The Effects of Aspirin in Gestation and Reproduction (EAGeR) Trial: A Story of Discovery

Abstract

Human reproduction is an inefficient process. There are several drivers of complications along the path to and during pregnancy, one of which is inflammation. Treatments to mitigate the deleterious effects of aberrant inflammation with something inexpensive and widely available like aspirin could have dramatic global impact. The Effects of Aspirin in Gestation and Reproduction (EAGeR) trial enrolled women aged 18 to 40 years with one to two prior pregnancy losses and no diagnosis of infertility. Patients were randomized to either low-dose aspirin or placebo. Here, we review the collective findings of the EAGeR trial to date and discuss several important lessons learned from the unique data resulting from this groundbreaking trial. Findings reported from this trial provide significant advances in the understanding of aspirin’s potential mechanisms in modulating reproductive processes and the role of inflammation in these processes. This review describes the collective findings of the EAGeR trial in the context of the existing literature regarding aspirin and inflammation in reproduction to inform relevant next steps in fertility and obstetric research, as well as potential implications for clinical care.

Human reproduction is an inefficient process with the chance of conception per menstrual cycle thought to be approximately 25%^1–3 and almost 30% of clinically recognized pregnancies ending in a loss.⁴ Furthermore, the riskof a spontaneous pregnancy loss increases with the number of prior pregnancy losses.⁵ Loss from aneuploidy is the most common cause of early loss (accounting for 50%).⁶ Other causes of loss include uterine anomalies,⁷ antiphospholipid antibody syndrome,⁸ and poorly controlled endocrine disorders such as overt hypothyroidism and diabetes.^9,10 However, the cause of pregnancy loss in euploid pregnancies is unknown in the majority of cases. Abnormally decreased blood flow to the reproductive tract and aberrant inflammation may play a role in the pathologic process.¹¹

Furthermore, while a balance of pro- and anti-inflammatory factors is involved in implantation¹² and parturition,¹³ overactive inflammation has been shown to contribute to preterm birth,¹³ spontaneous early pregnancy loss,¹⁴ gestational diabetes,¹⁵ and preeclampsia.¹⁶ Excess inflammation has also been implicated in many causes of infertility including poor oocyte quality,¹⁷ pelvic inflammatory disease,¹⁸ polycystic ovarian syndrome,¹⁹ and endometriosis.²⁰ Through modulation of excess inflammation, aspirin therapy has the potential to reduce or correct poor obstetric outcomes. Indeed, aspirin has the ability to increase blood flow and decrease inflammationinreproductiveorgans.²¹Further, aspirin is very low cost, has relatively few side effects, and is widely available.

Studies of aspirin and reproduction have generated mixed results and recent evidence indicates timing of initiating therapy may be critical to its effectiveness. When initiated at the start of the cycle for in vitro fertilization (IVF), aspirin improved endometrial growth and vascularization²² and when initiated at stimulation a recent meta-analysis showed an improvement in clinical pregnancy rate, though not live birth.²³ A clinical trial randomizing women with two or more pregnancy losses to low-dose aspirin (LDA), heparin and LDA, or placebo starting preconception or before 6 weeks’ gestation showed no benefit to LDA or heparin for increasing livebirth.²⁴ Postconception LDA has been studied extensively in recurrent pregnancy loss patients without benefit shown.^25,26 Collectively, these findings suggest that LDA initiated preconceptionally may have beneficial effects on downstream events. However, the use of LDA preconceptionally has not been extensively studied outside of infertility treatment.

Given the suggestive evidence of benefit from preconception aspirin treatment on improving ART outcomes, perhaps through improvements in endometrial vascularization and implantation, in addition to a biological rationale for a potential benefit on preventing pregnancy loss and other adverse outcomes, the recently completed Effects of Aspirin in Gestation and Reproduction (EAGeR) trial investigated the impact of preconception-initiated LDA treatment on live birth in 1,228 women recruited from four U.S. university medical centers who were attempting spontaneous conception after experiencing one or two prior pregnancy losses. Findings reported from this trial provide significant advances in the understanding of LDA’s potential mechanisms in modulating reproductive processes and the role of inflammation in these processes. This review describes the collective findings of the EAGeR trial in the context of the existing literature regarding aspirin and inflammation in reproduction to inform relevant next steps in fertility and obstetric research, as well as potential implications for clinical care.

