Impact of mode of delivery on maternal and fetal outcomes in pregnant patients with heart failure or pulmonary hypertension

Abstract

Background

Data on outcomes associated with various modes of delivery in pregnant patients with heart failure (HF) or pulmonary hypertension (pHTN) are limited.

Objective

We aim to investigate the association between mode of delivery on maternal and fetal outcomes in pregnant patients with HF or pHTN by conducting a multicenter, large scale and nationwide retrospective cohort study.

Methods

This retrospective population-based cohort study used the Nationwide Readmission Database to identify all hospitalized pregnant patients who were primarily admitted for vaginal or cesarean delivery from 2011 to 2019. Pregnant patients with HF or pHTN were identified. The primary outcomes were in-hospital maternal and fetal mortality.

Results

23,359 hospitalized pregnant patients with HF or pHTN were identified: 7885 (33.8 %) underwent vaginal delivery, and 15,474 (66.2 %) underwent cesarean delivery. Patients undergoing cesarean delivery had lower odds of fetal mortality than vaginal delivery (OR: 0.18, 95 % CI: 0.07–0.45; p < 0.001). However, the odds of in-hospital maternal mortality (OR: 2.0, 95 % CI: 1.07–3.77; p = 0.03) were higher in the cesarean group than vaginal delivery group. When considering vaginal delivery as a control group, 167 pregnant women would have to undergo cesarean delivery to have one in-hospital maternal mortality and 164 pregnant women would have to undergo cesarean delivery to reduce one fetal mortality.

Conclusions

In this multicentric cohort study, cesarean delivery was associated with lower fetal mortality but higher in-hospital maternal mortality. Thus, the unique features of each patient case should be considered via a multidisciplinary team before determining the optimal delivery method for patients with HF or pHTN.

1. Introduction

Cardiovascular disease (CVD) and pregnancy overlap represent a unique milieu with a significantly elevated risk for adverse fetal, maternal, and obstetrical outcomes [1]. The Centers for Disease Control (CDC) established the Pregnancy Mortality Surveillance System in 1987 to collect data regarding trends in pregnancy-related mortalities. From 2016–2018, cardiovascular conditions and cardiomyopathy represented the first and third leading etiologies of pregnancy-related death, constituting 16.2 % and 12.5 % of deaths, respectively [2]. Consistent with these data, a recent study on nationwide data from 2010 to 2019 in the United States (US) showed an increasing age-adjusted trend of CVD in pregnant women, with the presence of CVD associated with a 15-fold higher odds of in-hospital mortality in compared with those without CVD [3]. Despite the significant impact of CVD on outcomes in pregnant women, data and guidelines to optimize both fetal and maternal outcomes remain sparse. In 2018, the European Society of Cardiology (ESC) published guidelines for managing CVD during pregnancy [4]. The recommended mode of delivery endorsed by these guidelines was vaginal whenever possible, with cesarean delivery (C-section) reserved primarily for obstetric indications [4]. Vaginal deliveries were endorsed due to less blood loss and a lower risk of surgical site infections. A C-section approach was recommended for pre-term labor in the setting of oral anticoagulant use, aggressive aortic pathology, severe pulmonary hypertension (pHTN), or intractable heart failure (HF) [4]. Nonetheless, this suggestion primarily relied on expert consensus, with scant scientific evidence to substantiate it. The safest mode of delivery for mother and fetus has not been established in mothers with cardiac disease, particularly in the highest risk cardiac conditions such as HF (cardiomyopathy) or pHTN. The ESC Guidelines on pregnancy and heart disease acknowledge the limitations of the current available data and the complexity of the decisions made in individual case scenarios. Hence, we sought to investigate the association between mode of delivery on maternal, fetal, and obstetric outcomes in pregnant patients with HF or pHTN.

2. Methods

2.1. Data source

This study utilized the nationwide readmission database (NRD) from 2011 to 2019. The NRD consists of all-payer databases for the Healthcare Cost and Utilization Project, established by the Agency for Healthcare Research and Quality. It comprises discharge records from 31 geographically dispersed states, with approximately 32 million weighted discharges annually (excluding rehabilitation and long-term acute care facilities), regardless of the payer for the hospital stay. The NRD represents approximately 62.2 % of the US resident population and 60.8 % of all US hospitalizations. The study was exempt from the University of Kansas institutional review board approval as the database contained de-identified data sets with prior ethics committee approval. Informed consent was not required because data were deidentified. This study followed the reporting guidelines specified by the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines (Supplemental Table-1) [5].

