Abstract
Aims
We sought to identify factors associated with right ventricular (RV) dysfunction and elevated pulmonary artery systolic pressure (PASP) and association with adverse outcomes in peripartum cardiomyopathy (PPCM).
Methods and results
We conducted a multi‐centre cohort study to identify subjects with PPCM with the following criteria: left ventricular ejection fraction (LVEF) < 40%, development of heart failure within the last month of pregnancy or 5 months of delivery, and no other identifiable cause of heart failure with reduced ejection fraction. Outcomes included a composite of (i) major adverse events (need for extracorporeal membrane oxygenation, ventricular assist device, orthotopic heart transplantation, or death) or (ii) recurrent heart failure hospitalization. RV function was obtained from echocardiogram reports. In total, 229 women (1993–2017) met criteria for PPCM. Mean age was 32.4 ± 6.8 years, 28% were of African descent, 50 (22%) had RV dysfunction, and 38 (17%) had PASP ≥ 30 mmHg. After a median follow‐up of 3.4 years (interquartile range 1.0–8.8), 58 (25%) experienced the composite outcome of adverse events. African descent, family history of cardiomyopathy, LVEF, and PASP were significant predictors of RV dysfunction. Using Cox proportional hazards models, we found that women with RV dysfunction were three times more likely to experience the adverse composite outcome: hazard ratio 3.21 (95% confidence interval: 1.11–9.28), P = 0.03, in a multivariable model adjusting for age, race, body mass index, preeclampsia, hypertension, diabetes, kidney disease, and LVEF. Women with PASP ≥ 30 mmHg had a lower probability of survival free from adverse events (log‐rank P = 0.04).
Conclusions
African descent and family history of cardiomyopathy were significant predictors of RV dysfunction. RV dysfunction and elevated PASP were significantly associated with a composite of major adverse cardiac events. This at‐risk group may prompt closer monitoring or early referral for advanced therapies.
Keywords: Peripartum cardiomyopathy, Right ventricular dysfunction, Pulmonary hypertension, Adverse outcomes, Pregnancy‐associated heart failure
Introduction
Maternal mortality has been rising in the United States over the past 20 years, despite an overall decline worldwide. 1 Peripartum cardiomyopathy (PPCM) is a type of pregnancy‐associated heart failure with reduced ejection fraction of unknown aetiology that is associated with significant morbidity and mortality in young women. 2 , 3 , 4 , 5 It is a leading cause of peripartum maternal death in the United States, with mortality as high as 25% in some cohorts, 2 , 5 and accounts for approximately 50% of cases of maternal cardiogenic shock. 6 The incidence of PPCM ranges from 1:100 to 1:3000 births with high rates in Nigeria and Haiti. 2 , 3 , 4 , 5 Only 40–67% of patients recover left ventricular (LV) systolic function by 1 year. 7 , 8 , 9
Prior studies have found that severe LV systolic dysfunction and LV dilatation are risk factors for adverse outcomes in PPCM. 2 However, few studies have examined the prognostic value of right ventricular (RV) dysfunction and clinical predictors of RV dysfunction in this population. RV dysfunction has been found to be an important determinant of functional impairment and cardiovascular outcomes in patients with left‐sided heart failure in general. 10 , 11 , 12 In PPCM, RV dysfunction can be due to shared pathophysiology with LV myopathy or a result of increased RV afterload resulting from pulmonary hypertension associated with LV dysfunction.
Prior small cohort studies found that RV systolic dysfunction was an independent predictor of absence of LV systolic function recovery and greater need for mechanical support in PPCM. 13 , 14 , 15 In addition, pulmonary hypertension, which can be assessed via pulmonary artery systolic pressure (PASP), is known to result in diminished exercise capacity, and quality of life, and a high mortality in patients with left‐sided heart disease overall. 16 , 17 , 18 , 19 Even a slight increase in pulmonary artery pressures within the upper end of the traditional normal range has been found to correlate with greater mortality. 19 However, studies are lacking on the prognostic role of RV systolic dysfunction and pulmonary hypertension in PPCM. Additionally, the predictors and impact of RV dysfunction and pulmonary hypertension in PPCM are not known. Thus, the purpose of our study is (i) to determine predictors associated with RV dysfunction and elevated PASP on presentation and (ii) to examine whether RV dysfunction and pulmonary hypertension are associated with adverse outcomes (death, need for LV assist device, orthotopic heart transplantation, extracorporeal membrane oxygenation, or recurrent heart failure hospitalization) in patients with PPCM.
Methods
Study population
We conducted a multi‐centre retrospective study across three hospitals: Brigham and Women's Hospital, Massachusetts General Hospital, and Beth Israel Deaconess Medical Center to identify subjects with PPCM. Our study protocol was approved by the institutional review boards at each of the participating centres. The following criteria were used for diagnosis of PPCM: adult women (>18 years of age) with new‐onset LV systolic dysfunction (ejection fraction < 40%) at the time of diagnosis and development of heart failure within the last month of pregnancy or within 5 months of delivery and no other identifiable cause of heart failure with reduced ejection fraction. A total of 237 adult women from January 1993 to December 2017 met inclusion criteria for PPCM, of which 229 had RV function recorded at baseline. The cohort selection process has been previously described in detail. 9
Data collection
We collected the following data from electronic medical records (EMRs): demographic information—date of birth, date of admission, gravidity, parity, height, weight, race/ethnicity, past medical history, prenatal care and delivery information if available, social history including tobacco use, alcohol use, and illicit drug use, family history of cardiomyopathy, signs and symptoms of heart failure and therapies, transthoracic echocardiographic parameters: LV ejection fraction (LVEF), date of transthoracic echocardiography, LV end‐diastolic and end‐systolic volumes, RV systolic pressure (if sufficient tricuspid regurgitation jet was present), and RV function, vital signs and laboratory values (haemoglobin, haematocrit, creatinine, and magnesium potassium), and discharge information (including medications on discharge and follow‐up appointments).