The EAGeR Trial: Lessons from Preconception Aspirin Therapy

As described earlier, prior trials of LDA treated women who were either already pregnant^24,26–30 or undergoingIVF.^22,31–33 Furthermore, LDA is currently recommended to begin at 12 weeks’ gestation for women who are high risk for preeclampsia.³⁴ Thus, the EAGeR trial findings produced novel information regarding the impacts of aspirin and the role of inflammation in the preconception period for women attempting spontaneous conception. Women included in EAGeR were aged 18 to 40 years with up to two prior live births. Women with one prior loss less than 20 weeks of gestation within the preceding 12 months with no infertility diagnosis were enrolled under the “original” eligibility stratum. Additionally, women with one or two prior losses at any gestational age and at any time in the past were enrolled into the “expanded” eligibility stratum, more closely representing a general obstetric population with a history of pregnancy loss. Patients were followed up for six menstrual cycles plus the duration of pregnancy if applicable.

First, the findings regarding the primary outcome of live birth in the overall trial and by eligibility stratum indicated some beneficial impact of aspirin, but only among certain women. Among all participants, there was no statistically significant improvement in live birth attributable to LDA compared with placebo (relative risk [RR]: 1.10, 95% confidence interval [CI]: 0.98, 1.22); however, there was a statistically significant difference in live birth between the LDA and placebo group within women with a single recent loss that made up original stratum (62 vs. 53%, p = 0.045).³⁵ In the expanded stratum alone, there was no difference between treatment groups (54 vs. 52%, p = 0.74). Interestingly, there were significantly more positive pregnancy tests and ultrasound-confirmed pregnancies in the LDA versus placebo group, accounting for the increased live birth rate observed in the original stratum and the marginally increased livebirths in the trial overall (►Fig. 1a, b). Concordantly, despite a hypothesis that aspirin may decrease pregnancy loss through a potential improvement in endometrial blood flow and reduced inflammation,²¹ there was no effect of LDA on pregnancy loss at any stage of pregnancy, either in the overall trial cohort or within eligibility strata.³⁵ Moreover, to further investigate aspirin’s potential effect on supporting existing pregnancy, the type of loss (i.e., implantation failure, early pregnancy loss, or stillbirth) was examined in detail, and LDA was not found to have any effect on any particular subtype of loss. Notably, an analysis of chromosomal abnormalities determined from products of conception collected from early pregnancy losses further demonstrated that LDA did not decrease the rate of euploid pregnancy loss.³⁶ Collectively, these primary trial findings suggested that initiating LDA therapy prior to conception may have important impacts in successful establishment, as opposed to maintenance, of pregnancy. Of note, such effects could not be observed in previous trial initiating LDA therapy after pregnancy confirmation.

Fig. 1.

(a) Overall cohort by pregnancy outcome. There is a slight nonstatistically significant increase in live births with low-dose aspirin (LDA) compared with placebo, with no difference in pregnancy loss. (b). Original stratum by pregnancy outcome. There is a statistically significant increase in live births with LDA compared with placebo, with no difference in pregnancy loss. (Adapted from Schisterman et al.³⁵)

Indeed, a specific analysis of LDA’s effect on time to pregnancy in the overall cohort reflected thefindings of pregnancy and live birth rates in that LDA increased fecundability by 28% in the LDA compared with placebo group within the original stratum. This was true for either a human chorionic gonadotropin–positive pregnancy or an ultrasound-confirmed pregnancy. This would support the hypothesis that aspirin’s effect is not on supporting an established pregnancy but rather exerting its effects on earlier events, such as ovulation or embryo survival and successful implantation.³⁷ Though direct assessment of each of these underlying mechanisms is impossible in humans, an examination of the effect of LDA on anovulation was recently reported.

Anovulation

In eumenorrheic women, reports of the per-cycle chance of sporadic anovulation have ranged from 0 to 11%.^38–42 Prior evidence for an effect of LDA on ovulation is indirect and mixed.^22,43,44 To better answer whether LDA impacts ovulatory function and to better elucidate the mechanism of increased fecundability observed among certain EAGeR participants, a secondary analysis of EAGeR trial data was performed to evaluate the impact of LDA compared with placebo on the occurrence of sporadic anovulation.

