Bromocriptine treatment and outcomes in peripartum cardiomyopathy: the EORP PPCM registry

Abstract

Background and Aims

Peripartum cardiomyopathy (PPCM) remains a serious threat to maternal health around the world. While bromocriptine, in addition to standard treatment for heart failure, presents a promising pathophysiology-based disease-specific treatment option in PPCM, the evidence regarding its efficacy remains limited. This study aimed to determine whether bromocriptine treatment is associated with improved maternal outcomes in PPCM.

Methods

Peripartum cardiomyopathy patients from the EORP PPCM registry with available follow-up were included. The main exposure of this exploratory non-randomized analysis was bromocriptine treatment, and the main outcome was a composite endpoint of maternal outcome [death or hospital readmission within the first 6 months after diagnosis, or persistent severe left ventricular dysfunction (left ventricular ejection fraction < 35%) at 6-month follow-up]. Inverse probability weighting was used to minimize the effects of confounding by indication. Multiple imputation was used to account for the missing data.

Results

Among the 552 patients with PPCM, 85 were treated with bromocriptine (15%). The primary endpoint was available in 491 patients (89%) and occurred in 18 out of 82 patients treated with bromocriptine in addition to standard of care (22%) and in 136 out of 409 patients treated with standard of care (33%) (P = .044). In complete case analysis, bromocriptine treatment was associated with reduced adverse maternal outcome [odds ratio (OR) 0.29, 95% confidence interval (CI) 0.10–0.83, P = .021]. This association remained after applying multiple imputation and methods to correct for confounding by indication (inverse probability weighted model on imputed data: OR 0.47, 95% CI 0.31-0.70, P < 0.001). Thromboembolic events were observed in 6.0% of the patients in the bromocriptine group vs. 5.6% in the standard of care group (P = .900).

Conclusions

Among women with PPCM, bromocriptine treatment in addition to standard of care was associated with better maternal outcomes after 6 months.

Structured Graphical Abstract

The association between bromocriptine treatment and maternal outcome in patients with peripartum cardiomyopathy. CI, confidence interval; LV, left ventricular.

Open in new tabDownload slide

See the editorial comment for this article ‘Bromocriptine in the treatment of peripartum cardiomyopathy: is it ready for prime time?', by U. Elkayam, https://doi.org/10.1093/eurheartj/ehae875.

Introduction

Peripartum cardiomyopathy (PPCM) is a significant cause of global maternal mortality, particularly in some regions.¹ It is defined as a cardiomyopathy with left ventricular ejection fraction (LVEF) < 45% that occurs late during pregnancy or in the early postpartum period.^2,3 While the disease course may be mild in some cases, mortality rates reach as high as 20%, despite the young age of the patients.⁴ Recently, a simple score was derived from the EORP PPCM registry to predict left ventricular (LV) recovery in patients with PPCM.⁵ Although guideline-directed medical therapy (GDMT) for heart failure with reduced ejection fraction is recommended in patients with PPCM, the disease lacks proven disease-specific therapies.¹

From a mechanistic perspective, oxidative stress in combination with elevated concentrations of prolactin results in the generation of a pro-apoptotic 16 kDa prolactin fragment.⁶ Dopamine 2D receptor agonists, such as bromocriptine, are prolactin-suppressing agents used to cease nursing and have successfully attenuated PPCM in mouse models^6,7 and have therefore been investigated as a potential treatment of PPCM, yet existing evidence remains limited.^1,8,9 Regional disparities persist in recommendations regarding its use, with no consensus reached in the USA while the European Society of Cardiology (ESC) guideline suggests that it may be considered (IIb recommendation).^10,11 The ongoing Randomized Evaluation of Bromocriptine in Myocardial Recovery Therapy for Peripartum Cardiomyopathy (REBIRTH) trial aims to address this uncertainty, but results are not expected until 2029, underscoring the need for additional evidence to inform treatment decisions in PPCM (ClinicalTrials.gov number, NCT05180773).

In this prospective, multicentre, cohort study, we aim to bridge this gap in evidence by investigating the association between bromocriptine treatment and adverse maternal outcomes 6 months post-diagnosis in women with PPCM.