2.2. Study population

We queried the NRD from 2011 to 2019 using the International Classification of Diseases (ICD)- 9 and 10 Clinical Modification codes to identify all pregnant patients who were primarily admitted for vaginal delivery or cesarean delivery. We applied ICD codes to identify patients with HF or pHTN (Supplemental Table-2). We excluded patients with a period of gestation ≤28 weeks, those with intrauterine fetal demise, or who were hospitalized for abortion (termination of pregnancy) (n = 1803). After exclusion, we reached our final cohort of 23,359 pregnant patients with HF or pHTN admitted for delivery. We divided the cohort into those who underwent vaginal delivery or C-section using appropriate ICD procedure codes (Supplemental Table-2). Vaginal delivery was defined as a composite of normal vaginal delivery and assisted vaginal delivery.

2.3. Patient and hospital characteristics

We used NRD variables to identify patient ages (in years), median household income in quartiles, type of insurance (Medicare, Medicaid, or private insurance), and day of admission (weekend/weekday). ICD- 9 & 10 CM codes were used to define the prior history of stroke/transient ischemic attack (TIA), ischemic and non-ischemic cardiomyopathy, atrial fibrillation, valvular heart disease or presence of prosthetic valves, congenital heart disease (CHD), history of percutaneous coronary intervention (PCI), history of coronary artery bypass graft (CABG), and history of defibrillator/pacemaker placement. Obstetrical comorbidities including antepartum hemorrhage, gestational hypertension, pre-eclampsia, eclampsia, gestational diabetes, multiparity, nulliparity, HELLP (Hemolysis, Elevated Liver enzymes, and Low Platelets) syndrome, multiple gestations, and birth defects (non-CHD) were defined using ICD- 9 & 10 CM diagnosis codes (Supplemental Table-2). Elixhauser comorbidities were used to define hypertension, diabetes, hyperlipidemia, valvular heart disease, smoking, alcohol abuse, obesity, chronic pulmonary disease, renal failure, liver disease, cancer, fluid and electrolyte imbalance, coagulopathy, use of steroids, use of anticoagulants, hormonal replacement therapy, thrombophilia, autoimmune disorders, and connective tissue diseases. Hypertension included chronic hypertension and pregnancy-related hypertension (gestational hypertension, pre-eclampsia, and eclampsia). Similarly, a composite of diabetes was defined by combining pre-existing diabetes and gestational diabetes variables. Composite patient characteristics have been defined in Supplemental Table-3. Under hospital characteristics, we included hospital size according to the number of beds (small/medium/large), teaching status of the hospital (teaching/non-teaching), and location of the hospital (urban/rural).

2.4. Study outcomes

The co-primary outcomes for this study were in-hospital maternal and fetal mortality. The secondary outcomes included other in-hospital outcomes and 6-month readmission outcomes. The other in-hospital outcomes of interest were acute coronary syndrome (ACS), acute stroke, deep vein thrombosis (DVT)/pulmonary embolism (PE), and obstetrical outcomes, including postpartum hemorrhage (PPH), and puerperal sepsis. ACS was defined as a composite of ST-elevation myocardial infarction (STEMI), non-ST-elevation myocardial infarction (NSTEMI), and unstable angina (UA). 6-month readmission outcomes included readmissions due to HF, stroke, DVT/PE, and ACS, as well as all-cause mortality during readmission. We defined readmission as hospital admission for any principal diagnosis from the discharge date of the index admission. In the case of multiple readmissions, only the first readmission from the index admission discharge date was counted. Mortality over 6 months includes death during hospitalization only, as out-of-hospital death is not registered in NRD databases.

2.5. Statistical analysis

Descriptive statistics are presented as numbers with percentages for categorical variables. Means and standard deviations (SD) are used to report continuous variables. Baseline patient and hospital characteristics were compared between vaginal delivery and cesarean delivery using the chi-square test for categorical variables and the Student’s T-test or Wilcoxon rank-sum test for continuous variables. Logistic regression analysis was used to calculate the odds ratio (OR) with a 95 % confidence interval (CI) for in-hospital and 6-month readmission outcomes. A multivariable logistic regression analysis was done to adjust for potential confounders (list of the adjusted variables is presented in Supplemental Table-4). Factors adjusted in the model were selected based on both forward and backward selection models and based on clinical significance. A subgroup analysis was also performed for patients with fetal distress, HF and pHTN, separately. The number needed to harm (NNH) and number needed to treat (NNT) were calculated for in-hospital maternal and fetal mortality, respectively, using vaginal delivery as a control group. All statistical analyses were performed with appropriate weighting, stratification, and clustering of samples using the svy package in STATA, version 17 (StataCorp., College Station, TX, USA). This approach ensured the production of nationally representative, unbiased results, variance estimates, and P-values. A two-sided P-value of <0.05 was considered statistically significant.

