Objective: To examine the effectiveness of aspirin in preventing perinatal death and preeclampsia in women with predisposing historical risk factors, such as previous history of preeclampsia, chronic hypertension, diabetes, and renal disease.

Data Sources: Searches were conducted in MEDLINE, EMBASE, Cochrane Library, National Research Register, SCISEARCH, and ISI Conference Proceedings without any language restriction, using the following medical subject headings and text words: "aspirin," "antiplatelet*," "salicyl*," "acetylsalicyl*," "platelet aggregation inhibitors," "pre-eclamp*," "preeclamp*," and "hypertens*."

Methods of Study Selection: We included all randomized trials that evaluated the effectiveness of aspirin compared with placebo or no treatment in women with predisposing historical risk factors and reported clinically relevant perinatal or maternal outcomes. Study selection, quality appraisal, and data extractions were performed independently and in duplicate.

Tabulation, Integration, and Results: We identified 14 relevant trials, including a total of 12,416 women. Meta-analysis showed a significant benefit of aspirin therapy in reducing perinatal death (odds ratio [OR] 0.79, 95% confidence interval [CI] 0.64, 0.96) and preeclampsia (OR 0.86, 95% CI 0.76, 0.96). Aspirin was also associated with a reduction in rates of spontaneous preterm birth (OR 0.86, 95% CI 0.79, 0.94), and an increase of 215 g in mean birth weight (weighted mean difference 215, 95% CI 90, 341). There was no increase in the risk of placental abruption with aspirin (OR 0.98, 95% CI 0.79, 1.21). Funnel plot analysis indicated that publication and related biases were unlikely (Egger test, P = .84).

Conclusion: Aspirin reduces the risk of perinatal death and preeclampsia in women with historical risk factors. Given the importance of these outcomes and the safety and low cost of aspirin, aspirin therapy should be considered in women with historical risk factors.

^aEducation Resource Centre, Birmingham Women's Hospital, Birmingham, United Kingdom

All contributors to this review, except HH, are employees of the National Health Service of United Kingdom. HH holds a WellBeing Research Fund from the Royal College of Obstetricians and Gynaecologists (United Kingdom). No private or commercial funding was obtained.

AC and KSK conceived the review. AC, HH, and SP collected, analyzed, and interpreted the data and drafted the manuscript. KSK and HG made critical revisions. AC and HH are the guarantors for this article.

(Obstet Gynecol 2003:101:1319-1332. © 2003 by The American College of Obstetricians and Gynecologists.)

Readily obtainable historical information, such as previous history of preeclampsia,^1-4 history of chronic hypertension,^3,5,6 diabetes,^3,7 renal disease,^8,9 and family history¹⁰ of preeclampsia may be used to identify groups of women at high risk for preeclampsia. Such an inexpensive and practical risk identification process, which can be done at the primary care level, assumes significant importance if there were an effective therapy to reduce the risk of preeclampsia and its related complications.

There are several randomized trials that assess the effectiveness of aspirin in women with historical risk factors.^11-19 However, they generally show a statistically nonsignificant inclination toward benefit, which may be owing to lack of power. A meta-analysis of this group of studies should allow us to generate more precise estimates of effect. We therefore conducted a systematic review and meta-analysis of randomized trials to assess the effectiveness of aspirin in women identified to be at risk from historical risk factors.

Sources

We searched MEDLINE (1966-2001), EMBASE (1980-2001), Cochrane Library (2001:3), National Research Register (2001:3), SCISEARCH (1974-2001), and conference proceedings (ISI Proceedings, 1990-2001) for relevant citations. A combination of medical subject headings and text words were used to generate two subsets of citations, one including studies of aspirin ("aspirin," "antiplatelet*," "salicyl*," "acetylsalicyl*," and "platelet aggregation inhibitors") and the other studies of preeclampsia ("pre-eclamp*," "preeclamp*," and "hypertens*"). These subsets were combined using "AND" to generate a subset of citations relevant to our research question. The reference lists of all known primary and review articles were examined to identify cited articles not captured by electronic searches. Articles frequently cited were used in the Science Citation Index to identify additional citations. We also made inquiries about unpublished studies to researchers investigating in this field. No language restrictions were placed in any of our searches.