Information was recorded for subsequent heart failure hospitalizations and follow‐up transthoracic echocardiograms. The date of last follow‐up was determined as (i) date of censoring (if an outcome event occurred such as death, LV assist device implantation, and orthotopic heart transplantation) OR (ii) date of last follow‐up by a healthcare provider in the EMR OR (iii) end of study. Data collection was performed by four independent physician reviewers. In addition, each chart was cross‐checked for accuracy and completion.
Outcomes
The outcome was a composite of adverse cardiovascular outcomes: (i) need for extracorporeal membrane oxygenation, (ii) ventricular assist device implantation, (iii) orthotopic heart transplantation, (iv) death, or (v) recurrent heart failure hospitalization. Recurrent heart failure hospitalization was defined as one or more hospitalizations occurring for acutely decompensated heart failure after the first month of diagnosis. Outcomes were extracted from the EMR chart review, validated by four independent physician reviewers, and cross‐checked for accuracy.
Covariates
Covariates were extracted from the EMR. Age (years) was recorded at the time of diagnosis—either date of admission or outpatient visit. Age was also categorized as younger than 35 years or ≥35 years for advanced maternal age, as per the American College of Obstetricians and Gynecologists and Society for Maternal‐Fetal Medicine clinical consensus guideline definition. 20 Race was characterized as White, African American, Haitian of African descent, African, American Indian/Alaskan Native, or Asian as recorded in the EMR. Gravidity was defined as the number of times the participant had been pregnant, and parity referred to the number of times the pregnancy was carried to viable gestational age. Body mass index was calculated from weight in kilograms over height in metres squared as recorded in the EMR prior to pregnancy. The diagnosis year refers to the calendar year during which the patient was diagnosed with PPCM. Past medical history (preeclampsia, hypertension, gestational hypertension, diabetes mellitus, and gestational diabetes), social history (smoking, alcohol, and illicit drug use), and family history were extracted from the EMR through medical physician or healthcare provider notes, discharge summaries, history and physical examination notes, outpatient clinic notes, emergency department assessments, and problem lists.
Vital signs including systolic blood pressure (mmHg), diastolic blood pressure (mmHg), heart rate (beats per minute), and temperature were recorded from the initial inpatient or outpatient visit. Laboratory values, including haematocrit (%), platelets (per microlitre), white blood cell count (cells per litre), serum creatinine (mg/dL), serum sodium (mEq/L or mmol/L), serum magnesium (mg/dL), and serum potassium (mEq/L), N‐terminal pro‐B‐type natriuretic peptide (NT‐proBNP) (pg/mL), and heart failure treatment (intravenous diuretics, intravenous inotropes, intravenous vasodilators, initiation, or increased dosage of oral diuretic) were recorded from EMR lab results, medication lists, and physician clinical notes.
Echocardiographic parameters (LVEF, end‐diastolic and end‐systolic volumes and RV systolic function, and RV systolic pressure) at baseline were extracted from index transthoracic echocardiogram reports in the medical record. RV systolic dysfunction was defined by tricuspid annular plane systolic excursion (TAPSE) < 16 mm or tissue Doppler‐derived tricuspid lateral annular systolic velocity S′ < 10 cm/s, or global visual assessment, and was interpreted by board‐certified cardiologists and echocardiographers based on contemporary guidelines for assessment of the right heart in adults. 21 PASP was extracted from echocardiogram reports by adding estimated right atrial pressure to the tricuspid regurgitation gradient using the modified Bernoulli equation [PASP = 4(TRV)2 + RAP]. 21 Participants did not have pulmonic stenosis or RV outflow obstruction.
Statistical analysis
Continuous variables were tested for normality using the Kolmogorov–Smirnov test and are presented as mean ± standard deviation (SD), whereas skewed variables are presented as median and interquartile range (IQR). Categorical variables are expressed as count and percentage. The independent t‐test or the paired samples t‐tests were used for continuous variables, and χ 2 tests were used for categorical variables. Logistic regression was used for univariable analysis of each predictor for the composite outcome. We examined the association of RV function and the composite outcome of adverse cardiovascular outcomes and recurrent heart failure hospitalization using multivariable models in Cox proportional hazards regression models. Clinical covariates for multivariable models were defined a priori. We created multivariable models adjusting for age, race, body mass index, history of preeclampsia, hypertension, diabetes mellitus, kidney disease, and LVEF at diagnosis.
To examine the association between PASP and RV function, we examined PASP as both a continuous and a categorical variable using logistic regression. We generated Kaplan–Meier survival curves to compare those with normal RV function and RV dysfunction and those with PASP of <30 and ≥30 mmHg using the PROC LIFETEST statement in the SAS software. The log‐rank test was used to determine whether there was a statistically significant difference between the groups. Time to event was defined as the time from the baseline echocardiogram on presentation to the event or date of last follow‐up. We conducted all analyses using SAS 9.4 software (SAS Institute, Cary, NC, USA). An alpha < 0.05 was considered statistically significant.
Results
From 1993 to 2017, a total of 229 women with PPCM and a median follow‐up of 3.4 years (IQR 1.0–8.8), 722 person‐years, and RV function noted on initial echocardiogram met criteria for inclusion. The study inclusion process has been previously described and is illustrated in Figure 1 . 9 In all, 237 women were eligible but 8 were missing RV function assessment, leaving 229 for analysis. Of these 229, 213 had 1 month or longer follow‐up. Participants had a mean age of 32.4 ± 6.8 years, 28% were Black women of African descent, and 50 (22%) had RV dysfunction on presentation. Baseline characteristics of participants by RV function are shown in Table 1 . Those with RV dysfunction on presentation were more likely to be of African descent (44% vs. 23%), have higher gravidity, were more likely to have current or former tobacco use (34% vs. 22%), were less likely to have received prenatal care (44% vs. 56%), and were more likely to have a family history of cardiomyopathy (22% vs. 8%). They were also more likely to give birth to an infant requiring the neonatal intensive care unit (20% vs. 14%). Those with RV dysfunction also had lower systolic blood pressure, higher heart rate, and higher NT‐proBNP levels on presentation. They were more likely to require intravenous diuretics (74% vs. 54%), inotropes (34% vs. 6%), vasodilators, and extracorporeal membrane oxygenation and ultimately undergo implantable cardioverter defibrillator (ICD) implantation (24% vs. 7%).