For the purposes of this analysis in EAGeR, a cycle was considered ovulatory if the urinary luteinizing hormone was 2.5-fold greater than the mean of the previous 5 days.⁴⁰ Anovulation occurred in 12.2% of all cycles and LDA did not have an effect on the rate of anovulation overall (13.4% LDA vs 11.1% placebo; RR: 1.16, 95% CI: 0.88, 1.52) or among the original stratum (11.1% LDA vs. 10.1% placebo, RR: 1.06, 95% CI: 0.68, 1.54).⁴⁵ These data suggest that the increase in fecundability seen in the LDA group previously³⁷ was not mediated by decreasing risk of having an anovulatory cycle, but rather some other mechanism.

Inflammation and Implantation

Given the potential role of aspirin to affect endometrial receptivity and implantation through modulating inflammation, a secondary analysis stratifying by tertile of high-sensitivity C-reactive protein (hsCRP) at study entry, as a marker of chronic low-grade inflammation, was next undertaken. Indeed, among women in the highest tertile (hsCRP:1.95–9.99mg/L, excluding women with hsCRP ≥10 mg/L) of preconception hsCRP, LDA significantly increased clinicallyconfirmedpregnanciesby31% and increased live births by 35% compared with the placebo group. In contrast, among women in the middle and lower tertiles of hsCRP, LDA did not affect clinically confirmed pregnancies or live birth (►Fig. 2).⁴⁶ Interestingly, the chance of clinically confirmed pregnancy and live birth decreased with increasing level of hsCRP, and among the highest tertile of hsCRP, LDA appeared to restore the pregnancy and live birth rates to similar rates as women with lower hsCRP levels (►Fig. 2). Also, since obesity is known to be an inflammatory state which may be contributing to otherwise healthy women having low-grade inflammation,⁴⁷ women with hsCRP in the top tertile (hsCRP ≥1.95 mg/L) were further stratified by adiposity status, considering both waist:hip ratio, a marker of central obesity, and body mass index (BMI) as an overall measure of adiposity. Interestingly, LDA showed a significant positive effect on pregnancy and live birth among leaner women only, including those below the cohort median waist: hip ratio and also among those with normal BMI (<25 kg/m²) (RR: 1.60 95% CI: 1.11, 2.30). In contrast, women above the cohort median waist:hip ratio, or with BMI ≥25 kg/m², exhibited an attenuated LDA effect (RR: 1.18 95% CI: 0.89, 1.56).⁴⁶ Thus, inflammation level at study entry appeared to significantly modify the effect of LDA on pregnancy chances, with hsCRP acting as a useful biomarker to identify women most likely to benefit from LDA therapy, similar to its application in cardiovascular medicine research. While potentially important for future clinical practice, these findings may also help shed light on the reproductive stage or mechanism through which LDA may impact reproduction. Importantly, it is thought that inflammatory mediators exchanged between the embryo and endometrial surface are an important component of the cross talk which drives the endometrium’s function as a sensor of the preimplantation embryo’s viability, with such communication enabling or disrupting successful embryonic implantation.⁴⁸

Fig. 2.

Clinical pregnancy based on tertile of high-sensitivity C-reactive protein (hsCRP). In the highest tertile, clinical pregnancy was restored in the low-dose aspirin (LDA) group versus the placebo group. In the lower two tertiles, there was no difference between LDA and placebo. (Reprinted with permission from Sjaarda et al.⁴⁶)