Study Selection

Studies were selected if the target population was women with historical risk factors for preeclampsia, the therapeutic intervention was low-dose aspirin (any definition) compared with placebo or no drug treatment, and the studies were of randomized design. The historical risk factors included previous preeclampsia, chronic (preexisting) hypertension, diabetes, renal disease, and extremes of age at conception.^3,20,21 The main outcomes were perinatal death and preeclampsia (proteinuric hypertension). However, all studies that reported on any clinically relevant perinatal or maternal outcomes, such as preterm birth, birth weight, and antenatal bleeding or placental abruption, were also included.

Studies were selected in a two-stage process. First, the titles and abstracts from the electronic searches were scrutinized by two reviewers independently (AC and HH), and full manuscripts of all citations that were likely to meet the predefined selection criteria were obtained. Second, final inclusion or exclusion decisions were made on examination of the full manuscripts. In cases of duplicate publication, the most recent and complete versions were selected. The assessment of English-language manuscripts was performed independently by two reviewers (AC and HH) and other language manuscripts by people who had command of the language. Any disagreements about inclusion were resolved by consensus or arbitration by a third reviewer (KSK).

The selected studies were assessed for methodologic quality by using the components of study design that are related to internal validity.²² Information on the adequacy of randomization, concealment, control, blinding, intention to treat, and follow-up rates was sought.

Percentage agreement and κ statistic were used for analysis of agreement between the reviewers. A minimum acceptable agreement was set at a κ level of 0.7, as this represents substantial strength of agreement.²³ From each study outcome data were abstracted in 22 tables. Odds ratios (ORs) from individual studies were pooled using the Peto method.²⁴ However, as there is no clear consensus as to the most appropriate method for meta-analysis,²⁵ we examined the sensitivity of our results to variation in statistical approach by performing meta-analysis using fixed-effects²⁶ and random-effects²⁷ models. Weighted mean difference was calculated for continuous variables using means and standard deviations from individual studies. Heterogeneity of treatment effects was evaluated graphically using forest plot²⁷ and statistically using ² test.²⁸ Exploration of the causes of heterogeneity was planned using variation in features of the population, intervention, outcome, and study quality. To assess for publication bias, we performed a funnel plot analysis,²⁹ using the Egger test³⁰ to evaluate for asymmetry. All statistical analyzes were performed using Stata 7.0 statistical software (Stata Corp., College Station, TX).

Results

Figure 1 summarizes the process of literature identification and selection. Of the 729 citations identified, 47 were selected during the initial screening (agreement 99%; κ 0.91), and on examination of the full manuscripts of these 47, 12 articles,^{11-15,17-19,31-34} and two abstracts (Azar R, Turpin D. Effect of antiplatelet therapy in women at high risk for pregnancy induced hypertension [abstract]. Proceedings of 7th World Congress, International Society for the Study of Hypertension in Pregnancy, Perugia, Italy. Newcastle upon Tyne, United Kingdom: ISSHP, 1990:257) (August P, Helseth G, Edersheim TG, Hutson JM, Druzin M. Sustained release, low-dose aspirin ameliorates but does not prevent preeclampsia (PE) in a high risk population [abstract]. Proceedings of 9th International Congress, International Society for the Study of Hypertension in Pregnancy, Sydney, Australia. Newcastle upon Tyne, United Kingdom: ISSHP, 1994:352), including a total of 12,416 women, were identified to satisfy the selection criteria for our review (agreement 96%; κ 0.89). The quality and the main characteristics of the included studies are presented in Tables 1 and 2.

Figure 1. Study selection process for systematic review of aspirin to prevent preeclampsia in women with historical risk factors for preeclampsia. Coomarasamy. Aspirin in Historical Risk Factors. Obstet Gynecol 2003.

Pooling of results from these studies showed a significant benefit of aspirin therapy in reducing perinatal death (OR 0.79, 95% confidence interval [CI] 0.64, 0.96, Figure 2) and preeclampsia (OR 0.86, 95% CI 0.76, 0.96, Figure 3). The results were robust to alternative approaches to statistical pooling using either fixed- or random-effects models (results available from authors on request). The results were also found to be homogeneous (P = .56 for perinatal death and .54 for preeclampsia).