Figure 1.
Flow diagram illustrating the process of cohort selection.
Table 1.
Baseline characteristics of participants with peripartum cardiomyopathy by right ventricular function at baseline (n = 229)
| Characteristic a | Normal RV function (n = 179) | RV dysfunction (n = 50) |
|---|---|---|
| Age (years) | 32.4 ± 6.7 | 32.6 ± 7.0 |
| Advanced maternal age (>35 years) | 64 (35.8%) | 17 (34.0%) |
| Race | ||
| White | 117 (65.4%) | 24 (48.0%) |
| Black, African descent | 41 (22.9%) | 22 (44.0%) |
| Other b | 21 (11.7%) | 4 (8.0%) |
| Gravidity | 2.8 ± 2.3 | 3.4 ± 2.9 |
| Med (IQR): 2 (1–3) | Med (IQR): 2 (2–5) | |
| Parity | 1.5 ± 1.3 | 1.9 ± 2.1 |
| Med (IQR): 1 (1–2) | Med (IQR): 1 (1–2) | |
| Body mass index (kg/m2) | 31.0 ± 8.0 | 29.7 ± 6.8 |
| Diagnosis year | ||
| 1993–2004 | 5 | 0 |
| 2004–2007 | 57 | 14 |
| 2008–2011 | 58 | 13 |
| 2011–2017 | 59 | 23 |
| Medical history | ||
| Hypertension (prior to pregnancy) | 33 (18.4%) | 12 (24.0%) |
| Preeclampsia (current diagnosis) | 38 (21.2%) | 6 (12.0%) |
| Gestational diabetes mellitus | 15 (8.4%) | 5 (10.0%) |
| Tobacco use (former) | 39 (21.8%) | 17 (34.0%) |
| Ethanol use (prior to pregnancy) | ||
| None | 60 (33.5%) | 15 (30.0%) |
| Daily to weekly | 23 (12.8%) | 2 (4.0%) |
| Monthly or less | 54 (30.2%) | 23 (46.0%) |
| Illicit drug use (former) | 11 (6.1%) | 5 (10.0%) |
| Received prenatal care | 96 (53.6%) | 22 (44.0%) |
| Family history | ||
| Coronary artery disease | 37 (20.7%) | 8 (16.0%) |
| Diabetes mellitus | 26 (14.5%) | 5 (10.0%) |
| Hypertension | 38 (21.2%) | 9 (18.0%) |
| Cardiomyopathy | 14 (7.8%) | 11 (22.0%) |
| Delivery | ||
| Gestational age (weeks) | 36.8 ± 3.3 | 35.4 ± 4.4 |
| Vaginal delivery | 80 (44.7%) | 25 (50.0%) |
| Infant APGAR score | 7.6 ± 1.9 | 7.4 ± 2.3 |
| NICU stay | 25 (14.0%) | 10 (20.0%) |
| Systolic blood pressure (mmHg) | 137.0 ± 30.0 | 119.5 ± 27.1 |
| Heart rate at presentation (b.p.m.) | 93.9 ± 25.5 | 109.0 ± 24.8 |
| Laboratory markers | ||
| Haemoglobin | 10.8 ± 2.1 | 11.4 ± 1.7 |
| Haematocrit | 32.4 ± 6.1 | 35.2 ± 3.2 |
| Platelets | 330.3 ± 127.1 | 311.8 ± 151.4 |
| Serum creatinine | 0.8 ± 0.3 | 1.0 ± 0.5 |
| Serum sodium | 139.1 ± 3.6 | 140.0 ± 3.1 |
| NT‐proBNP | Med (IQR): 1794 (1066–3542) | Med (IQR): 3559 (2362–5768) |
| Troponin | 0.13 ± 0.24 | 0.09 ± 0.13 |
| Med (IQR): 0.01 (0.01–0.10) | Med (IQR): 0.01 (0.01–0.10) | |
| Heart failure treatment | ||
| Intravenous diuretics | 96 (53.6%) | 37 (74.0%) |
| Inotropes | 11 (6.1%) | 17 (34.0%) |
| Intravenous vasodilators | 18 (10.0%) | 12 (24.0%) |
| ECMO | 3 (1.7%) | 2 (4.0%) |
| Initiation or increase of oral diuretic | 18 (10.1%) | 8 (16.0%) |
| ICD implantation | 12 (6.7%) | 12 (24.0%) |
| Discharge information | ||
| Cardiology follow‐up | 165 (92.2%) | 43 (86.0%) |
| ACE inhibitor or ARB | 108 (60.3%) | 39 (78.0%) |
| Aspirin | 13 (7.3%) | 6 (12.0%) |
| Aldactone | 19 (10.6%) | 12 (24.0%) |
| Beta‐blocker | 89 (49.7%) | 36 (72.0%) |
| Digoxin | 15 (8.4%) | 17 (34.0%) |
| Diuretic | 70 (39.1%) | 28 (56.0%) |
ACE, angiotensin‐converting enzyme; APGAR, appearance, pulse, grimace, activity, and respiration; ARB, angiotensin receptor blocker; ECMO, extracorporeal membrane oxygenation; ICD, implantable cardioverter defibrillator; IQR, interquartile range; NICU, neonatal intensive care unit; NT‐proBNP, N‐terminal pro‐B‐type natriuretic peptide; RV, right ventricular.