Methods

In 2011, the ESC invited over 100 national and affiliated member cardiac societies to contribute to a global registry on PPCM as part of the ESC EURObservational Research Programme (EORP).¹² These societies were tasked with identifying centres to participate in the registry. In low-income countries without national cardiac societies, potential study sites were identified through abstracts and publications. Participating centres were required to have clinical expertise in diagnosing PPCM, availability of echocardiography, and the ability to follow up with patients for at least 6 months.

Women newly diagnosed with PPCM were enrolled in the study between 2012 and 2018 with a mandatory 6-month follow-up. To be eligible, participants had to meet the following criteria at the time of diagnosis: (i) be in a peripartum state (i.e. last trimester of pregnancy up to the first 6 months postpartum), (ii) exhibit signs and/or symptoms of heart failure, (iii) have an LVEF ≤45%, and (iv) have no other identifiable causes of heart failure. From the initial registry of 739 patients,¹³ only those with known information regarding their bromocriptine therapy status (n = 552) were included in this study.

Baseline data collected included demographics, comorbidities, obstetric history, signs, symptoms, and results from blood tests, electrocardiography, chest radiography, echocardiography, and pharmacological therapy.

Poor outcome was determined by the composite endpoint of death; hospital readmission, within the first 6 months after diagnosis; or persistent severe LV dysfunction (LVEF < 35%) at 6-month follow-up. Secondary outcomes included thromboembolic events (either arterial or venous).

The study complied with the Declaration of Helsinki, with investigators participating voluntarily and without compensation. Approvals were obtained by participating centres from national or regional ethics committees and Institutional Review Boards, following local regulations. This observational study was supported by a central study management team and did not involve specific protocols or recommendations for diagnosis or management.

Statistical analysis

Descriptive statistics were used to summarize the data. Continuous variables were presented as means with standard deviations (SD) for parametric data or medians with interquartile ranges (IQR) for non-parametric data. Categorical variables were expressed as frequencies and percentages. Baseline characteristics were described and stratified based on whether patients received bromocriptine treatment. Continuous data were compared using independent t-tests for normally distributed data or the Mann–Whitney U test for non-normally distributed data. Binary data were compared using Pearson’s χ² test or Fisher’s exact test depending on the sample size.

Multiple imputation and inverse probability weighting

Missing data was imputed with the Multivariate Imputation with Chained Equations (mice, version 3.14.0) in R. Fifty imputed data sets were generated with five iterations per data set. Continuous variables were imputed with predictive mean matching, binary data with logistic regression, and categorical data with polytomous regression. No abnormalities in the imputed values were found.

Due to the non-randomized nature of this study, there is the risk of confounding by indication. To minimize the potential impact of this bias, inverse probability weighting and multivariable adjusted regression were used. In short, first, least absolute shrinkage and selection operator (LASSO) regression was used to identify variables related to adverse maternal outcome. The R package glmnet (version 4.1–4) was used to perform LASSO regression, and directed acyclic graphs were constructed to identify possible confounders for the association between bromocriptine treatment and adverse maternal outcome. Due to multicollinearity between LV end-diastolic and end-systolic diameter, only LV end-diastolic diameter (LVEDD) was included. History of diabetes mellitus, anticoagulant treatment, treatment with mineralocorticoid receptor antagonists (MRAs), treatment with loop diuretics, systolic blood pressure at baseline, QRS duration at baseline, LVEDD, LVEF, geographical region, right ventricular failure at baseline, and cardiomegaly at chest X-ray at baseline were associated with adverse maternal outcome.

In the following step, weights were generated for inverse probability weighting according to previously described methods.^14,15 The variables in the aforementioned paragraph were used to generate weights, except MRA, which was replaced by treatment with GDMT, which is treatment with angiotensin-converting enzyme inhibitors (ACEi) or angiotensin receptor blockers (ARBs), beta-blockers and MRAs.^13,14 Inverse probability weighting was performed using the R package MatchThem (version 1.10). Covariate balance was assessed using the absolute standardized mean differences across all imputed data sets.

Finally, two logistic regression models were fitted, one on the weighted data and one unweighted but adjusted for multiple confounding variables.