Figure 2. Meta-analysis of randomized trials evaluating the effectiveness of aspirin to prevent perinatal death in women with historical risk factors for preeclampsia. The results were not heterogeneous (²₁₂ = 10.61, P = .56). CI = confidence interval. Coomarasamy. Aspirin in Historical Risk Factors. Obstet Gynecol 2003.

Figure 3. Meta-analysis of randomized trials evaluating the effectiveness of aspirin to prevent preeclampsia in women with historical risk factors. The results were not heterogeneous (²₁₂ = 10.85, P = .54). CI = confidence interval. Coomarasamy. Aspirin in Historical Risk Factors. Obstet Gynecol 2003.

We also found a significant reduction in rates of spontaneous preterm births (OR 0.86, 95% CI 0.79, 0.94, Figure 4), and an increase of 215 g in mean birth weight with aspirin use (weighted mean difference 215, 95% CI 90, 341, Figure 5). There was no increase in the risk of abruptions with aspirin treatment (OR 0.98, 95% CI 0.79, 1.21, Figure 6). Although funnel plot analysis was limited because of the small number of studies, it indicated that publication and related biases were unlikely (Egger test, P = .84).

Figure 4. Meta-analysis of randomized trials evaluating the effectiveness of aspirin to prevent preterm birth in women with historical risk factors for preeclampsia. The results were not heterogeneous (²₆ = 1.54, P = .96). CI = confidence interval. Coomarasamy. Aspirin in Historical Risk Factors. Obstet Gynecol 2003.

Figure 5. Meta-analysis of randomized trials assessing neonatal birth weight in women treated with aspirin compared with placebo or no treatment among women with historical risk factors for preeclampsia. As the results were heterogeneous (²₇ = 40.85, P = .01), random-effects model was used for meta-analysis. CI = confidence interval. Coomarasamy. Aspirin in Historical Risk Factors. Obstet Gynecol 2003.

Figure 6. Meta-analysis of randomized trials evaluating rates of abruptions and other antepartum bleeding with aspirin compared with placebo or no treatment among women with historical risk factors for preeclampsia The results were not heterogeneous (²₆ = 4.00, P = .67). CI = confidence interval. Coomarasamy. Aspirin in Historical Risk Factors. Obstet Gynecol 2003.

Conclusion

Our systematic review shows that low-dose aspirin has a significant effect in reducing the rates of a number of clinically relevant outcomes, including perinatal deaths and preeclampsia, in women with historical risk factors for preeclampsia.

The validity of our findings depends on the methodologic rigor of our review and of the component primary studies. We used a prospective protocol and made a concerted effort to find all the evidence. Two independent reviewers assessed study quality and extracted data, and the agreement between the two reviewers was high. The studies we found were generally of good quality, and the results were homogeneous. The results were significant regardless of the statistical approach used for meta-analysis. In addition, there was no evidence of publication and related biases from funnel plot analysis. One potential source of bias in our review was the inclusion of studies without a placebo arm. To address this, we performed sensitivity analysis, excluding such studies from the analysis (data not shown), and found that the results were consistent with our inferences. We are, therefore, confident that our inferences are robust.

The debate about aspirin to prevent preeclampsia has been rekindled by a recent Cochrane review.^35,36 It showed a statistically significant reduction in preeclampsia (relative risk [RR] 0.85, 95% CI 0.78, 0.92) and perinatal death (RR 0.86, 95% CI 0.75, 0.98), but it concluded that the clinical significance of this finding was questionable, stating "as the reductions in risk are small to moderate, relatively large numbers of women will need to be treated to prevent a single outcome" (the number of women needed to be treated to prevent one case of preeclampsia was 100, and to prevent one case of perinatal death was 250). This is likely to be because studies including high-risk women who could benefit most from aspirin therapy were pooled together with those including low-risk women. In high-risk women, the baseline risk of preeclampsia and associated complications would be expected to be higher, thus resulting in higher absolute benefit (lower number needed to treat) from aspirin therapy. For example, a recent systematic review of studies exclusively of those women in whom high risk was identified by an abnormal uterine artery Doppler test showed that the average absolute benefit from aspirin therapy to prevent preeclampsia was substantial (number needed to treat was 16, 95% CI 8, 316).³⁷