2.6. Sensitivity analysis

To assess the robustness of our findings and mitigate potential confounding, we conducted a sensitivity analysis using propensity score matching (PSM) for the primary outcomes of in-hospital maternal and fetal mortality. Propensity scores were estimated using a logistic regression model, incorporating clinically relevant baseline characteristics and potential confounders significantly associated with mode of delivery or outcomes, including age, hypertension, diabetes, obesity, HELLP syndrome, multiple gestations, and other Elixhauser comorbidities (Supplemental Table-4). Variables were selected based on their established clinical relevance to maternal and fetal outcomes in high-risk pregnancies or their significant association with mode of delivery in univariate analyses or prior literature. A 1:1 nearest-neighbor matching algorithm with a caliper width of 0.2 standard deviations of the logit of the propensity score was applied to create balanced cohorts. After matching, we assessed the balance of covariates between the two groups using standardized mean differences (SMD), considering an SMD of ≤10 % as indicative of adequate balance (Supplemental Table 5). Variables with an SMD >10 % after matching (age, hypertension, diabetes, obesity, HELLP syndrome and multiple gestations) were further adjusted in a multivariate logistic regression model to account for residual confounding.

3. Results

We identified a total of 23,359 hospitalized pregnant patients with HF or pHTN who were admitted primarily for delivery between 2011 and 2019. Among them, 15,935 patients had HF and 3120 patients had pHTN. Of this cohort, 7885 (33.8 %) underwent vaginal delivery, and 15,474 (66.2 %) underwent C-section. Of the 15,935 patients with HF, 5315 (33.4 %) patients underwent vaginal delivery, and 10,620 (66.6 %) patients underwent c-section. Out of 3120 patients with pHTN, 1167 (37.4 %) patients underwent vaginal delivery, and 1953 (62.6 %) patients underwent c-section.

3.1. Baseline characteristics

Table 1 shows the baseline characteristics of pregnant patients with HF or pHTN who underwent vaginal or cesarean delivery. In this cohort, cesarean deliveries were performed more than vaginal deliveries (66.2 % vs. 33.8 %). Patients who had cesarean delivery were older than vaginal delivery (mean age in years [SD]: 29.8 [0.25] vs. 31.1 [0.15], p < 0.001). Patients who had a cesarean delivery had a higher prevalence of hypertension, diabetes mellitus, hyperlipidemia, atrial fibrillation/flutter, chronic kidney disease (CKD), vasculitis, and psychiatric disorders but a lower prevalence of a history of pacemaker or implantable cardioverter defibrillator (ICD) device implantation, non-ischemic cardiomyopathy, or documentation of an alcohol abuse disorder. Cesarean delivery was performed more frequently in patients with more antenatal and intrapartum obstetrical complications such as antepartum hemorrhage (1.9 % vs. 4.3 %, p < 0.001), HELLP syndrome (3.7 % vs. 9.5 %, p < 0.001), and multiple gestations (2.5 % vs. 6.6 %, p < 0.001). Additionally, cesarean deliveries were performed more frequently over the weekdays, and vaginal deliveries were performed more frequently over the weekends. Both vaginal delivery and cesarean delivery groups had a relatively equal distribution of comorbidities like chronic liver disease, chronic pulmonary disease, smoking, drug abuse, valvular heart diseases or presence of prosthetic heart valves, autoimmune diseases, other connective tissue disorders, thrombophilia, stroke/TIA, CHD, ischemic heart disease, and birth defects (non-CHD).

Table 1.

Baseline characteristics of pregnant patients with heart failure or pulmonary hypertension who delivered during hospitalization.