Results

Baseline characteristics

Of the 552 patients with PPCM, 85 received treatment with bromocriptine whereas 467 did not (Figure 1). There were no differences in age at baseline [bromocriptine vs. no bromocriptine: 31.5 years (± 5.6) vs. 30.9 years (± 6.2), P = .40], parity (>2 children 83.0% vs. 72.1%, P = .096), and previous presence of PPCM (7.5% vs. 7.6%, P = .98). There was a significant difference in the geographical region regarding bromocriptine prescription. For example, patients living in Africa were more likely to receive bromocriptine compared with other global regions (P < .001). Patients receiving bromocriptine experienced more severe heart failure symptoms (87.1% of women who received bromocriptine had a New York Heart Association functional class III/IV at time of diagnosis vs. 62.4% in those who did not, P < .001).

There were no significant differences in the presence of hypertension during pregnancy nor in LVEDD nor in QRS duration at baseline, all of which have been shown to be important predictors of LV recovery (both cohorts had a PPCM LV recovery prediction score of 4).⁵ Furthermore, no significant differences were observed in LV systolic function [LVEF 34% (IQR 26–40) in the bromocriptine group vs. LVEF 31% (IQR 23–38) in the no bromocriptine group; P = .09). However more patients not receiving bromocriptine had at baseline a LVEF ≤ 25% (36.2% vs. 20.2%, P = .014).

There were significant differences in the use of GDMT between the two groups. Patients that received bromocriptine were more likely to receive ACEi/ARBs and MRAs compared with those not receiving bromocriptine (96.5% vs. 82.9% for ACEi/ARB, P = .001, and 71.8% vs. 40.1% for MRA, P < .001). No differences in beta-blocker treatment (bromocriptine group 75.3% vs. 82.4%, P = .12) or oral anticoagulation at baseline were observed (bromocriptine group 20.2% vs. 15.3%, P = .26). Loop diuretics were significantly more often prescribed in patients receiving bromocriptine (69% vs. 98%, P < .001). Other baseline characteristics can be found in Table 1.

Maternal outcome

The primary composite endpoint was available in 491 of 552 patients (89%) and occurred in 18 out of 82 patients treated with bromocriptine (22.0%) and in 136 out of 409 patients treated with standard of care (33.3%, P = .044) (Figure 2). This was primarily driven by fewer patients with severe LV dysfunction (defined as an LVEF < 35%) at 6 months in the bromocriptine group (15.0% vs. 24.5%). The occurrence of the primary endpoint and its components are summarized in Table 2.

In complete case analysis (Table 3), bromocriptine treatment, adjusted for possible confounders, was associated with reduced adverse maternal outcome [odds ratio (OR) 0.29, 95% confidence interval (CI) 0.10–0.83, P = .021]. This association remained after applying multiple imputation and methods to correct for confounding by indication. The estimates for bromocriptine in the inverse probability weighted model, based on the imputed data for the primary endpoint, were as follows: OR 0.47, 95% CI 0.31–0.70, P < 0.001. The results are summarized in Figure 3 and Table 3. Thromboembolic events were observed in 6.0% of the patients in the bromocriptine group vs. 5.6% in the standard of care group (P = .900).

Discussion

Principal findings

The principal finding of this study is that bromocriptine treatment was associated with better maternal outcome which was primarily driven by fewer patients with severe LV dysfunction after 6 months (Structured Graphical Abstract). Furthermore, no differences in thromboembolic events were observed in the bromocriptine group vs. the standard of care group. These data may help guide treatment decisions until evidence from randomized studies that are currently being conducted becomes available.

Comparison with other studies

Clinical data regarding the efficacy of bromocriptine treatment in patients with PPCM remain limited. Only two randomized studies have been conducted that have provided some evidence that bromocriptine treatment might be beneficial; however, the evidence remains inconclusive due to the low number of patients that was included or the lack of a true control group in one of the studies.^8,9 A systematic review and meta-analysis of the available evidence, including these two randomized studies and six observational studies, encompassing 593 patients with PPCM, demonstrated that bromocriptine treatment was associated with significantly higher survival rates and greater improvement in LVEF.¹⁶ The results of our study are consistent with these previous findings, showing that bromocriptine treatment is associated with better maternal outcomes. In the current study, this result was mainly driven by a lower number of patients with severe LV dysfunction after 6 months, despite rigorous correction for all baseline characteristics known to impact LV recovery.