Mean and standard deviation are shown for continuous variables. Median and interquartile range are shown for skewed variables. Column percentages are shown.
Other race includes Asian, American Indian/Alaska Native, Native Hawaiian or Pacific Islander, and unknown.
Table 2 shows the echocardiographic parameters by RV dysfunction on presentation. Mean LVEF for those with RV dysfunction was 23.1% ± 10.0 vs. 35.6% ± 11.9 for those with preserved RV function. Those with RV dysfunction were less likely to recover LV systolic function during follow‐up (46% vs. 68%). Mean PASP was higher in the RV dysfunction group [median (IQR): 37 mmHg (24–43)] vs. the group with preserved RV function [median (IQR): 26 (23–36)]. Women with RV dysfunction had greater baseline and follow‐up LV end‐diastolic and end‐systolic diameter. Out of 50 women with RV dysfunction on presentation, 28 (56%) recovered RV function at follow‐up.
Table 2.
Echocardiographic parameters of participants with peripartum cardiomyopathy by right ventricular function at baseline and follow‐up
| Baseline echocardiogram (n = 229) | ||
|---|---|---|
| Echocardiographic parameter | Normal RV function (n = 179) | RV dysfunction (n = 50) |
| LVEF | 35.6 ± 11.9 | 23.1 ± 10.0 |
| LVEDD | 5.4 ± 0.6 | 5.7 ± 0.9 |
| LVESD | 4.3 ± 0.8 | 4.8 ± 0.9 |
| PASP (mmHg) | 29.4 ± 10.3 | 35.5 ± 13.8 |
| Med (IQR): 26 (23–36) | Med (IQR): 37 (24–43) | |
| PASP ≥ 30 mmHg | 21 (11.7%) | 17 (34.0%) |
| Follow‐up echocardiogram (n = 158) | ||
|---|---|---|
| Echocardiographic parameter | Normal RV function (n = 135) | RV dysfunction (n = 23) |
| LVEF | 50.0 ± 14.8 | 40.0 ± 17.9 |
| LVEF recovery | 121 (67.6%) | 23 (46.0%) |
| LVEDD | 4.9 ± 0.8 | 5.2 ± 1.1 |
| LVESD | 3.6 ± 0.9 | 4.0 ± 1.4 |
| Normal RV function at baseline | 107 (79.2%) | 10 (43.5%) |
LVEDD, left ventricular end‐diastolic dimension; LVEF, left ventricular ejection fraction; LVESD, left ventricular end‐systolic dimension; PASP, pulmonary artery systolic pressure; RV, right ventricular.
Column percentages are shown.
Predictors associated with right ventricular dysfunction
Table 3 reveals the univariate predictors associated with RV dysfunction on presentation. Race [odds ratio (OR) 2.02, 95% confidence interval (CI): 1.07–3.82, P = 0.03] and family history of cardiomyopathy (OR 3.30, 95% CI: 1.39–7.84, P = 0.007) were statistically significant predictors of RV dysfunction on presentation. Supporting Information, Table S1 shows the clinical factors associated with RV dysfunction on presentation. Significant clinical factors included systolic blood pressure, heart rate, haematocrit, serum creatinine, NT‐proBNP, LVEF on presentation, ICD implantation, PASP, LV end‐diastolic and end‐systolic diameter, and need for intravenous diuretics and inotropes (Supporting Information, Table S1 ). To explore whether race and family history relate to RV dysfunction independent of their association with LVEF and elevated RV afterload, we performed additional analyses adjusting for LVEF or PASP (Supporting Information, Table S2 ). LVEF and PASP may be mediators of higher frequency of RV dysfunction among patients of African descent or with family history of cardiomyopathy. Although the ORs were in the same direction, the results from these analyses did not meet the threshold of significance: race + LVEF, OR 2.12 (95% CI: 0.99–4.60), P = 0.05, and race + PASP, OR 2.06 (95% CI: 0.74–5.76), P = 0.08.
Table 3.
Predictors associated with right ventricular dysfunction and hazard ratios (95% confidence interval) of composite outcome by right ventricular dysfunction on presentation
| Predictor | Odds ratio (95% CI) | P‐value |
|---|---|---|
| Age (years) | 1.00 (0.96–1.05) | 0.83 |
| Race (Black of African descent) | 2.02 (1.07–3.82) | 0.03 |
| Body mass index (kg/m2) | 0.98 (0.92–1.04) | 0.45 |
| Gravidity | 1.10 (0.97–1.24) | 0.14 |
| Parity | 1.17 (0.95–1.43) | 0.14 |
| Tobacco use (current or former) | 1.71 (0.84–3.48) | 0.14 |
| Preeclampsia history | 2.00 (0.79–5.02) | 0.14 |
| Hypertension | 1.39 (0.66–2.94) | 0.39 |
| Diabetes mellitus | 1.21 (0.42–3.50) | 0.73 |
| Family history of cardiomyopathy | 3.30 (1.39–7.84) | 0.007 |
| Gestational age | 0.91 (0.82–1.00) | 0.05 |
CI, confidence interval.
Multivariable analyses—Right ventricular dysfunction
During the follow‐up period, 10 women died from all causes, 5 underwent LV assist device implantation and ultimately orthotopic heart transplantation, 5 required extracorporeal membrane oxygenation on presentation, and 55 had recurrent heart failure hospitalizations. Table 4 shows the Cox proportional hazards models of the composite outcome by RV dysfunction on presentation. In the age‐adjusted model, RV dysfunction was associated with the composite outcome: hazard ratio (HR) 2.42 (95% CI: 1.31–4.50), P = 0.005. This association persisted in the multivariable model adjusting for age, race, body mass index, history of preeclampsia or hypertension, diabetes mellitus, kidney disease, and initial LVEF: HR 3.21 (95% CI: 1.11–9.28), P = 0.03.
Table 4.