	Pregnant patients with HF/pHTN (n = 23,359)
	Vaginal delivery (n = 7885)	Cesarean delivery (n = 15,474)
Age (mean ± SD), years	29.8 ± 0.25	31.1 ± 0.15
Hypertension	3588 (45.5)	9637 (62.3)
Diabetes	664 (8.4)	2469 (16)
Hyperlipidemia	132 (1.7)	404 (2.6)
Obesity	1358 (17.2)	4328 (28)
Atrial fibrillation/flutter	139 (1.8)	370 (2.4)
CKD	155 (2.0)	556 (3.6)
History of ICD/Pacemaker placement	501 (6.4)	592 (3.8)
Chronic liver disease	145 (1.8)	333 (2.2)
Chronic pulmonary disease	2256 (28.6)	4299 (27.8)
Smoker	1672 (21.2)	3087 (19.9)
Drug abuse	720 (9.1)	1299 (8.4)
Alcohol use	68 (0.9)	70 (0.5)
Valvular heart diseases	957 (12.1)	1830 (11.8)
Autoimmune disorders	172 (2.2)	371 (2.4)
Other connective tissue disorders	45 (0.6)	101 (0.6)
Thrombophilia	116 (1.5)	204 (1.3)
Vasculitis	5 (0.06)	54 (0.3)
Stroke/TIA	173 (2.2)	248 (1.6)
Assisted reproductive technology	11 (0.1)	65 (0.4)
Psychiatric disorders	29 (0.4)	116 (0.7)
Prevalence of various CVDs
Congenital heart disease	439 (5.4)	624 (4)
Simple	253 (3.1)	371 (2.4)
Moderate	124 (1.5)	196 (1.3)
Complex	62 (0.8)	57 (0.4)
Non-ischemic cardiomyopathy	3708 (47)	6878 (44.4)
Ischemic heart disease	285 (3.6)	542 (3.5)
Obstetric conditions
Antepartum Hemorrhage	148 (1.9)	660 (4.3)
Multiparity	1057 (13.4)	2269 (14.7)
Nulliparity	118 (1.5)	486 (3.1)
HELLP syndrome	289 (3.7)	1472 (9.5)
Multiple gestations	200 (2.5)	1022 (6.6)
Birth defects (non-CHD)	127 (1.6)	273 (1.8)
Insurance
Medicare	374 (4.7)	910 (5.9)
Medicaid	4262 (54.1)	8221 (53.1)
Private	2852 (36.2)	5725 (37)
Self-pay	119 (1.5)	216 (1.4)
Income quartiles
<25th	2899 (36.8)	5916 (38.2)
26–50th	2006 (25.4)	3919 (25.3)
51–75th	1752 (22.2)	3300 (21.3)
76–100th	1170 (14.8)	2184 (14.1)
Day of admission
Weekday	6345 (80.5)	13,030 (84.2)
Weekend	1540 (19.5)	2445 (15.8)
Teaching status of urban hospital
Teaching	6366 (80.7)	12,241 (79.1)
Non-teaching	1199 (15.2)	2576 (16.6)
Hospital bed size
Small/Medium	2357 (29.9)	4518 (29.2)
Large	5528 (70.1)	10,957 (70.8)
Hospital location
Urban	4836 (61.3)	9927 (64.2)
Non-urban	3048 (38.7)	5548 (35.9)

*N = number of patients; SD = standard deviation; HF = heart failure, pHTN = pulmonary hypertension; CHD = congenital heart disease; CKD = chronic kidney disease; TIA = transient ischemic attack; NSAIDs = non-steroidal anti-inflammatory drugs; HRT = hormonal replacement therapy; CVDs = cardiovascular diseases; HELLP syndrome = hemolysis, elevated liver enzymes and low platelet counts.

*Vaginal delivery = composite of normal vaginal delivery and assisted vaginal delivery; Hypertension = composite of pre-existing hypertension and pregnancy-induced hypertension (pre-eclampsia and eclampsia); Diabetes = composite of pre-existing diabetes and gestational diabetes; Valvular heart diseases = composite of valvular heart diseases and presence of prosthetic valves; Other connective tissue diseases = composite of dermatopolymyositis, systemic sclerosis, and other systemic involvement of connective tissue disorders.

* Four out of six major categories of types of insurance from HCUP are presented. Medicare includes both fee-for-service and managed care Medicare patients. Medicaid includes both fee-for-service and managed care Medicaid patients. Private insurance includes Blue Cross, commercial carriers, and private HMOs and PPOs. (https://www.hcup-us.ahrq.gov/db/vars/pay1/nrdnote.jsp);.

* Median household income of residents within the patients’ zip code are presented in quartiles (https://www.hcup-us.ahrq.gov/db/vars/zipinc_qrtl/nrdnote.jsp);.

* Day of admission is categorized by HCUP as an admission on weekday or weekend. (https://www.hcup-us.ahrq.gov/db/vars/aweekend/nrdnote.jsp).