Two recent studies investigated which factors are associated with an improvement in LV function in patients with PPCM. In a Scottish registry comprising of 225 women with PPCM, only LVEDD was independently associated with an improved LV function at 1 year.¹⁷ Similar findings were observed in the EORP PPCM registry. Here, we found that the two strongest predictors of LV recovery, defined as a LVEF ≥ 50%, were also LVEDD and, in addition to that, QRS duration.^5,18,19 Both of these variables were comparable between patients receiving bromocriptine or no bromocriptine in our study. Another important predictor for LV recovery and maternal outcome in patients with PPCM is the development of hypertension during pregnancy. Especially the presence of pre-eclampsia is associated with a more than two-fold higher chance of LV recovery compared with women with PPCM with no gestational hypertension²⁰ In our current study, the presence of pre-eclampsia was not significantly different between groups.

Loop diuretics were used in significantly more patients receiving bromocriptine indicating that these patients suffered from more severe acute heart failure/congestion than patients not treated with bromocriptine. The fact that patients in the bromocriptine group tended to be sicker than in the non-treated group indicates that bias may not explain the beneficial effects associated with bromocriptine. This is also supported by several statistical methods, used to account for confounding by indication, which all confirm the main analysis.

We did not observe a reduction in mortality or heart failure rehospitalizations in the bromocriptine group, although it must be noted that overall mortality was relatively low. Interestingly, we did not observe differences in the incidence of thromboembolic events with the use of bromocriptine. Previously, several case reports described the occurrence of myocardial infarction and retinal vein occlusion in women with PPCM also treated with bromocriptine.^21,22 It is clear that women with PPCM similar to all women in the peripartal state are at increased risk for thromboembolic events; we have previously shown that 5% of the women with PPCM present with thromboembolism which is likely related to the hypercoagulable state, an evolutionary remnant to minimize postpartum bleeding.²³ This in combination with LV dilatation, endothelial injury, and immobility may explain this high rate. However, our study shows that the use of bromocriptine was not associated with incident thromboembolic events. Since only 20% of the patients on bromocriptine were also using anticoagulation, our data do not support long-term treatment with oral anticoagulation in women with PPCM. However, due to the relatively low incidence of thromboembolic events during follow-up, these results alone should not inform coagulation treatment decisions.

Strength and limitations

The study’s relatively large sample size and the methods to account for confounding by indication are its main strengths, and the results of this study may aid in bridging the gap until the results of the REBIRTH trial are published, which is expected to be in the beginning of 2029.

Several limitations must be acknowledged. First, although multiple strategies have been employed to correct for confounding by indication, these methods only allow for correction for known confounders. Therefore, there remains the risk of residual confounding due to unknown confounders. Second, patients with unknown bromocriptine status were not included in the study (n = 187) which introduces the risk of selection bias. Third, the results of this observational, non-randomized study should be interpreted cautiously, and further randomized studies are warranted to provide evidence of causality. However, future randomized trials will be limited by inadequate blinding as bromocriptine treatment leads to cessation of lactation.

Conclusion

Among women with PPCM, bromocriptine treatment was associated with better maternal outcomes after 6 months and no increased risk of thromboembolic events.

Acknowledgement

The EORP Oversight Committee and PPCM Registry Executive Committee are listed in Appendix 1. Data collection was conducted by the EORP department of the ESC, and includes Afiah Zabre and Emmanuelle Mouraux as Clinical Project Managers, Gabrielle Bonneville as Project Officer, and Sandrine Anglars as Data Manager. We acknowledge the support of Dr Fareda Jakoet-Bassier for the logistical aspects related to the study. We also acknowledge the participation of all investigators who entered patients into the study (also listed in Appendix 1).

Supplementary data

Supplementary data are available at European Heart Journal online.

