 214 (161 autoantibody-negative women and 53 autoantibody-positive women) received prednisone (Roussel, Paris, France) 20 mg/day up to 12 weeks, plus aspirin (100 mg per day) until the end of the seventh month. The other 63 antiphospholipid (aPL)-negative pregnant women were given only low-dose aspirin (100 mg per day) as a control group for prednisone. Each patient gave informed consent for the study, which was approved by the local Ethics Committee.

Biological investigations
All women were investigated for the presence of autoantibodies:

Immunoglobulins (Ig)
IgG and IgM aCL, IgG and IgM anti-ß2 glycoprotein 1 (anti-ß2GP1) and antiprothrombin antibodies (aPt) were detected by an enzyme-linked immunosorbent assay, as described previously (Abuaf et al., 1997Go). Briefly, 96-well microtitre plates (Maxisorp Nunc, Roskilde, Denmark) were coated with 30 ml of cardiolipin in ethanol at a concentration of 50 mg/ml (Sigma C-1649; Saint-Quentin Fallavier, France) or 120 ml of ß2GP1 at 15 mg/ml isolated from pooled normal human plasma, or 120 ml of prothrombin at 5 mg/ml (Diagnostic Stago, Asnières, France). Dilutions of a reference standard serum and known negative sera were used on every plate. Calibration sera for {gamma}G antiphospholipid (GPL) (IgG aPL) or {gamma}M antiphospholipid (MPL) (IgM aPL) units were obtained from Dr E.N.Harris, (Department of Medicine, University of Louisville, KY 40292, USA). A patient's serum with a titre of anti-ß2G or aPt was used as a standard serum to define arbitrary units; the threshold at 95 percentile was established with 130 blood donors. The patients were assigned to negative group: aCL <15 uGPL or MPL, anti-ß2G <6 uß2GP or MP and aPrt <60 uGP.

Lupus anticoagulant (LA)
Diagnosis was made on the basis of: (i) an abnormal tissue thromboplastin test using Innovin (Ortho Diagnostic, Issy les Moulineaux, France) diluted 1/100 as reagent. The test was considered to be positive when the clotting time ratio to normal pool plasma was >1.10; (ii) an abnormal activated partial thromboplastin time minus the altered incubation time performed as previously described (Robert, 1994Go): to be positive the difference in clotting times should be >9 s; and (iii) a positive plasma 1:1 mixing study with normal pooled plasma. This procedure for LA diagnosis met the requirement of the International Society for Thrombosis and Haemostasis (Brandt et al., 1992).

Antinuclear antibodies (ANA)
These were detected by indirect immunofluorescence on the Hep-2 cell line (larynx carcinoma cell line), a titre >1/100 was considered to be positive. Native anti-DNA antibody determination was performed by the Farr test (Amersham, Bucks, UK) with a threshold of 7 IU.

Antithyroid antibodies (aTh)
These were detected by determining the antimicrosomal and antithyroglobulin antibodies. Antimicrosomal antibodies were detected by immunofluorescence on human thyroid sections; an antibody titre >1/10 was considered as positive. Antithyroglobulin antibodies were assessed by passive haemagglutination with a threshold of 1/640.

Statistical analysis
Variables between series were compared using the {chi}2 test or by Fisher's exact test for low values. The comparison of means was performed using Student's t-test. P < 0.05 was considered to be statistically significant.


    Results
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
Autoantibodies were found to be positive in 230 patients out of 678 HA (33.9%) (Table IIGo); 66.1% of the patients were autoantibody-negative. The most frequent autoantibodies were the aCL which were found in 82.6% (190 out of 230). IgG aCL were present in almost three out of four cases and IgM aCL in about 35%. LA were observed in only 2% out of all the patients and mostly associated with aCL. 79% of the autoantibodies were found alone and 21% were found to be associated with one or two other autoantibodies (Table IIIGo).


View this table:
[in this window]
[in a new window]
 
Table II. Autoantibodies in 678 patients with habitual abortionsa
 

View this table:
[in this window]
[in a new window]
 
Table III. Association of autoantibodies in women with habitual abortions. Values in parentheses are percentagesa
 
In autoantibody-negative women, the rate of live births in the group of women treated with prednisone in the first trimester and aspirin until the end of the seventh month was significantly increased compared with women treated with aspirin alone: 90.7% of women (146 out of 161) treated with prednisone and aspirin had live births compared with 74.5% of women (47 out of 63) treated with aspirin alone (P < 0.01). In autoantibody-positive women (45 out of 53) the rate of live births was 84.9%; there was no significant difference in the rate of live births between autoantibody-negative and autoantibody-positive women (P > 0.5). The success rate was much lower when aCL was >30 uGPL or MPL (11 out of 53 positive women) with seven ongoing pregnancies (63%) and four abortions, compared with 38 successful pregnancies in 42 women with aCL <=30 uGPL or MPL (90%) (P > 0.05). In sero-negative women, aCL antibodies became positive in nine women out of 63 (14.3%) who received aspirin treatment only and in 19 out of the 161 women (11.8%) who received prednisone and aspirin treatment (P > 0.9).

