Frias, Antonio E. Jr MD*; Luikenaar, Rixt A. MD*; Sullivan, Amy E. MD*; Lee, Richard M. MD†; Porter, T Flint MD*; Branch, D Ware MD*; Silver, Robert M. MD*
First-trimester spontaneous abortion is one of the most common obstetric complications, occurring in 12–15% of clinically recognized pregnancies.1,2 Approximately 1% of couples suffer recurrent spontaneous abortion, which may be defined as 3 or more consecutive losses.3 This group of patients has been the subject of intensive investigation. However, relatively little is known about second-trimester pregnancy loss. Thus, it is difficult for clinicians to optimally evaluate, counsel, and manage patients with previous unexplained fetal death. The objectives of our study were to report the outcome of subsequent pregnancies in women referred for evaluation of fetal death and identify risk factors that may influence ensuing pregnancy loss.
MATERIALS AND METHODS
Subjects were identified from patients referred to the University of Utah for evaluation of pregnancy loss, which included at least one fetal death, and who had at least one subsequent pregnancy between 1998 and 2002. Fetal death was defined as the intrauterine demise of a conceptus known to be alive at or beyond 10 completed weeks of gestation documented by 1) detection of fetal heart tones, 2) the ultrasonographic detection of a conceptus with biometric measurements indicating a gestational age at or beyond 10 weeks of gestation, or 3) the delivery of a dead conceptus whose size indicated a gestational age at or beyond 10 weeks of gestation.4,5 All pregnancy losses before 10 weeks of gestation were considered spontaneous abortions of either an anembryonic or embryonic nature. The pregnancy immediately following each patient's first fetal death was considered to be the index pregnancy for this study. This was chosen to allow for analysis of only one pregnancy per subject and to minimize potential bias.
This retrospective cohort study was approved by the University of Utah Institutional Review Board. Medical and obstetric histories were obtained by review of the University of Utah Recurrent Pregnancy Loss database and the medical records. Patient entry into the University of Utah Recurrent Pregnancy Loss database began prospectively in 2001 after approval from the Institutional Review Board. Before that date, information was entered in the database by retrospective chart review. Data were ascertained regarding prior pregnancies, medical problems, and diagnostic evaluations of either individual or recurrent fetal losses. Testing for recognized causes of pregnancy loss and clinical management was at the discretion of the primary physicians.
All patients in this cohort were tested for lupus anticoagulant and anticardiolipin antibodies. Lupus anticoagulant was detected by using an activated partial thromboplastin time and a dilute Russel viper venom time and by mixing studies as previously described (Branch DW, Dudley DJ, Scott JR, Silver RM. Antiphospholipid antibodies and fetal loss [letter]. N Engl J Med 1992;326:952; author reply 953-4).6 Anticardiolipin antibodies were identified by means of an immunoassay that was calibrated by using serum standards from the Antiphospholipid Standardization Laboratory (Morehouse University, Atlanta, GA) (Branch et al. N Engl J Med).7 Patients with either lupus anticoagulant, medium- to high-positive levels of immunoglobulin G (IgG) anticardiolipin antibodies, or both, were considered to have antiphospholipid syndrome.8–10 These individuals were excluded from the current study to avoid undue bias, because the outcome of subsequent pregnancies in untreated patients with antiphospholipid syndrome is particularly poor.11
Some patients underwent testing for other recognized causes of recurrent pregnancy loss as outlined in the Results section. A number of patients, who were treated with a variety of medical therapies at the discretion of their physicians, including low-dose aspirin (81 mg/d), thromboprophylactic doses of heparin (10,000–15,000 U/d), or both, intended to improve pregnancy outcome. Treatment was empiric and occurred at the discretion of the primary physician.
Data were analyzed with SAS/STAT 8.2 (SAS Institute Inc, Cary, NC). The logistic regression model using the generalized estimating equation method was used to test the association of various factors on the outcome of unsuccessful pregnancy versus live birth (Table 1). The results of the single independent variable analysis identified the variables included as independent variables in the multifactor logistic regression model using the generalized estimating equation method.
Two-factor interactions were also considered for inclusion in the multivariable model. Data were analyzed for all subsequent pregnancies in the cohort, as well as for the first pregnancy after the first fetal death, which was considered the index pregnancy. This latter analysis was performed so as to include only one pregnancy outcome for each patient.
Two-hundred thirty white women were identified, who had a prior fetal death, at least one subsequent pregnancy with known outcome, and negative testing for antiphospholipid antibodies. Subsequent pregnancy outcomes for these women are shown in Table 2. The rate of recurrent pregnancy loss was high, with fewer than 25% resulting in live births.