Hazard ratios and 95% confidence intervals of composite outcome by right ventricular dysfunction on presentation
| Model | Hazard ratio (95% CI) | P‐value |
|---|---|---|
| Crude | 2.38 (1.29–4.38) | 0.006 |
| Age adjusted | 2.42 (1.31–4.50) | 0.005 |
| Multivariable model a | 3.21 (1.11–9.28) | 0.032 |
CI, confidence interval.
Multivariable model adjusts for age, body mass index, race, history of hypertension or preeclampsia, diabetes mellitus, kidney disease, and initial left ventricular ejection fraction.
Figure 2 A shows the Kaplan–Meier survival curves for those with preserved RV function and RV dysfunction. Those with preserved RV function on presentation were more likely to have survival free of adverse events. Figure 2 B demonstrates the survival curves for those with preserved RV function and RV dysfunction by African ancestry.
Figure 2.
(A, B) Kaplan–Meier curves for survival free from adverse cardiovascular outcomes by right ventricular (RV) function.
Pulmonary artery systolic pressure
As RV dysfunction was associated with increased RV afterload (PASP) likely due to LV dysfunction (Group 2 pulmonary hypertension), we studied the effect of elevated PASP (≥30 mmHg) on outcomes. In logistic regression, we found that PASP was associated with RV dysfunction: OR 1.05 (95% CI: 1.00–1.09), P = 0.013, using PASP as a continuous variable. Out of 50 women with PPCM who had RV dysfunction on presentation, only 17 (34%) had PASP ≥ 30 mmHg. Clinical characteristics of women with PPCM by PASP of ≥30 and <30 mmHg are shown in Supporting Information, Table S3 , and echocardiographic parameters are shown in Supporting Information, Table S4 . Those with PASP ≥ 30 mmHg were more likely to be of advanced maternal age (≥35 years), more likely to be of African descent, multiparous [median (IQR): 3 (2–6) vs. 1 (1–2)], more likely to have preeclampsia (29% vs. 17%), less likely to have received prenatal care (29% vs. 45%), and more likely to have greater NT‐proBNP levels and ultimately undergo ICD implantation (29% vs. 7%). In survival analysis, those with PASP < 30 mmHg were more likely to have survival free of adverse events (log‐rank P = 0.04) (Figure 3 ). We chose 30 mmHg as the cut‐off for PASP as prior studies have found that PASP ≥ 30 mmHg is associated with higher mortality. 18 , 19
Figure 3.
Kaplan–Meier curves for survival free from adverse cardiovascular outcomes by pulmonary artery systolic pressure (PASP).
Discussion
Our study found that women with PPCM who had RV dysfunction at diagnosis were three times more likely to develop a composite of adverse events independent of initial LV systolic function and comorbid conditions on multivariable analysis. Also, we show that African ancestry and family history of cardiomyopathy were predictors associated with RV systolic dysfunction on presentation. To our knowledge, this is the first study to examine predictors of RV dysfunction on presentation and to evaluate the association of pulmonary hypertension with outcomes in a large diverse cohort of patients with PPCM over long‐term follow‐up. Our findings show that women with PPCM who have RV dysfunction or pulmonary hypertension at presentation are a high‐risk group and may warrant monitoring for advanced therapies and guiding treatment.
Only a handful of studies have previously examined RV dysfunction in patients with PPCM and have been limited by sample size and follow‐up time. In the Investigations of Pregnancy‐Associated Cardiomyopathy (IPAC) study (n = 84), RV dysfunction was associated with lack of LV systolic function recovery and major adverse events such as death, transplant, or LV assist device implantation (HR 0.95, 95% CI: 0.92–0.99). 13 In a cohort of 53 patients, moderate to severe RV dysfunction was independently associated with a composite outcome of LV assist device implantation, cardiac transplantation, or death in a model adjusting for LVEF and LV end‐diastolic dimension (HR 3.21, 95% CI: 1.13–9.10). 14 In a retrospective analysis from the Bromocriptine in Heart Failure (BRO‐HF) study (n = 67), those with RV dysfunction had higher odds of the need for mechanical support (OR 10.10, 95% CI: 1.86–54.8) but not LVEF recovery, cardiac transplantation, heart failure hospitalization, or death at 6 months. 15 A study of 45 patients in Nigeria did not find a significant association with RV dysfunction and mortality at 6 months. 22 Another study of women with PPCM (n = 34) found that RV dysfunction was associated with lower chance of LVEF recovery at 5 months. 23 In contrast, our study (n = 229, multi‐centre retrospective cohort) found that RV dysfunction at diagnosis was associated with a composite of adverse outcomes including need for LV assist device, extracorporeal membrane oxygenation, orthotopic heart transplantation, death, or recurrent heart failure hospitalization (HR 3.21, 95% CI: 1.11–9.28) in a multivariable model including initial LVEF, with a median follow‐up of 3.4 years (IQR 1.0–8.8). Our findings are in the same direction as those of some prior studies, and the greater sample size allows for adequate power to examine the association in a multivariable model with adequate follow‐up, thus offering new insights. These results support the finding that RV systolic dysfunction is an independent predictor of adverse outcomes in PPCM.