Our review did not show any evidence of harm from aspirin therapy. This was also confirmed in the Cochrane review.^35,36 In particular, there was no increase in placental abruptions, fetal intraventricular hemorrhage, or other neonatal bleeding. However, as randomized trials and systematic reviews of randomized trials may involve patients studied for a short length of time, they are not likely to detect delayed adverse events.³⁸ We therefore supplemented the evidence on harm with a systematic search for large observational studies using the following words and their word variants in MEDLINE (1966-2001) and EMBASE (1988-2001) bibliographic databases: "aspirin [adverse effects]" combined with "pregnancy." Three large cohort studies^39-41 and one large case-control study⁴² were identified (Table 3). These observational studies, with more than 96,000 women between them, did not provide any evidence of teratogenicity or long-term adverse effects of aspirin use in pregnancy. The only evidence of harm came from a case-control study that showed evidence of increased risk of miscarriage when nonsteroidal anti-inflammatory drugs were started early in gestation, although this risk was not present if the drugs were started after 12 weeks.⁴¹ However, no adjustment was made in this study for the indication of nonsteroidal anti-inflammatory use, and this would be expected to introduce substantial bias. This is because women with conditions such as systemic lupus erythematosus and rheumatoid arthritis are more likely to be users of nonsteroidal anti-inflammatory drugs during pregnancy but have an inherently higher risk of miscarriage as a result of the underlying condition rather than the drug itself. Thus, we regard this evidence of association between nonsteroidal drugs and miscarriage as unreliable.

The baseline risks of perinatal death and preeclampsia among the randomized trials, represented by the control event rates, varied substantially between trials. In this situation, calculating average numbers needed to be treated from pooled meta-analysis results can be seriously misleading.⁴³ This is because numbers needed to be treated are sensitive to the levels of baseline risk, and will, therefore, need to be individualized to appropriate baseline risks.⁴⁴ Consequently, we have given a range of numbers needed to be treated appropriate for various baseline risks for the several outcomes, as examples, in Table 4. When multiple historical risk factors are involved, the numbers needed to be treated would be expected to be even lower. However, these numbers needed to be treated need to be interpreted with caution, as there may have been changes in baseline risks over time, and the CIs around the numbers needed to be treated do not incorporate the uncertainty around the base-line risks (as these were not reported in most prevalence studies).

Based on our findings and the established safety of aspirin, it seems reasonable to recommend aspirin therapy to women who are historically at high risk for preeclampsia, particularly those with multiple risk factors. Those who may benefit include women with previous severe or early-onset preeclampsia, chronic hypertension, severe diabetes, and moderate to severe renal disease (Table 4). Our findings assume particular significance as the diagnosis of historical risk factors for preeclampsia can be made inexpensively (without the need for any laboratory tests) and at the primary care level. This is of critical importance in the developing world, particularly Africa, where preeclampsia is more prevalent, and resources for sophisticated investigations and expensive treatment are not generally available. Further research needs to address the possible cost-benefits of this approach to health care systems and the society.

References

1. Sibai BM, Mercer B, Sarinoglu C. Severe preeclampsia in the second trimester: Recurrence risk and long-term prognosis. Am J Obstet Gynecol 1991;165:1408-12.

2. Sibai BM, el Nazer A, Gonzalez-Ruiz A. Severe preeclampsia-eclampsia in young primigravid women: Subsequent pregnancy outcome and remote prognosis. Am J Obstet Gynecol 1986;155:1011-6.

3. Caritis S, Sibai B, Hauth J, Lindheimer M, VanDorsten P, Klebanoff M, et al. Predictors of pre-eclampsia in women at high risk. Am J Obstet Gynecol 1998;179:946-51.

4. Hnat MD, Sibai BM, Caritis S, Hauth J, Lindheimer MD, MacPherson C, et al. Perinatal outcome in women with recurrent preeclampsia compared with women who develop preeclampsia as nulliparas. Am J Obstet Gynecol 2002;186:422-6.