* Bed size of hospital is divided into small/medium, and large, so that approximately one-third of the hospitals in a given region, location, and teaching status combination would fall within each bed size category (https://www.hcup-us.ahrq.gov/db/vars/hosp_bedsize/nrdnote.jsp);.

* Teaching hospital is defined as a hospital which has an American Medical Association–approved residency program (https://www.hcup-us.ahrq.gov/db/vars/hosp_ur_teach/nrdnote.jsp);.

* Urban-nonurban categorization of a hospital is defined by the American Hospital Association and includes four categories that differentiate between large and small metropolitan, micropolitan, and a non-urban residual. (https://www.hcup-us.ahrq.gov/db/vars/hosp_urcat4/nrdnote.jsp). We have combined the categories into two broader categories: Urban and non-urban hospital.

3.2. In-Hospital Outcomes

Among hospitalized pregnant patients with HF or pHTN, 30 (0.38 %) patients died after undergoing vaginal delivery and 152 (0.98 %) patients died after undergoing cesarean delivery. Moreover, there were 75 (0.95 %) fetus deaths with vaginal delivery and 52 (0.34 %) fetus deaths with cesarean delivery. Cesarean delivery was associated with lower odds of fetal mortality than vaginal delivery (cesarean delivery: 0.34 % vs. vaginal: 0.95 %, OR: 0.18, 95 % CI: 0.07–0.45; p < 0.001) (Fig. 1A). However, it was associated with a higher odds of in-hospital maternal mortality (cesarean delivery: 0.98 % vs. vaginal: 0.38 %, OR: 2.0, 95 % CI: 1.07–3.77; p = 0.03) (Fig. 1A), ACS (cesarean delivery: 1.95 % vs. vaginal: 0.88 %, OR: 2.17, 95 % CI: 1.44–3.27; p < 0.001) and puerperal sepsis (cesarean delivery: 2.2 % vs. vaginal: 1.3 %, OR: 1.55, 95 % CI: 1.09–2.21; p = 0.01) compared with vaginal delivery. Cesarean delivery was associated with lower odds of PPH compared with vaginal delivery (cesarean delivery: 6.9 % vs. vaginal: 8.4 %, OR: 0.73, 95 % CI: 0.61–0.87; p < 0.001). There were no differences in the odds of in-hospital stroke and DVT/PE between the two groups. We found that 167 women with HF or pHTN would have to undergo cesarean delivery to result in one in-hospital maternal death (NNH=167), and 164 women with HF or pHTN would have to undergo cesarean delivery to decrease one fetal mortality (NNT=164). (Table 2) (Fig. 1B).

Fig. 1.

Comparison of clinical outcomes between vaginal delivery versus cesarean delivery in pregnant patients with heart failure or pulmonary hypertension who delivered during hospitalization.

(A) Comparison of in-hospital maternal and fetal mortality

(B) Comparison of in-hospital outcomes

Abbreviations: DVT/PE: deep venous thrombosis/pulmonary embolism.

Table 2.

Comparison of in-hospital outcomes between vaginal delivery versus cesarean delivery among pregnant patients admitted for delivery.

Clinical outcomes	Vaginal (n = 7885)	CD (n = 15,474)	p-value		CD (Vaginal delivery is reference)
	N ( %)	N ( %)			Adjusted OR	95 % CI		p-value
Maternal outcomes				NNH
In-hospital maternal mortality	30 (0.38)	152 (0.98)	0.001	167	2.00	1.07	3.77	0.03
ACS	69 (0.88)	302 (1.95)	<0.001		2.17	1.44	3.27	<0.001
Acute stroke	98 (1.2)	261 (1.7)	0.13		1.14	0.76	1.71	0.53
DVT/PE	51 (0.6)	131 (0.8)	0.21		0.96	0.61	1.53	0.86
Obstetrical outcomes
PPH	660 (8.4)	1062 (6.9)	0.006		0.73	0.61	0.87	<0.001
Puerperal sepsis	104 (1.3)	339 (2.2)	0.002		1.55	1.09	2.21	0.01
Fetal outcomes				NNT
Fetal mortality	75 (0.95)	52 (0.34)	<0.001	164	0.18	0.07	0.45	<0.001
Length of stay in days (median, IQR)	3 (2–5)	5 (4–9)	N/A	N/A	N/A	N/A	N/A	N/A

*CD = cesarean delivery; OR = odds ratio; CI = confidence interval; ACS = acute coronary syndrome; DVT/PE = deep vein thrombosis / pulmonary embolism; PPH = post-partum hemorrhage.