The mean birth weights in women who received prednisone and aspirin or aspirin alone was similar in the two groups, 3083 ± 564.5 g and 2956 ± 431.5 g, as was the mean term delivery which was respectively 39 ± 2.0 and 38. 8 ± 1.6 weeks (not significant; Student's t-test). In the group of autoantibody-negative women treated with prednisone and aspirin, three cases of hypertension (2%) were observed and one of gestational diabetes (0.6%), and hypertension was observed in one patient (2%) with autoantibody-positive women, which is in accordance with the general population.


    Discussion
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
The proportions of autoantibodies in our population of healthy HA patients was 33.9%, in which aCL were present in the majority of these cases (82.6%) or 28% of all the HA. This percentage is in accordance with findings previously published (Rai et al., 1995Go; Stephenson, 1996Go; Geva et al., 1998Go). IgG aCL were twice as frequent as IgM. IgG anti-ß2G were present in <1% of the cases and always associated with other aPL. They probably have little, if any, implication in the pathogenesis of HA and this is in accordance with an earlier study (Maejima et al., 1997Go). In the same way antithyroid antibodies were observed in only 1.3% of the patients. LA were found in very few cases and generally associated with aCL. Elevated ANA with anti-DNA antibody were observed in only 0.02% of patients. These results differ very much from those of a recent study (Laskin et al., 1997Go). In their cohort of patients with two or more habitual abortions and almost 50% autoantibodies, they observed 40% ANA with anti-DNA antibodies in 18%, but only 6–14% of aCL. These observations may not be due to the differences in techniques but probably in the way patients were recruited.

Prednisone and aspirin treatment was superior to aspirin treatment alone in autoantibody-negative women: the success rate was 90.7%, compared with 74.5% in patients with aspirin alone (P < 0.01). Previously, the outcome of pregnancies was followed in a similar cohort of 102 HA patients without treatment (Reznikoff-Etievant, 1991Go) and 58% resulted in births. This observation would mean that low-dose aspirin could foster pregnancy in autoantibody-negative patients. However, in a recent study (Tuppala et al., 1997) no difference was found in the success rate in HA women treated with either aspirin (70%) or placebo (70%).

Prednisone and aspirin treatment have the same success rate in both autoantibody-negative and autoantibody-positive women. The results presented here confirm those of a recent study (Geva et al., 1998Go). However, in another study (Laskin et al., 1997Go), a double-blind trial in patients with two or more unexplained abortions with low autoantibodies, prednisone until delivery associated with aspirin was not significantly different from placebo: 65% live infants versus 56% with placebo (P = 0.19). The success rate in this work was much lower than in our study. There are two possible reasons: firstly, our work is not a randomized controlled trial and so the results may be overestimated. Secondly, the autoimmune profile is different in the two studies discussed above, and the same treament may not have the same efficacy in these two cohorts. In this study, prednisone does not seem to act by preventing the seroconversion of aCL as no difference was found between the seroconversion rate in the two treatment groups, aspirin alone or aspirin and prednisone. This would mean that prednisone may interfere with factors which were not investigated in this work, such as an embryotoxic factor (Hill et al., 1992Go) or factors associated with human leukocyte antigens (HLA) (Christiansen, 1996Go). As observed in women with aCL >30 uGPL, prednisone plus aspirin appeared to be less successful than in a group of aCL <=30 uGPL, it is suggested that in the former group, after prednisone is stopped, heparin should be considered if the aCL titre rises. With lupus anticoagulant, heparin should be the treatment of choice as heparin is superior to low-dose aspirin (Kutteh et al., 1996; Di Simone et al., 1997Go).

This series of 214 treated patients established that prednisone and aspirin treatment is safe, in accordance with previous studies (Geva et al., 1998Go). This contrasts with other studies (Laskin et al., 1997Go) which described adverse side-effects (e.g. hypertension, diabetes mellitus and prematurity) in the prednisone and aspirin-treated group. This contrast may be due to the difference in the duration of prednisone treatment which was 2 months in this study, 3.5 months in Geva's study and throughout pregnancy for Laskin.

After this preliminary study, we began a double-blind trial in women with or without very low aPL; this study is in progress.