5. Rey E, Couturier A. The prognosis of pregnancy in women with chronic hypertension. Am J Obstet Gynecol 1994;171:410-6.

6. Mabie WC, Pernoll ML, Biswas MK. Chronic hypertension in pregnancy. Obstet Gynecol 1986;67:197-205.

7. Garner PR, D'Alton ME, Dudley DK, Huard P, Hardie M. Preeclampsia in diabetic pregnancies. Am J Obstet Gynecol 1990;163:505-8.

8. Cunningham FG, Cox SM, Harstad TW, Mason RA, Pritchard JA. Chronic renal disease and pregnancy outcome. Am J Obstet Gynecol 1990;163:453-9.

9. Jones DC, Hayslett JP. Outcome of pregnancy in women with moderate or severe renal insufficiency. N Engl J Med 1996;335:226-32.

10. Mogren I, Hogberg U, Winkvist A, Stenlund H. Familial occurrence of preeclampsia. Epidemiology 1999;10:518-22.

11. Beaufils M, Uzan S, Donsimoni R, Colau JC. Prevention of pre-eclampsia by early antiplatelet therapy. Lancet 1985;1:840-2.

12. Sureau C. Prevention of perinatal consequences of pre-eclampsia with low-dose aspirin: Results of the Epreda trial. Eur J Obstet Gynecol Reprod Biol 1991;41:71-3.

13. Parazzini F, Benedetto C, Frusca T, Gregorini G, Bocciolone L, Marozio , et al. Low-dose aspirin in prevention and treatment of intrauterine growth retardation and pregnancy-induced hypertension. Lancet 1993;341:396-400.

14. Rogov VA, Tareeva IE, Sidorova IS, Androsova SO, Katamadze KT, Nikiforova OV. Acetylsalicylic acid and kurantil in the prevention of pregnancy complications in glomerulonephritis and hypertension [in Russian]. Ter Arkh 1993;65:65-8.

15. Viinikka L, Hartikainen-Sorri A-L, Lumme R, Hiilesmaa V, Ylikorkala O. Low dose aspirin in hypertensive pregnant women: Effect on pregnancy outcome and prostacyclin-thromboxane balance in mother and newborn. Br J Obstet Gynaecol 1993;100:809-15.

16. Goffinet F, Breart G, Uzan S, Atallah A. ECPPA: Randomised trial of low dose aspirin for the prevention of maternal and fetal complications in high risk pregnant women. Br J Obstet Gynaecol 1996;103:719-20.

17. Gallery EDM, Ross MR, Hawkins M, Leslie G, Gyory AZ. Low-dose aspirin in high-risk pregnancy? Hypertens Pregnancy 1997;16:229-38.

18. Byaruhanga RN, Chipato T, Rusakaniko S. A randomized controlled trial of low-dose aspirin in women at risk from pre-eclampsia. Int J Gynecol Obstet 1998;60:129-35.

19. Caritis S, Sibai B, Hauth J, Lindheimer MD, Klebanoff M, Thom E, et al. Low-dose aspirin to prevent preeclampsia in women at high risk. National Institute of Child Health and Human Development Network of Maternal-Fetal Medicine Units. N Engl J Med 1998;338:701-5.

20. Dumont A, Merviel P, Berkane N, Gaudet R, Uzan S. Risk factors for pre-eclampsia. Presse Med 1999;28:2189-96.

21. Visser W, Wallenburg HC. Prediction and prevention of pregnancy-induced hypertensive disorders. Baillieres Best Pract Res Clin Obstet Gynaecol 1999;13:131-56.

22. NHS Centre for Reviews and Dissemination. Undertaking systematic reviews of research on effectiveness. CRD's guidance for carrying out or commissioning reviews. CRD report number 4. (2nd edition) York, United Kingdom: University of York, NHS Centre for Reviews and Dissemination, 2001.

23. Landis JR, Koch GG. The measurement of observer agreement for categorical data. Biometrics 1977;33:159-74.

24. Yusuf S, Peto R, Lewis J, Collins R, Sleight P. Beta blockade during and after myocardial infarction: An overview of the randomized trials. Prog Cardiovasc Dis 1985;27:335-71.