Indeed, a crucial purpose of the decidua’s sensor function is to recognize chromosomally abnormal embryos by their increased metabolic activity and prevent their implantation with a proteotoxic stress response.⁴⁸ Thus, it was hypothesized that this appropriate inflammatory response may become pathological in the presence of an underlying disordered maternal inflammatory milieu, resulting in rejection of chromosomally normal embryos. Moreover, male embryos may be more vulnerable to rejection by an overactive maternal inflammatory environment, as preimplantation male embryos have demonstrated greater metabolic activity than female embryos.^49,50 Further, in mice and bovine models, preimplantation embryos have been shown to exhibit a sexually dimorphic response to maternal inflammation, with male embryos exhibiting greater vulnerability.^51,52 As such, the resulting sex ratio of women treated with LDA could lend a clue to whether the key effects of an anti-inflammatory agent such as aspirin occur at the time of implantation when this delicate embryo–endometrium crosstalk is occurring. Examining impacts of exogenous factors on sex ratio is not new, as the ratio of male-to-female newborn infants has trended downward globally over the past few decades.^53–57 Parental exposure to smoking,⁵⁸ dioxin,⁵⁹ lead,⁶⁰ methylmercury,⁶¹ pesticides,⁶² and earthquakes^63,64 have all been linked to reduced male-to-female sex ratio.

Thus, in a prespecified secondary analysis of the EAGeR trial⁶⁵ to investigate the impacts of LDA therapy on offspring sex ratio, LDA was shown to increase male live births overall (RR: 1.31, 95% CI: 1.07, 1.59) and had no effect on female live birth. A further analysis found a similar positive effect of LDA on pregnancy with male offspring by including the outcome of pregnancy with a male embryo, whether the pregnancy ended in live birth or loss. Karyotype or microarray determined the offspring sex from pregnancy losses.⁶⁵ To further evaluate the hypotheses that inflammation was hazardous to male embryos, and that LDA was subsequently protective, the analysis was stratified by hsCRP level at study entry. Congruent with the findings of the stratified analysis of LDA and live birth, the percentage male at birth decreased with increasing tertile of hsCRP among the placebo group, but the proportion of males remained at the expected level among the LDA-treated group (►Fig. 3). Thus, the lower pregnancy rates observed in women with higher inflammation appeared to be driven by a discordant lack of male pregnancies in women with chronic inflammation, and this discordance was restored with preconception LDA therapy.

Fig. 3.

Percentage of male live births by high-sensitivity C-reactive protein (hsCRP) tertile. In the highest tertile, the placebo group was less likely to have a male live birth than the low-dose aspirin (LDA) group. (Adapted from Radin et al.⁴⁵)

Collectively, these data support and help elucidate LDA’s effect on implantation, a key regulatory step tied to delicate inflammatory mediators, in successful establishment of a healthy, ongoing pregnancy. There is good biologic plausibility to these findings given the intricately involved mechanisms of inflammatory mediators and endometrial receptivity and implantation.⁴⁸ LDA’s inhibition of COX-2 function may lead to downstream inhibition of chronic inflammatory pathways.⁶⁶ It has been well documented that patients with central obesity have higher levels of endogenous inflammation compared with individuals with normal BMI.¹⁷ However, the impact of LDA on obesity-driven inflammation is still unclear, as a minimal effect of LDA was observed in patients with higher levels of hsCRP who were overweight/obese. It may be that LDA, at 81 mg, is too low a dose to overcome the overall greater level of inflammation in women with excess adiposity, or that obesity-driven inflammation represents a different inflammatory milieu that may require treatments impacting different pathways than those inhibited by aspirin. As the percentage of overweight and obese females of reproductive age is almost 60% in the United States,⁶⁷ further work to elucidate the complex interplay of adiposity, inflammation, and the role of anti-inflammatory treatments to improve reproduction in this population is needed.

Adverse Pregnancy Outcomes

In addition to the effect of LDA on pregnancy and live birth rates, stratified by level of inflammation, the effect of LDA on hsCRP concentrations through pregnancy was also completed. Indeed, inflammation has been implicated in many adverse pregnancy outcomes, including preeclampsia and gestational diabetes.^15,16,68 Examining the effect of LDA on hsCRP at 8, 20, and 36 weeks of gestation revealed that women in the highest tertile of hsCRP at study entry treated with LDA demonstrated a significant reduction in hsCRP’s elevation throughout pregnancy, whereas LDA had no effect on hsCRP concentrations through pregnancy in women beginning the study in the lower tertiles (►Fig. 4).⁴⁶ Given this potential role for LDA to lower excessive elevation of hsCRP concentrations during pregnancy, an obvious potential application of LDA therapy may be to prevent adverse pregnancy outcomes such as preterm birth, which is a major cause of neonatal mortalityand morbidity.^69,70 Furthermore, LDA is currently recommended for women at high risk for preeclampsia but not until 12 weeks’ gestation.³⁴ A large systematic review of LDA initiated during pregnancy among more than 30,000 women noted that the relative risk of delivery before 37 weeks of gestation was 0.92 (95% CI: 0.88, 0.97) compared with no LDA.⁷¹ Aspirin therapy was mostly started after the first trimester in these studies, however. Within this large prior analysis, the decrease in preterm birth is thought to be from a decrease in medically indicated births, though this is not specified. Thus, given these prior findings and the observation of a reduction in excess inflammation during pregnancy in women in the third hsCRP tertile in EAGeR, an important step was also to examine the impact of LDA on adverse pregnancy outcome.