View this table:
[in this window]
[in a new window]
 
Table IV. Treatment with prednisone in the first trimester and or low-dose aspirin until the end of the seventh month, in women with habitual abortion
 

    Notes
 
4 To whom correspondence should be addressed Back


    References
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
Abuaf, N., Laperche, S., Rajoely, B. et al. (1997) Autoantibodies to phospholipids and to the coagulation proteins in AIDS. Thromb. Haemost., 77, 856–861.[ISI][Medline]

Brandt, J.T., Triplett, D., Alving, B. et al. (1995) Criteria for the diagnosis of lupus anticoagulant: an update. Thromb. Haemost., 74, 1185–1190.[ISI][Medline]

Christiansen, O.B. (1996) A fresh look at the causes and treatments of recurrent miscarriage, especially its immunological aspects. Hum. Reprod. Update, 2, 271–293.[Abstract]

Di Simone, N., Ferrazzani, S., Castellani, R. et al. (1997) Heparin and low-dose aspirin restore placental human chorionic gonadotrophin secretion abolished by antiphospholipid antibody-containing sera. Hum. Reprod., 12, 2061–2065.[Abstract]

German RSA/IVIG Group (1994) Intravenous immunoglobulin in the prevention of recurrent miscarriage. Br. J. Obstet. Gynecol., 101, 1072–1077.[ISI][Medline]

Geva, E., Amit, A., Lerner-Geva, L., et al. (1998) Prevention of pregnancy loss in autoantibody seropositive women. Lancet, 351, 34–35.

Gleicher, N. (1998) Autoantibodies in infertility: current opinion. Hum. Reprod. Update, 4, 169–176.[Free Full Text]

Hill, J.A., Polgar, K., Harlow, B.L. et al. (1992) Evidence of embryo and trophoblastotoxic cellular immune response(s) in women with recurrent spontaneous abortion. Am. J. Obstet. Gynecol., 166, 1044–1052.[ISI][Medline]

Kutteh, W.H. (1996) Antiphospholipid antibody associated recurrent pregnancy loss: treatment with heparin and low-dose aspirin is superior to low-dose aspirin alone. Am. J. Obstet. Gynecol., 174, 1584–1589.[ISI][Medline]

Laskin, C.A., Bombardier, C., Hannah, M.E. et al. (1997) Prednisone and aspirin in women with autoantibodies and unexplained recurrent abortion. N. Engl. J. Med., 337, 148–153.[Abstract/Free Full Text]

Maejima, M., Fujii, T., Okai, T. et al. (1997) Beta 2-glycoprotein I-dependent anticardiolipin antibody in early recurrent spontaneous abortion. Hum. Reprod., 12, 2140–2142.[Abstract]

Rai, R.S., Regan, L., Clifford, K. et al. (1995) Antiphospholipid antibodies and beta 2 glycoprotein-1 in 500 women with miscarriage: results of a comprehensive screening approach. Hum. Reprod., 10, 2001–2005.[Abstract]

Recurrent Miscarriage Immunotherapy Trialists Group (1994) World collaborative observational study and meta-analysis on allogenic leucocyte immunotherapy for recurrent spontaneous abortion. Am. J. Immunol., 33, 55–72.

Reznikoff-Etievant, M.F. (1991) Results of 15 centers in France. In Chaouat, G. and Mowbray, J. (eds), Cellular and Biology of the Materno–Fetal Relationship. Colloque ISERM/John Libbey, 212, pp. 303–306.

Robert, A. (1994) Two different incubation times for the activated partial thromboplastin time (APTT): a new criterion for diagnosis of lupus anticoagulant. Thromb. Haemost., 71, 220–224.[ISI][Medline]

Stephenson, M.D. (1996) Frequency of factors associated with habitual abortion in 197 couples. Fertil. Steril., 66, 24–29.[ISI][Medline]

Stirrat, G.M. (1990) Recurrent miscarriage II: Clinical associations, causes, and management. Lancet, 336, 728–733.[ISI][Medline]

Tulppala, M., Marttunen, M., Söderström-Anttila, V. et al. (1997) Low-dose aspirin in prevention of miscarriage in women with unexplained or autoimmune related recurrent miscarriage: effect on prostacyclin and thromboxane A2 production. Hum. Reprod., 12, 1567–1572.[Abstract]

Yamada, H., Kishida, T., Kobayashi, N. et al. (1998) Massive immunoglobulin treatment in women with four or more recurrent spontaneous primary abortions of unexplained aetiology. Hum. Reprod., 13, 2620–2623.[Abstract]

Submitted on December 29, 1998; accepted on April 22, 1999.