*Number needed to treat (NNT), and number needed to harm (NNH) were calculated for maternal and fetal mortality with vaginal delivery taken as a control group and cesarean delivery taken as a treatment group.

*A multivariable logistic regression analysis was done to adjust for potential confounders (list of the adjusted variables is presented in Supplemental Table-4).

Among pregnant patients with HF only, cesarean delivery was associated with lower odds of fetal mortality (OR: 0.25, 95 % CI: 0.08–0.72; p = 0.01) and higher odds of in-hospital maternal mortality (OR: 3.09, 95 % CI: 1.44–6.65; p = 0.004) compared with vaginal delivery. Among pregnant patients with pHTN only, cesarean delivery was associated with lower odds of fetal mortality (OR: 0.04, 95 % CI: 0.003–0.54; p = 0.02) compared with vaginal delivery, while there was no statistically significant difference in the odds of maternal mortality between the two groups (OR: 0.7, 95 % CI: 0.23–2.09; p = 0.52). (Table 3).

Table 3.

Subgroup analysis in heart failure and pulmonary hypertension between vaginal delivery versus cesarean delivery.

Pregnant patients with heart failure (N = 15,931)
Clinical outcomes	Vaginal (n = 5315)	CS (n = 10,617)	P-value		CS (Vaginal delivery is reference)
	N ( %)	N ( %)		NNH	Adjusted OR	CI		P
In-hospital maternal mortality	17 (0.32)	121 (1.14)	<0.001	122	3.09	1.44	6.65	0.004
				NNT
Fetal mortality	53 (1)	44 (0.41)	0.006	170	0.25	0.08	0.72	0.01
Pregnant patients with pulmonary hypertension (N=7425)
	Vaginal (n=2570)	CS (n=4855)	P-value		CS (Vaginal delivery is reference)
	N ( %)	N ( %)		NNH	Adjusted OR	CI		P
In-hospital maternal mortality	13 (0.5)	31 (0.64)	0.62	714	0.7	0.23	2.09	0.52
				NNT
Fetal mortality	21 (0.28)	<11 (0.16)	0.002	833	0.04	0.003	0.54	0.02

* CD = cesarean delivery; OR = odds ratio; CI = confidence interval.

* Number needed to treat (NNT), and number needed to harm (NNH) were calculated for maternal and fetal mortality with vaginal delivery taken as a control group and cesarean delivery taken as a treatment group.

*A multivariable logistic regression analysis was done to adjust for potential confounders (list of the adjusted variables is presented in Supplemental Table-4).

A subgroup analysis of patients with fetal distress showed that patients undergoing cesarean delivery had significantly lower rates of fetal mortality as compared with those undergoing vaginal delivery (cesarean delivery: 0.1 % vs. vaginal: 0.6 %, p = 0.04), while there was no statistically significant difference in rates of maternal mortality between the two groups (cesarean delivery: 0.7 % vs. vaginal: 0.1 %, p = 0.06). (Supplemental Table-6)

3.3. Six-month readmission outcomes

There were increased odds of all-cause readmissions within 6 months in patients who had cesarean delivery compared with those who had vaginal deliveries (Cesarean: 14.4 % vs. vaginal: 11.0 %, OR: 1.17, 95 % CI: 1.02–1.35; p = 0.03). There was no difference in risk of acute exacerbation of heart failure, readmission mortality, ACS, stroke, and DVT/PE between the two groups at 6 months. (Supplemental Table-7)

3.4. Sensitivity analysis

After propensity score matching and multivariable regression analysis, the primary outcomes of in-hospital maternal mortality (OR: 2.33, 95 % CI: 1.06–5.13; p = 0.04) and fetal mortality (OR: 0.16, 95 % CI: 0.06–0.43; p < 0.001) did not change. (Supplemental Table-8)

4. Discussion

In this large, real-world, retrospective cohort study of in-hospital and 6-month maternal, fetal, and obstetrical outcomes in gravid females with HF or pHTN stratified by mode of delivery, we showcase the following important findings: (1) cesarean delivery was associated with a higher odds of in-hospital maternal mortality and a lower odds of fetal mortality; 2) cesarean delivery was associated with a lower odds of PPH, but a higher odds of puerperal sepsis and inhospital ACS, as well as greater all-cause readmission rates when compared with vaginal delivery; and (3) the higher maternal mortality seen in cesarean delivery was only seen in women with HF but not in women with pHTN. (Fig. 2).