Fig. 4.

Difference in high-sensitivity C-reactive protein (hsCRP) concentrations in low-dose aspirin (LDA) versus placebo groups during pregnancy. Symbols indicate geometric means of log-transformed high-sensitivity C-reactive protein (hsCRP) baseline (pre-randomization) and throughout pregnancy. Squares indicate the lower hsCRP tertile (closed, LDA; open, placebo), triangles indicate middle hsCRP tertile (closed, LDA; open, placebo), and diamonds indicate higher hsCRP tertile (closed, LDA; open, placebo). Notation of significance indicates treatment group (LDA vs. placebo) difference across pregnancy (excluding baseline) from generalized estimating equations (GEE) models within each hsCRP tertile (assigned at baseline). (Reprinted with permission from Sjaarda et al.⁴⁶)

In this secondary analysis in EAGeR participants,⁷² no significant reduction in the preterm birth rate was observed among the overall study population (RR: 0.72; 95% CI: 0.42, 1.23). A separate analysis of medically indicated and spontaneous preterm births demonstrated a nonsignificant trend toward lower risk in the LDA group for spontaneous preterm birth. However, when restricting to only patients achieving a clinically confirmed pregnancy, a reduction in preterm birth in the original stratum (RR: 0.39; 95% CI: 0.16, 0.94) but not in either the expanded cohort or in the overall population was observed. Interestingly, these lower rates of preterm birth tended to be more from spontaneous births than medically indicated births, which is consistent with a recent metaanalysis.⁷³ Originally, it was postulated that the main mechanism for this reduction in preterm birth is aspirin’s effects on reducing the rate of preeclampsia and intrauterine growth restriction.⁷¹ Thus, these data from the EAGeR trial are of particular interest, as they clearly demonstrate that preconception LDA may reduce spontaneous preterm birth more so than medically indicated preterm birth, and also demonstrate no reduction in preeclampsia, though cases of preeclampsia in EAGeR were few (33 LDA vs. 22 placebo, p = 0.17). The inflammatory nature of labor¹³ lends biologic plausibility to a greater impact on spontaneous preterm birth. Parturition involves moving from a uterine quiescent state to a proinflammatory state which then promotes uterine contraction, cervical dilation, and rupture of membranes.¹³ Aspirin’s ability to inhibit cyclooxygenase decreases prostaglandin formation, which is known to play a role in both normal and abnormal labor.⁷⁴ Furthermore, cyclooxygenase inhibitors are used as tocolytics⁷⁵ and regular-dose aspirin (325 mg) is associated with a delay in the onset of labor.⁷⁶ It is important to note that the EAGeR trial was not powered to detect an effect of LDA on preterm birth, however, and overall rates of pregnancy complications were low, as this population was selected to be without chronic medical conditions. As such, an examination of the impact of LDA on preterm birth and other adverse pregnancy outcomes stratified by hsCRP was not possible and remains an important area of future study, given the findings of LDA’s effect on lowering excess hsCRP throughout pregnancy in women beginning the study with low-grade inflammation.

Finally, as aspirin is inexpensive and widely available, its potential to decrease spontaneous preterm birth, a leading cause of neonatal morbidity and mortality, would have dramatic global implications. With this in mind, the Aspirin Supplementation for Pregnancy Indicated Risk Reduction in Nulliparous (ASPIRIN) trial (NCT02409680) currently conducted by the NICHD Global Network for Women’s and Children’s Health is open and actively enrolling patients to better elucidate the effect of aspirin on preterm birth.