The first fetal death occurred during the first pregnancy in 20.9% (48/230), the second pregnancy in 19.6% (45/230), the third pregnancy in 22.5% (51/230), and the fourth or later pregnancy in 37.5% (86/230) of the subjects. The mean age during the first pregnancy following the first fetal death was 29.3 years ± 5.5 (standard deviation), with a range of 14.4 to 43.2. After the first fetal death, the median number of subsequent pregnancies was 3, with an interquartile range of 2 to 5 (range 1–18).
The gestational ages at time of fetal death are shown in Fig. 1. Over two thirds occurred during the second trimester, with the mode between 16 and 18 weeks of gestation. The majority (174/230; 76%) of first fetal deaths also occurred in the second trimester, with the mode between 10 and 12 weeks of gestation.
We next sought to examine the association of patient characteristics with pregnancy. Table 1 shows the odds ratio estimate for each variable that was tested for association with the outcome of unsuccessful pregnancy in a univariate logistic regression model. Surprisingly, previous live births were not associated with improved outcome in subsequent pregnancies, and increasing numbers of spontaneous abortions were not associated with worse outcome. Older women were less likely to suffer pregnancy loss than their younger counterparts (P = .009). Thyroid disease was not associated with subsequent pregnancy loss, although patients were treated for their disorder. There were too few patients with other medical conditions to evaluate their effects on pregnancy outcome.
Table 3 shows the outcome of treated pregnancies after the first fetal demise. Patients were treated at the discretion of their providers without randomization or controls. Patients treated with low-dose aspirin had improved outcomes compared with those who were not treated, with a 39% live birth rate. Univariate analysis (Table 4) demonstrated that women treated with low-dose aspirin had an odds ratio (OR) of 0.41 (95% confidence interval [CI] 0.25–0.68) for subsequent pregnancy loss (P = .001). The favorable effect of low-dose aspirin persisted whether patients were treated with low-dose aspirin alone or in combination with other therapies. Pregnancy outcome was not significantly improved after treatment with progesterone, heparin, or combinations of these regimens.
A multivariable logistic regression model was used to assess the independent effect of variables significantly associated with pregnancy outcome in the univariate analysis. These variables included age at pregnancy and treatment with low-dose aspirin. In the multivariable logistic regression model, low-dose aspirin was associated with a reduction in the risk for unsuccessful pregnancy only in women 35 years of age or older (OR 0.12, 95% CI 0.05–0.32).
In a majority of cases, the evaluation for potential causes of fetal death was incomplete. One hundred forty-two patients were evaluated by hysterosalpingography. Seventeen had abnormalities, including 8 müllerian anomalies, 6 leiomyomata or polyps, and 3 synechiae. Abortus karyotype was obtained in 100 patients, with 20 abnormal results (3 45 XO, 3 trisomy 21, 3 triploidy, 10 other trisomy, and 1 mosaic). Eight of 88 (10%) fetuses that underwent autopsy had congenital malformations and/or evidence of genetic syndromes. Maternal and paternal karyotype was evaluated in 102 and 93 individuals, respectively. There were 2 abnormal maternal karyotypes, including a 45 XO mosaic and a balanced translocation. There were 2 abnormal paternal karyotypes (1 balanced translocation, 1 deletion). As previously stated, all subjects tested negative for antiphospholipid antibodies. Other thrombophilias were not assessed in a systematic fashion. One of 34 patients tested was heterozygous for the 20210 prothrombin gene mutation. Fifty-four women were tested for the factor V Leiden mutation. Three were heterozygous for the mutation.
Women with recurrent pregnancy loss and fetal death have an extremely poor prognosis for subsequent pregnancies, with fewer than 25% resulting in surviving infants. The risk of recurrent loss in these patients is higher than we anticipated and is substantially worse than the 60–70% rate of successful pregnancies in women with recurrent first-trimester loss.2,12–14 The risk of subsequent pregnancy loss in this cohort was also higher than for patients suffering a sporadic second-trimester loss (11% rate of pregnancy loss).15
Although a majority of subsequent pregnancies resulted in spontaneous abortion, almost one third resulted in fetal death. This observation suggests that the mechanism of pregnancy loss may be different in these couples than in those with recurrent first-trimester loss. Indeed, there is accumulating evidence that the etiologies of fetal deaths are different from those responsible for early pregnancy failures.5 For example, antiphospholipid syndrome and other thrombophilias are more common in women with fetal death than in those with recurrent spontaneous abortion.4,16 In contrast, aneuploidy is much more prevalent in first-trimester losses.17,18
In contrast to other studies,19 we did not find increased maternal age to be associated with an increased risk of pregnancy loss. This may be due to the fact that maternal age increases the risk for aneuploidy and, thus, the risk for spontaneous abortion as opposed to fetal death. Nonetheless, Fretts and coworkers20 noted advancing maternal age to be a risk factor for deaths occurring in fetuses weighing more than 500 g. This observation suggests that the causes of the often recurrent second-trimester losses seen in our population may differ from those responsible for sporadic late fetal deaths. In addition, our cohort was likely affected by the selection bias inherent in a referral population. We speculate that older women were referred after fewer losses (compared with younger women) because of increased anxiety. Also, older women may have had a longer interval between pregnancies (than younger women), accounting for the apparent protective effect of maternal age. The interval between pregnancies was not ascertained in this cohort.