Declarations

Disclosure of Interest

P.v.d.M. is supported by a grant from the European Research Council (ERC CoG 101045236, DISSECT-HF). The UMCG, which employs P.v.d.M., received consultancy fees and/or grants from Novartis, Pharmacosmos, Vifor Pharma, AstraZeneca, Pfizer, Pharma Nord, BridgeBio, Novo Nordisk, Daiichi Sankyo, Boehringer Ingelheim, and Ionis. C.V. received honoraria for lectures from Biotronik not related to this article. J.H. received honoraria for lectures from Boehringer Ingelheim not related to this article. A.M. received research contracts from 4TEEN4, Roche, Sphingotec, Abbott Diagnostics, and Windtree; consultation fee from Roche Corteria, Adrenomed, and Fire; honorarium for lecture from Merck, Novartis, Roche, and Bayer; is co-inventor of patient on combined therapies to treat dyspnoea; is a co-inventor of a patent owned by S-Form Pharma; and is a member of Committee of Trials for Secret-HF, sponsored by the French Government, for S-Form Pharma and Implicity. M.C.P. has received research funding form Boehringer Ingelheim, Roche, SQ Innovations, AstraZeneca, Novartis, Novo Nordisk, Medtronic, Boston Scientific, and Pharmacosmos and consulted or served on committees for Abott, Akero, Applied Therapeutics, Amgen, AnaCardio, Biosensors, Boehringer Ingelheim, Corteria, Novartis, AstraZeneca, Novo Nordisk, Abbvie, Bayer, Horizon Therapeutics, Foundry, Takeda, Cardiorentis, Pharmacosmos, Siemens, Eli Lilly, Vifor, New Amsterdam, Moderna, Teikoku, LIB Therapeutics, 3R Lifesciences, Reprieve, FIRE 1, Corvia, and Regeneron. S.G. has received honoraria for lectures/consulting/serving on committees from AstraZeneca, Novo Nordisk, BMS, and Cardurion and a research grant from BMS unrelated to the article. W.O. received consulting from Us2.ai and owns patent US-10702247-B, unrelated to this article. J.T. is supported by the National University of Singapore Start-Up grant, the tier 1 grant from the Ministry of Education, and the CS-IRG New Investigator Grant from the National Medical Research Council; has received research support from AstraZeneca and consulting or speaker fees from Us2.ai; and owns patent US-10702247-B2 unrelated to the present work. J.B. received honoraria for lectures/consulting from Novartis, Vifor, Bayer, Pfizer, Boehringer Ingelheim, AstraZeneca, Cardior, CVRx, BMS, Amgen, Corvia, Norgine, Edwards, and Roche not related to this article and research support for the department from Zoll, CVRx, Abiomed, Norgine, and Roche, not related to this article. M.B. is supported by the Deutsche Forschungsgemeinschaft (German Research Foundation; TTR 219, project number 322 900 939) and reports personal fees from Abbott, Amgen, AstraZeneca, Bayer, Boehringer Ingelheim, Cytokinetics, Edwards, Medtronic, Novartis, Pharmacosmos, Recor, Servier, and Vifor during the conduct of the study. P.M.S. received consultancy agreement and honorarium for lecture from Boehringer Ingelheim, Novartis, and Menarini and honorarium for lecture from AstraZeneca. All other authors declare no conflicts of interest.

Data Availability

Data are available upon request.

Funding

Since the start of EORP, the following companies have supported the whole research programme: Abbott Vascular Int. (2011–21), Amgen Cardiovascular (2009–18), AstraZeneca (2014–21), Bayer AG (2009–18), Boehringer Ingelheim (2009–19), Boston Scientific (2009–12), The Bristol Myers Squibb and Pfizer Alliance (2011–19), Daiichi Sankyo Europe GmbH (2011–20), The Alliance Daiichi Sankyo Europe GmbH and Eli Lilly and Company (2014–17), Edwards (2016–19), Gedeon Richter Plc. (2014–16), Menarini Int. Op. (2009–12), MSDMerck & Co. (2011–14), Novartis Pharma AG (2014–20), ResMed (2014–16), Sanofi (2009–11), SERVIER (2009–21), and Vifor (2019–22). P.v.d.M. is supported by a grant from the European Research Council (ERC CoG 101045236, DISSECT-HF).

Ethical Approval

The study complied with the Declaration of Helsinki, with investigators participating voluntarily and without compensation. Approvals were obtained by participating centres from national or regional ethics committees and Institutional Review Boards, following local regulations. This observational study was supported by a central study management team and did not involve specific protocols or recommendations for diagnosis or management.

Pre-registered Clinical Trial Number

Not applicable.

References

Author notes