25. Thompson SG, Pocock SJ. Can meta-analyses be trusted? Lancet 1991;338:1127-30.

26. Mantel N, Haenszel W. Statistical aspects of the analysis of data from retrospective studies of disease. J Natl Cancer Inst 1959;22:719-48.

27. DerSimonian R, Laird N. Meta-analysis in clinical trials. Control Clin Trials 1986;7:177-88.

28. Clarke M, Oxman MD, eds. Investigating heterogeneity. Cochrane reviewers handbook (updated June 2000). Section 8.6. Oxford, United Kingdom: The Cochrane Collaboration, 2000.

29. Light RJ, Pillemer DB. Summing up: The science of reviewing research. Cambridge, Massachusetts: Harvard University Press, 1984.

30. Egger M, Davey SG, Schneider M, Minder C. Bias in meta-analysis detected by a simple, graphical test. BMJ 1997;315:629-34.

31. Benigni A, Gregorini G, Frusca T, Chiabrando C, Ballerini S, Valcamonico A, et al. Effect of low-dose aspirin on fetal and maternal generation of thromboxane by platelets in women at risk for pregnancy-induced hypertension. N Engl J Med 1989;321:357-62.

32. Beroyz G, Casale R, Farreiros A, Palermo M, Margulies M, Voto L, et al. CLASP: A randomised trial of low-dose aspirin for the prevention and treatment of pre-eclampsia among 9364 pregnant women. Lancet 1994;343:619-29.

33. ECPPA (Estudo Colaborativo para Prevencao da Pre-eclampsia com Aspirina) Collaborative Group. ECPPA: Randomised trial of low dose aspirin for the prevention of maternal and fetal complications in high risk pregnant women. Br J Obstet Gynaecology 1996;103:39-47.

34. Volpicelli T, D'Anto V, Faticato A, Galante L, Civitillo R. Trial prospecticco sull'uso profillatico dell'aspirina in donne gravide ad alto rischio di pre-eclampsia. Gestosi 1999;99:159-60.

35. Knight M, Duley L, Henderson-Smart DJ, King JF. Antiplatelet agents for preventing and treating pre-eclampsia. Cochrane Database Syst Rev 2000;2:CD000492.

36. Duley L, Henderson-Smart D, Knight M, King J. Antiplatelet drugs for prevention of pre-eclampsia and its consequences: Systematic review. BMJ 2001;322:329-33.

37. Coomarasamy A, Papaioannou S, Gee H, Khan KS. Aspirin for the prevention of preeclampsia in women with abnormal uterine artery Doppler: A meta-analysis. Obstet Gynecol 2001;98:861-6.

38. Ernst E, Pittler MH. Assessment of therapeutic safety in systematic reviews: Literature review. BMJ 2001;323:546.

39. Slone D, Siskind V, Heinonen OP, Monson RR, Kaufman DW, Shapiro S. Aspirin and congenital malformations. Lancet 1976;1:1373-5.

40. Klebanoff MA, Berendes HW. Aspirin exposure during the first 20 weeks of gestation and IQ at four years of age. Teratology 1988;37:249-55.

41. Nielsen GL, Sorensen HT, Larsen H, Pedersen L. Risk of adverse birth outcome and miscarriage in pregnant users of non-steroidal anti-inflammatory drugs: Population based observational study and case-control study. BMJ 2001;322:266-70.

42. Werler MM, Mitchell AA, Shapiro S. The relation of aspirin use during the first trimester of pregnancy to congenital cardiac defects. N Engl J Med 1989;321:1639-42.

43. Smeeth L, Haines A, Ebrahim S. Numbers needed to treat derived from meta-analyses—sometimes informative, usually misleading. BMJ 1999;318:1548-51.

44. Cook RJ, Sackett DL. The number needed to treat: A clinically useful measure of treatment effect. BMJ 1995;310:452-4.

Address reprint requests to: Aravinthan Coomarasamy, MRCOG, Research Fellow in Evidence-Based Obstetrics, Birmingham Women's Hospital, Education Resource Center, Metchley Park Road, Birmingham, B15 2TG, United Kingdom; E-mail: arricoomar@blueyonder.co.uk

Received September 13, 2002.
Received in revised form December 4, 2002.
Accepted December 12, 2002.