It was also surprising to find no association between either previous live births or the number of spontaneous abortions and subsequent pregnancy outcome. These data also differ from data obtained in cohorts of women with recurrent first-trimester losses19 and underscore the difference between our cohort of women with recurrent fetal death and those with recurrent first-trimester losses. Another possible explanation for these findings was an increased use of empiric treatment during the index pregnancy in women with worse obstetric histories (more losses and fewer live births).
We also did not find underlying medical conditions such as diabetes, hypertension, or systemic lupus erythematous to be associated with fetal death. This was likely due to the small number of patients with these conditions in our study and the exclusion of women with antiphospholipid antibodies.
Unfortunately, most of our patients had incomplete evaluations for possible causes of fetal death. Fewer than 3% of women underwent a comprehensive evaluation, and over half had no investigation into possible etiologies of recurrent pregnancy loss or fetal death. Accordingly, a cause of fetal death was identified in only 9% of cases. In contrast, Ahlenius and colleagues21 identified a potential cause of death in 88% of cases of late second- and third-trimester stillbirths using an extensive and consistent protocol. The most common causes of death in their series included infections, preterm premature rupture of membranes, fetal malformations, and preeclampsia. Some of these etiologies are less likely to cause fetal death before 20 weeks of gestation.
Pregnancy outcomes were dramatically improved in women receiving medical therapy with low-dose aspirin. Our retrospective and uncontrolled study design does not permit us to make any conclusions regarding the efficacy of low-dose aspirin in patients with previous unexplained fetal death. Nonetheless, our data support the merit of determining treatment efficacy in randomized controlled trials. It is uncertain why treatment with low-dose aspirin may improve outcome in patients with unexplained fetal death, but one possibility is that it improves the utero-placental circulation. Conditions that have been associated with placental thrombosis and insufficiency, such as antiphospholipid syndrome and heritable thrombophilias, are specifically associated with fetal death.4,11,16,22–24 Moreover, anticoagulant therapy improves outcome in pregnancies complicated by antiphospholipid syndrome (Branch et al. N Engl J Med).25,26 It is possible that some of our patients have unrecognized heritable thrombophilias. This cohort was not routinely tested for these conditions because the relationship between thrombophilia and fetal death has only recently become apparent and is still controversial (Dizon-Townson D, for the NICHD MFMU Network. Factor V Leiden mutation does not increase risk of pregnancy-related venous thromboembolism [abstract]. Am J Obstet Gynecol 2002;187:S159)23,27,28 Also, placental insufficiency is more common in older women, providing a potential reason for the observation that aspirin was only effective in this population.
In summary, women with recurrent pregnancy loss and fetal death are at high risk for subsequent pregnancy loss, with less than 25% of subsequent pregnancies resulting in surviving infants. Traditional risk factors for recurrent spontaneous abortion do not appear to influence pregnancy outcome in this group of patients. Our population included a referral population with primarily second-trimester losses and recurrent losses. Therefore, our results may be biased and not generalizable. Nevertheless, these data underscore the need for additional research into the pathophysiology and prevention of recurrent fetal death.
1. Warburton D, Fraser FC. Spontaneous abortion risks in man: data from reproductive histories collected in a medical genetics unit. Am J Hum Genet 1964;16:1–25.
2. Poland BJ, Miller JR, Jones DC, Trimble BK. Reproductive counseling in patients who have had a spontaneous abortion. Am J Obstet Gynecol 1977;127:685–91.
3. Lee RM, Silver RM. Recurrent pregnancy loss: summary and clinical recommendations. Semin Reprod Med 2000;18:433–40.
4. Oshiro BT, Silver RM, Scott JR, Yu H, Branch DW. Antiphospholipid antibodies and fetal death. Obstet Gynecol 1996;87:489–93.
5. Branch DW, Silver RM. Criteria for antiphospholipid syndrome: early pregnancy loss, fetal loss, or recurrent pregnancy loss? Lupus 1996;5:409–13.