Fig. 2.

Graphical abstract: Impact of mode of delivery on maternal and fetal outcomes in pregnant patients with heart failure or pulmonary hypertension.

Maternal cardiac disease is associated with higher rates of maternal and fetal mortality, but the current guidelines do not recommend one mode of delivery over another [4]. Nonetheless, women with cardiac disease have greater odds of undergoing cesarean delivery [6,7]. In the Registry of Pregnancy And Cardiac disease (ROPAC) participants, cesarean delivery was planned for 31 % of deliveries, which was 160 % greater than the expected rate in patients without cardiac disease [7]. Gavin et al. observed a cesarean delivery rate of 42.2 %, with the majority of indications for cesarean delivery being obstetric, such as repeat elective cesarean delivery and concerning fetal heart tracing [6]. Similarly, a prospective cohort study by Easter et al. at a single tertiary care center of 276 women with cardiovascular disease reported a cesarean delivery rate of 23.9 %, with all cesarean delivery indications being obstetric rather than related to maternal cardiac disease [8]. Using the NRD, we observed a higher rate of cesarean delivery in women with HF or pHTN at 66.2 %, which suggests our cohort may be of higher obstetrical risk. This is reflected by the higher rate of cesarean delivery in patients with high-risk pregnancies.

Amongst 894 ROPAC participants with all forms of cardiovascular disease, after propensity matching there was no statistically significant difference in maternal mortality observed after stratification for mode of delivery, most likely secondary to the small study cohort (2.7 % with vaginal delivery vs. 1.1 % with C-section, p = 0.14). However, cesarean delivery was associated with higher rates of postpartum heart failure [7]. Additionally, no difference in the incidence of PPH was found across different delivery modalities [7]. Similarly, Easter et al. examined a similar population of women with cardiovascular diseases, and reported maternal adverse cardiac outcomes and neonatal outcomes were similar after stratification for the mode of delivery; however, rates of PPH (1.9 % vs. 10.6 %, p < 0.01) and transfusion (1.9 % vs. 9.1 %, p = 0.01) were both lower in the vaginal delivery group [8]. Gavin et al. performed a retrospective analysis of maternal cardiac disease in two hospitals and reported that for those undergoing cesarean delivery, the rates of severe maternal morbidity were significantly higher than those undergoing vaginal delivery [6]. Several potential explanations exist for the discordance between our data and the results presented by both of these prior studies. Regarding maternal outcomes and mode of delivery, this current study included many more patients who may be more clinically homogenous, given we investigated only HF and pHTN. In contrast, other studies included patients with a wider variety of cardiac diseases. This may have enhanced our power to detect a change in outcomes with stratification for delivery method. Regarding PPH, traditional teachings suggest the risk for PPH is elevated with cesarean delivery, but our findings demonstrated significantly greater PPH in vaginal deliveries. In the ROPAC cohort, there was no difference in PPH observed based on delivery mode [9,10]. The definition of PPH has evolved and was previously dependent on the mode of delivery, whereas contemporary definitions no longer include such consideration. How the evolution of this definition has influenced our results is unclear, particularly given the dependence of our study on ICD codes and coding practices, which evolved during the duration of these collected data.

Both vaginal and cesarean delivery are tied to complex changes in maternal hemodynamics in mothers without cardiovascular disease. Prior to the onset of labor and delivery, pregnancy is associated with the development of significant cardiovascular adaptations, including increases in blood volume, stroke volume, and cardiac output (CO) [11,12]. Vaginal delivery is associated with continued increases in CO, with maximal values reaching in the first 10 min of the postpartum period. Epidural anesthesia attenuates this increase in CO; however, a net increase still exists. Vaginal delivery is associated with use of the Valsalva maneuver, which further impacts hemodynamics and can reduce preload while increasing afterload [13]. Cesarean delivery is associated with continued increases in CO until immediately after delivery [11,12]. This increase in CO occurs alongside decreases in total peripheral resistance (TPR) and mean arterial pressure (MAP) [11,12]. Spinal anesthesia likely contributes significantly to the decrease in TPR associated with cesarean delivery [12]. The suggested benefits of vaginal delivery over cesarean delivery in patients with HF and pHTN include less blood loss and decreased infection or thrombogenic risk. Our findings are concordant with this recommendation, as vaginal delivery was associated with lower odds of maternal in-hospital mortality. However, we observed a higher rate of PPH with vaginal delivery than cesarean delivery.