We found that African ancestry was a significant predictor of RV dysfunction on presentation. In all, 44% of those with RV dysfunction were of Black African descent vs. 23% of those in the normal RV function group. Women with PPCM of African descent were twice as likely to have RV dysfunction on presentation, which may represent a more severe form of disease. The incidence of PPCM is high in women of African ancestry and carries a worse prognosis. 2 , 24 , 25 , 26 The reasons for this difference in presentation may be multifactorial including genetic differences, variation in phenotypic presentation, prevalence of cardiovascular risk factors such as chronic hypertension, diabetes mellitus, and obesity, healthcare disparities, and socioeconomic factors. African American women were more likely to be diagnosed later in the postpartum period and have LVEF < 30% and chronic hypertension. 26
Our study is the first to find family history of cardiomyopathy as a significant predictor of RV dysfunction on presentation. It was present in 22% of those with RV dysfunction and 8% of those with normal RV function on presentation. Genetic studies have demonstrated that 10% of women with PPCM have truncating variants in TTN (encoding for the sarcomeric protein titin), and some have overrepresentation of truncating variants in Filamen C, desmoplakin, and BAG3, similar to patients with non‐ischaemic dilated cardiomyopathy (DCM), suggesting a shared genetic predisposition between the two. 27 , 28 Those with TTN truncating variant had lower LVEF on presentation. 27 , 28 Clustering of PPCM and DCM has been reported in some families, and it is postulated that PPCM may even be part of a spectrum of DCM. 29 Thus, family history of cardiomyopathy is an important predisposing factor that may predict those likely to develop PPCM and a more severe form of disease. It may be considered as a screening question in pregnant patients. Genetic testing may benefit patients with PPCM for prognosis and risk stratification and may also provide benefit to family members. 30
Preclinical studies suggest that PPCM is likely a vascular disease triggered by hormonal changes that occur during late pregnancy and the postpartum period. 31 , 32 , 33 , 34 Pregnancy triggers the release of prolactin and soluble FLT1, leading to proteolytic conversion to 16 kDa PRL (which is toxic to the cardiac microvasculature), or inhibition of vascular endothelial growth factor signalling, which leads to cardiac ischaemia, metabolic insufficiency, and apoptosis. 31 However, the mechanism underlying the development of RV dysfunction on presentation in PPCM remains to be elucidated. Race and family history of cardiomyopathy were significant predictors, and LVEF and PASP were strong clinical factors associated with RV dysfunction in our study. It is likely that RV dysfunction may be related to left‐sided heart failure, pulmonary hypertension, or a genetic predisposition in some patients. Studies have also found differences between the LV and RV at the molecular level, which may in part explain predisposition to RV dysfunction in some patients. The ingenuity pathway analysis (a bioinformatics application integrating omics with genomic and clinical data) from the Human Heart Cell Atlas identified some molecular differences between the RV and LV including higher levels of sarcomeric proteins titin, beta oestradiol, and the ryanodine receptor in the RV. 35 Future studies are needed to further understand these genotype–phenotype correlations.
We found that women with PPCM who had PASP < 30 mmHg were more likely to have survival free from these adverse events as compared with those with PASP ≥ 30 mmHg. Although to our knowledge, no prior studies have evaluated pulmonary hypertension in patients with PPCM, it is known that pulmonary hypertension reduces exercise capacity, worsens prognosis, and is associated with a high mortality in patients with left‐sided heart disease in general. 16 , 17 , 18 , 19 Even a mild elevation in pulmonary artery pressures within the traditional normal range has been found to be associated with higher mortality. 18 , 19 Thus, women with PPCM represent a higher risk group and may warrant closer monitoring. In addition, patients who have severely impaired RV function may not be able to mount elevated pulmonary artery pressures. This represents a small subset of patients who also have a poor prognosis. 36
This study is limited in that it is a retrospective cohort study and residual confounding due to unmeasured variables cannot be excluded. Although the study is multi‐centre and we attempted to also include data from outside hospitals scanned into our electronic medical systems, we were likely not able to capture all potential healthcare utilization outside of our centres. RV dysfunction was assessed by an echocardiographer reading the echocardiograms for clinical indications. Since the start of the study (1993–2017), there have been advances in the echocardiographic assessment of the right heart; thus, it is possible that milder forms of RV dysfunction may not have been captured during the earlier years. However, there were only a handful of women with PPCM from the first decade in our cohort. In addition, we only had follow‐up echocardiograms available on 158 of the 229 participants, thus limiting our ability to examine RV dysfunction in the long term. Despite these limitations, our study has several strengths. We include one of the largest and diverse cohorts of PPCM with long‐term follow‐up, which allows for greater statistical power to examine associations and adequate follow‐up time to capture events. In addition, our study is the first to identify predictors of RV dysfunction prior to presentation and examine the association of pulmonary hypertension with outcomes in PPCM. Future studies with large racially diverse cohorts are needed to better understand the predictors of RV dysfunction in PPCM to identify a higher risk group who may benefit from further prognostication and early referral for advanced therapies and to better elucidate the genotype–phenotype correlation in this population.
Conflict of interest
The authors have no conflict of interest.
Funding
The study was funded by the National Institute of General Medical Sciences (NIGMS), Project Number: 5P20GM103652‐10, Sub‐project ID: 9499.
Supporting information
Table S1. Clinical factors associated with right ventricular dysfunction on presentation.
Table S2. The association of race and family history of cardiomyopathy with right ventricular dysfunction on presentation adjusting for LVEF and PASP.
Table S3. Clinical characteristics of women with peripartum cardiomyopathy by pulmonary artery systolic pressure (n = 229).
Table S4. Echocardiographic parameters of participants with peripartum cardiomyopathy by pulmonary artery systolic pressure at baseline and follow up (n = 229).
Imran, T. F. , Ataklte, F. , Khalid, M. , Lopez, D. , Mohebali, D. , Bello, N. A. , Gaziano, J. M. , Djousse, L. , Arany, Z. , Sabe, M. A. , French, K. , Poppas, A. , Wu, W.‐C. , and Choudhary, G. (2024) Clinical predictors of right ventricular dysfunction and association with adverse outcomes in peripartum cardiomyopathy. ESC Heart Failure, 11: 422–432. 10.1002/ehf2.14583.