6. Martin BA, Branch DW, Rodgers GM. The preparation of a sensitive partial thromboplastin reagent from bovine brain. Blood Coagul Fibrinolysis 1992;3:287–94.
7. Harris EN. Special report. The Second International Anti-cardiolipin Standardization Workshop/the Kingston Anti-Phospholipid Antibody Study (KAPS) group. Am J Clin Pathol 1990;94:476–84.
8. Wilson WA, Gharavi AE, Piette JC. International classification criteria for antiphospholipid syndrome: synopsis of a post-conference workshop held at the Ninth International (Tours) aPL Symposium. Lupus 2001;10:457–60.
9. Silver RM, Porter TF, van Leeuween I, Jeng G, Scott JR, Branch DW. Anticardiolipin antibodies: clinical consequences of “low titers”. Obstet Gynecol 1996;87:494–500.
10. Harris EN, Chan JK, Asherson RA, Aber VR, Gharavi AE, Hughes GR. Thrombosis, recurrent fetal loss, and thrombocytopenia: predictive value of the anticardiolipin antibody test. Arch Intern Med 1986;146:2153–6.
11. Rai RS, Clifford K, Cohen H, Regan L. High prospective fetal loss rate in untreated pregnancies of women with recurrent miscarriage and antiphospholipid antibodies. Hum Reprod 1995;10:3301–4.
12. Warburton D, Fraser FC. On the probability that a woman who has had a spontaneous abortion will abort in subsequent pregnancies. J Obstet Gynaecol Br Emp 1961;68:784–8.
13. Regan L. Recurrent miscarriage. BMJ 1991;302:543–4.
14. Regan L, Braude PR, Trembath PL. Influence of past reproductive performance on risk of spontaneous abortion. BMJ 1989;299:541–5.
15. Goldenberg RL, Mayberry SK, Copper RL, Dubard MB, Hauth JC. Pregnancy outcome following a second-trimester loss. Obstet Gynecol 1993;81:444–6.
16. Preston FE, Rosendaal FR, Walker ID, Briet E, Berntorp E, Conard J, et al. Increased fetal loss in women with heritable thrombophilia. Lancet 1996;348:913–6.
17. Boue J, Bou A, Lazar P. Retrospective and prospective epidemiological studies of 1500 karyotyped spontaneous human abortions. Teratology 1975;12:11–26.
18. Warburton D. Chromosomal causes of fetal death. Clin Obstet Gynecol 1987;30:268–77.
19. Coulam CB, Clark DA, Beer AE, Kutteh WH, Silver R, Kwak J, et al. Current clinical options for diagnosis and treatment of recurrent spontaneous abortion. Clinical Guidelines Recommendation Committee for Diagnosis and Treatment of Recurrent Spontaneous Abortion. Am J Reprod Immunol 1997;38:57–74.
20. Fretts RC, Schmittdiel J, McLean FH, Usher RH, Goldman MB. Increased maternal age and the risk of fetal death. N Engl J Med 1995;333:953–7.
21. Ahlenius I, Floberg J, Thomassen P. Sixty-six cases of intrauterine fetal death: a prospective study with an extensive test protocol. Acta Obstet Gynecol Scand 1995;74:109–17.
22. Rai R, Regan L, Hadley E, Dave M, Cohen H. Second-trimester pregnancy loss is associated with activated C resistance. Br J Haematol 1996;92:489–90.
23. Kupferminc MJ, Eldor A, Steinman N, Many A, Bar-Am A, Jaffa A, et al. Increased frequency of genetic thrombophilia in women with complications of pregnancy [published erratum appears in N Engl J Med 1999;341:384]. N Engl J Med 1999;340:9–13.
24. Martinelli I, Taioli E, Cetin I, Marinoni A, Gerosa S, Villa MV, et al. Mutations in coagulation factors in women with unexplained late fetal loss. N Engl J Med 2000;343:1015–8.
25. Branch DW, Silver RM, Blackwell JL, Reading JC, Scott JR. Outcome of treated pregnancies in women with antiphospholipid syndrome: an update of the Utah experience. Obstet Gynecol 1992;80:614–20.
26. Rai R, Cohen H, Dave M, Regan L. Randomised controlled trial of aspirin and aspirin plus heparin in pregnant women with recurrent miscarriage associated with phospholipid antibodies (or antiphospholipid antibodies). BMJ 1997;314:253–7.
27. Saade GR, McLintock C. Inherited thrombophilia and stillbirth. Semin Perinatol 2002;26:51–69.
28. Lindqvist PG, Svensson PJ, Marsaal K, Grennert L, Luterkort M, Dahlback B. Activated protein C resistance (FV:Q506) and pregnancy. Thromb Haemost 1999;81:532–7.