4.1. Limitations

There are several limitations associated with our study. Due to the use of an administrative database, errors related to coding discrepancies may influence our results as the fidelity of our data is dependent on the rigor of coding practices. However, HCUP quality control procedures are routinely performed to confirm that NRD data values are valid, consistent, and reliable [14]. Furthermore, this database lacks information regarding important clinical data, such as prenatal care, laboratory results and medications. Severity of HF (NYHA class) or pHTN prior to/during pregnancy was not available in the database, thus misclassification remains a potential limitation. Milder cases of HF or pHTN may have been underdiagnosed, leading to an underestimation of event rates, whereas more severe cases could be associated with higher event rates than reported. This limitation may affect the generalizability of our findings to individual patient risk. We cannot confirm the etiology of HF or pHTN, however we have mentioned the associated comorbidities such as ischemic or non-ischemic cardiomyopathy, valvular heart diseases, and congenital heart diseases in table-1 which may represent the possible etiologies of these conditions. We do not have information regarding the indication for cesarean delivery and are unable to determine which patients underwent cesarean delivery for obstetrical indications as endorsed by ESC guidelines vs the perceived risk associated with maternal cardiovascular disease [4]. Additionally, while propensity score matching was employed to reduce confounding from observed variables, residual confounding and confounding by indication may still persist due to unmeasured factors, such as clinical severity of HF or pHTN, provider preferences for delivery mode, or higher baseline risk in patients selected for cesarean delivery, which could bias the comparison of outcomes between delivery modes. Moreover, the NRD does not capture deaths or events occurring outside the index delivery or readmission hospitalizations, which may lead to an underestimation of maternal and fetal event rates, particularly for post-discharge complications or mortality not resulting in readmission. The extent of this missing data is not quantifiable within the NRD, but it may reduce the observed associations’ strength, especially for maternal mortality, given that some pregnancy-related deaths occur outside inpatient settings. Non-teaching hospitals were under-represented in the NRD database and the results were mainly generated from teaching hospitals of large size. Finally, the inpatient nature of this data set does not allow to capture the intensity or quality of care before and after the delivery hospitalization for mothers or neonates.

5. Conclusions

While Cesarean delivery showed a correlation with reduced fetal mortality and decreased post-partum hemorrhage rates, it was also associated with elevated in-hospital maternal mortality, acute coronary syndrome, and puerperal sepsis. Interestingly, Cesarean delivery showed similar lower fetal mortality but no difference in maternal mortality in patients with pulmonary hypertension. Highlighting the necessity for a multidisciplinary approach that considers the unique patient characteristics is essential when determining the most appropriate delivery method for individuals with HF or pHTN. The current data presented here will aid in this decision-making process given the difficulty of performing prospective studies/randomized controlled trials in pregnant patients.

Funding

None.

Authorship contribution

All authors participated in the research and preparation of the manuscript as per the International Committee of Medical Journal Editors (ICMJE).

CRediT authorship contribution statement

Monil Majmundar: Writing – review & editing, Visualization, Supervision, Resources, Methodology, Formal analysis, Conceptualization, Writing – original draft, Validation, Software, Project administration, Investigation, Data curation. Kunal N. Patel: Writing – original draft, Validation, Software, Project administration, Investigation, Data curation, Writing – review & editing, Visualization, Supervision, Resources, Methodology, Formal analysis, Conceptualization. Tasveer Khawaja: Writing – review & editing, Visualization, Writing – original draft. Rajkumar Doshi: Writing – review & editing, Supervision, Methodology, Validation, Project administration, Conceptualization. Harshvardhan Zala: Software, Formal analysis, Methodology, Data curation. Ankur Kalra: Validation, Project administration, Investigation, Writing – review & editing, Supervision, Methodology, Conceptualization. Megan McKee Thomas: Validation, Methodology, Writing – review & editing, Supervision. Matthew Lippmann: Validation, Methodology, Writing – review & editing, Supervision. Martha Gulati: Validation, Project administration, Investigation, Writing – review & editing, Supervision, Methodology, Conceptualization.

Declaration of competing interest

The authors declare the following financial interests/personal relationships which may be considered as potential competing interests:

Martha Gulati reports a relationship with Novartis, Esperion and Boehringer Ingelheim that includes: consulting or advisory. Dr. Martha Gulati serves on a DSMB for Merck Inc. If there are other authors, they declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.