References
- 1. Declercq E, Zephyrin L. Maternal Mortality in the United States: A Primer. Commonwealth Fund; 2020. doi: 10.26099/ta1q-mw24 [DOI] [Google Scholar]
- 2. Davis MB, Arany Z, McNamara DM, Goland S, Elkayam U. Peripartum cardiomyopathy: JACC state‐of‐the‐art review. J Am Coll Cardiol 2020;75:207‐221. doi: 10.1016/j.jacc.2019.11.014 [DOI] [PubMed] [Google Scholar]
- 3. Elkayam U. Clinical characteristics of peripartum cardiomyopathy in the United States: Diagnosis, prognosis, and management. J Am Coll Cardiol 2011;58:659‐670. doi: 10.1016/j.jacc.2011.03.047 [DOI] [PubMed] [Google Scholar]
- 4. Bauersachs J, König T, van der Meer P, Petrie MC, Hilfiker‐Kleiner D, Mbakwem A, et al. Pathophysiology, diagnosis and management of peripartum cardiomyopathy: A position statement from the Heart Failure Association of the European Society of Cardiology Study Group on peripartum cardiomyopathy. Eur J Heart Fail 2019. Jul;21:827‐843. doi: 10.1002/ejhf.1493 [DOI] [PubMed] [Google Scholar]
- 5. Arany Z, Elkayam U. Peripartum cardiomyopathy. Circulation 2016;133:1397‐1409. doi: 10.1161/CIRCULATIONAHA.115.020491 [DOI] [PubMed] [Google Scholar]
- 6. Banayan J, Rana S, Mueller A, Tung A, Ramadan H, Arany Z, et al. Cardiogenic shock in pregnancy: Analysis from the national inpatient sample. Hypertens Pregnancy 2016;1–7:117‐123. doi: 10.1080/10641955.2016.1242606 [DOI] [PubMed] [Google Scholar]
- 7. Elkayam U, Akhter MW, Singh H, Khan S, Bitar F, Hameed A, et al. Pregnancy‐associated cardiomyopathy: Clinical characteristics and a comparison between early and late presentation. Circulation 2005;111:2050‐2055. doi: 10.1161/01.CIR.0000162478.36652.7E [DOI] [PubMed] [Google Scholar]
- 8. Goland S, Modi K, Bitar F, Janmohamed M, Mirocha JM, Czer LS, et al. Clinical profile and predictors of complications in peripartum cardiomyopathy. J Card Fail 2009;15:645‐650. doi: 10.1016/j.cardfail.2009.03.008 [DOI] [PubMed] [Google Scholar]
- 9. Imran TF, Mohebali D, Lopez D, Goli RR, DeFilippis EM, Truong S, et al. NT‐proBNP and predictors of event free survival and left ventricular systolic function recovery in peripartum cardiomyopathy. Int J Cardiol 2022;357:48‐54. doi: 10.1016/j.ijcard.2022.03.052 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Raina A, Meeran T. Right ventricular dysfunction and its contribution to morbidity and mortality in left ventricular heart failure. Curr Heart Fail Rep 2018;15:94‐105. doi: 10.1007/s11897-018-0378-8 [DOI] [PubMed] [Google Scholar]
- 11. Gulati A, Ismail TF, Jabbour A, Alpendurada F, Guha K, Ismail NA, et al. The prevalence and prognostic significance of right ventricular systolic dysfunction in nonischemic dilated cardiomyopathy. Circulation 2013;128:1623‐1633. doi: 10.1161/CIRCULATIONAHA.113.002518 [DOI] [PubMed] [Google Scholar]
- 12. Juillière Y, Barbier G, Feldmann L, Grentzinger A, Danchin N, Cherrier F. Additional predictive value of both left and right ventricular ejection fractions on long‐term survival in idiopathic dilated cardiomyopathy. Eur Heart J 1997;18:276‐280. doi: 10.1093/oxfordjournals.eurheartj.a015231 [DOI] [PubMed] [Google Scholar]
- 13. Blauwet LA, Delgado‐Montero A, Ryo K, Marek JJ, Alharethi R, Mather PJ, et al. Right ventricular function in peripartum cardiomyopathy at presentation is associated with subsequent left ventricular recovery and clinical outcomes. Circ Heart Fail 2016;9:e002756. doi: 10.1161/CIRCHEARTFAILURE.115.002756 [DOI] [PubMed] [Google Scholar]
- 14. Peters A, Caroline M, Zhao H, Baldwin MR, Forfia PR, Tsai EJ. Initial right ventricular dysfunction severity identifies severe peripartum cardiomyopathy phenotype with worse early and overall outcomes: A 24‐year cohort study. J Am Heart Assoc 2018;7:e008378. doi: 10.1161/JAHA.117.008378 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Pacheco C, Tremblay‐Gravel M, Marquis‐Gravel G, Couture E, Avram R, Desplantie O, et al. Association between right ventricular dysfunction and adverse outcomes in peripartum cardiomyopathy: Insights from the BRO‐HF Quebec cohort study. CJC Open 2022;4:913‐920. doi: 10.1016/j.cjco.2022.05.004 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Maron BA, Hess E, Maddox TM, Opotowsky AR, Tedford RJ, Lahm T, et al. Association of borderline pulmonary hypertension with mortality and hospitalization in a large patient cohort: Insights from the veterans affairs clinical assessment, reporting, and tracking program. Circulation 2016;133:1240‐1248. doi: 10.1161/CIRCULATIONAHA.115.020207 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Kjaergaard J, Akkan D, Iversen KK, Kjoller E, Kober L, Torp‐Pedersen C, et al. Prognostic importance of pulmonary hypertension in patients with heart failure. Am J Cardiol 2007;99:1146‐1150. doi: 10.1016/j.amjcard.2006.11.052 [DOI] [PubMed] [Google Scholar]
- 18. Kolte D, Lakshmanan S, Jankowich MD, Brittain EL, Maron BA, Choudhary G. Mild pulmonary hypertension is associated with increased mortality: A systematic review and meta‐analysis. J Am Heart Assoc 2018;7:e009729. doi: 10.1161/JAHA.118.009729 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Strange G, Stewart S, Celermajer DS, Prior D, Scalia GM, Marwick TH, et al. Threshold of pulmonary hypertension associated with increased mortality. J Am Coll Cardiol 2019;73:2660‐2672. doi: 10.1016/j.jacc.2019.03.482 [DOI] [PubMed] [Google Scholar]