Safety

While there is much yet to discover for the best clinical role for aspirin in fertility and obstetric care, a discussion of its role would be incomplete without a careful examination of risks of aspirin during pregnancy. Aspirin has repeatedly been shown to be of low risk to both mother and fetus⁷⁷ and its use during pregnancy has been found to reduce the risk of preeclampsia and is recommended for women at high risk for developing preeclampsia.³⁴ Based on these data, it is currently recommended that daily LDA therapy be initiated after the 12th week of pregnancy, as is recommended for preeclampsia prevention.⁷⁸ This recommendation comes from several randomized controlled trials starting aspirin after the first trimester.³⁴ The EAGeR trial afforded the opportunity to assess the effects of preconception aspirin use and its safety during early pregnancy as well.⁷⁹ With the evidence of aspirin’s effects on fecundity discussed here and potential effects on other reproductive outcomes, there may be an increase in its preconception use.

Overall there was no difference in symptoms related to aspirin use (i.e., gastrointestinal discomfort, nausea or vomiting, allergic reaction, and difficulty breathing) in the LDA versus placebo groups. In terms of maternal complications, vaginal bleeding and subchorionic hematoma (26 vs. 20%, p = 0.01) occurred more often in the LDA group than in the placebo group. There was no difference between treatment groups in regard to subchorionic hemorrhage, epistaxis, placental abruption, and postpartum hemorrhage. It is important to note that there was no difference in pregnancy loss between the two groups despite higher rates of vaginal bleeding in the LDA group.³⁵ Very few fetal or neonatal complications occurred in the trial overall and there were no differences in neonatal deaths, fetal deaths, or congenital anomalies by treatment group observed.⁷⁹

These data are important in that aspirin was well tolerated while trying to conceive and then after conception was tolerated by the fetus and neonate. These findings are consistent with previous pooled data on aspirin started after the first trimester.^71,77 Our noted increase in vaginal bleeding is in agreement with Truong et al, who found that LDA use increased the risk of subchorionic hematoma.⁸⁰ This retrospective analysis did not report on birth outcomes, whereas the EAGeR trial prospectively followed up the pregnancies and showed no increase in pregnancy loss in the aspirin group.³⁵ Importantly, neonatal and fetal complications were similar in each group. This was of particular importance in that the embryo is exposed to aspirin during organogenesis. This study had insufficient sample size to detect a difference in birth defects, but the data are reassuring and are in agreement with other large study findings.^81,82 Overall, these data should reassure providers using LDA in the preconception period.

Conclusion

The EAGeR trial set out to examine the effects of LDA initiated prior to conception on live birth among women attempting spontaneous conception with a history of pregnancy loss, but no history of infertility. The unfolding story of discovery of LDA’s impact on the trial’s primary and secondary outcomes (live birth, time to pregnancy, preterm birth, sex ratio, and safety) combined with secondary analyses focused on the modifying effect of low-grade inflammation has led to significant advances in knowledge of aspirin, inflammation, and reproduction. Specifically, four important lessons can be gleaned from this evolving body of work. First, the timing of medication likely plays a very important role. This trial is unique in that LDA was started preconceptionally. Most trials, aside from ones involving IVF, started LDA after the establishment of pregnancy; given the data reviewed, beginning LDA after pregnancy confirmation may be of no benefit as the increase in live birth observed was from establishment of pregnancy and not loss prevention. Second, not all patients benefitted in the same way, and the use of a biomarker such as hsCRP may provide an avenue for personalized medicine to identify women most likely to benefit from LDA therapy. Moreover, participants with higher hsCRP and low BMI received the greatest benefit from preconception LDA, leading to new hypotheses regarding the interplay of adiposity, inflammation, and reproduction in the majority of reproductive age women who suffer from excess body weight. Third, because LDA is inexpensive, widely available, and generally regarded as safe, its expanded use in the field of reproductive care could have dramatic global impact. Fourth, inflammation and the modulation of inflammation play a key role in female reproduction with higher inflammation contributing to lower pregnancy rates, particularly pregnancies with male embryos, and excess hsCRP elevation throughout gestation. In conclusion, the findings from EAGeR trial to date have initiated many new paths of clinical discovery in women’s reproductive health that have the potential to improve the lives of women and children globally.