- 20. Gantt A, Metz TD, Kuller JA, Louis JM, Society for Maternal‐Fetal Medicine , Cahill AG, et al. Obstetric care consensus #11, Pregnancy at age 35 years or older. Am J Obstet Gynecol 2022;S0002‐9378(22)00576‐00572. [DOI] [PubMed] [Google Scholar]
- 21. Rudski LG, Lai WW, Afilalo J, Hua L, Handschumacher MD, Chandrasekaran K, et al. Guidelines for the echocardiographic assessment of the right heart in adults: A report from the American Society of Echocardiography endorsed by the European Association of Echocardiography, a registered branch of the European Society of Cardiology, and the Canadian Society of Echocardiography. J Am Soc Echocardiogr 2010;23:685‐713. doi: 10.1016/j.echo.2010.05.010 [DOI] [PubMed] [Google Scholar]
- 22. Karaye KM, Lindmark K, Henein M. Right ventricular systolic dysfunction and remodelling in Nigerians with peripartum cardiomyopathy: A longitudinal study. BMC Cardiovasc Disord 2016;16:27. doi: 10.1186/s12872-016-0204-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Haghikia A, Rontgen P, Vogel‐Claussen J, Schwab J, Westenfeld R, Ehlermann P, et al. Prognostic implication of right ventricular involvement in peripartum cardiomyopathy: A cardiovascular magnetic resonance study. ESC Heart Fail 2015;2:139‐149. doi: 10.1002/ehf2.12059 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Gentry MB, Dias JK, Luis A, Patel R, Thornton J, Reed GL. African‐American women have a higher risk for developing peripartum cardiomyopathy. J Am Coll Cardiol 2010;55:654‐659. doi: 10.1016/j.jacc.2009.09.043 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Goland S, Modi K, Hatamizadeh P, Elkayam U. Differences in clinical profile of African‐American women with peripartum cardiomyopathy in the United States. J Card Fail 2013;19:214‐218. doi: 10.1016/j.cardfail.2013.03.004 [DOI] [PubMed] [Google Scholar]
- 26. Irizarry OC, Levine LD, Lewey J, Boyer T, Riis V, Elovitz MA, et al. Comparison of clinical characteristics and outcomes of peripartum cardiomyopathy between African American and non‐African American women. JAMA Cardiol 2017;2:1256‐1260. doi: 10.1001/jamacardio.2017.3574 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. Goli R, Li J, Brandimarto J, Levine LD, Riis V, McAfee Q, et al. Genetic and phenotypic landscape of peripartum cardiomyopathy. Circulation 2021;143:1852‐1862. doi: 10.1161/CIRCULATIONAHA.120.052395 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. Ware JS, Li J, Mazaika E, Yasso CM, DeSouza T, Cappola TP, et al. Shared genetic predisposition in peripartum and dilated cardiomyopathies. N Engl J Med. 2016;374:233‐241. doi: 10.1056/NEJMoa1505517 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29. van Spaendonck‐Zwarts KY, van Tintelen JP, van Veldhuisen DJ, van der Werf R, Jongbloed JD, Paulus WJ, et al. Peripartum cardiomyopathy as a part of familial dilated cardiomyopathy. Circulation 2010;121:2169‐2175. doi: 10.1161/CIRCULATIONAHA.109.929646 [DOI] [PubMed] [Google Scholar]
- 30. Arany Z. It is time to offer genetic testing to women with peripartum cardiomyopathy. Circulation 2022;146:4‐5. doi: 10.1161/CIRCULATIONAHA.122.059177 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31. Arany Z, Hilfiker‐Kleiner D, Karumanchi SA. Animal models of cardiovascular complications of pregnancy. Circ Res 2022;130:1763‐1779. doi: 10.1161/CIRCRESAHA.122.320395 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32. Bello NA, Arany Z. Molecular mechanisms of peripartum cardiomyopathy: A vascular/hormonal hypothesis. Trends Cardiovasc Med 2015;25:499‐504. doi: 10.1016/j.tcm.2015.01.004 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33. Hilfiker‐Kleiner D, Kaminski K, Podewski E, Bonda T, Schaefer A, Sliwa K, et al. A cathepsin D‐cleaved 16 kDa form of prolactin mediates postpartum cardiomyopathy. Cell 2007;128:589‐600. doi: 10.1016/j.cell.2006.12.036 [DOI] [PubMed] [Google Scholar]
- 34. Patten IS, Rana S, Shahul S, Rowe GC, Jang C, Liu L, et al. Cardiac angiogenic imbalance leads to peripartum cardiomyopathy. Nature 2012;485:333‐338. doi: 10.1038/nature11040 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35. Prisco SZ, Kazmirczak F, Thenappan T, Prins KW. Ingenuity pathway analysis of the human cardiac cell Atlas identifies differences between right and left ventricular cardiomyocytes. Pulm Circ 2022;12:e12011. doi: 10.1002/pul2.12011 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36. Frea S, Pidello S, Bovolo V, Iacovino C, Franco E, Pinneri F, et al. Prognostic incremental role of right ventricular function in acute decompensation of advanced chronic heart failure. Eur J Heart Fail 2016;18:564‐572. doi: 10.1002/ejhf.504 [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Table S1. Clinical factors associated with right ventricular dysfunction on presentation.
Table S2. The association of race and family history of cardiomyopathy with right ventricular dysfunction on presentation adjusting for LVEF and PASP.
Table S3. Clinical characteristics of women with peripartum cardiomyopathy by pulmonary artery systolic pressure (n = 229).
Table S4. Echocardiographic parameters of participants with peripartum cardiomyopathy by pulmonary artery systolic pressure at baseline and follow up (n = 229).