Abstract
Background
Preterm birth is associated with an increased rate of serious maternal complication compared with those who deliver at term. However, data are limited on risk factors for maternal morbidity and mortality specifically in the setting of preterm birth. Improved understanding of risk factors for morbidity in patients with preterm birth can help optimize interventions to best support patients experiencing preterm birth.
Objective
To assess associations between obstetric factors and maternal morbidities and mortality among patients who gave birth less than 35 weeks’ gestation.
Study design
This is a secondary analysis of a multicenter registry of gravidas carrying singletons or twins who gave birth at 16 0/7 to 34 6/7 weeks’ gestation January through August 2019. The primary outcome was a maternal morbidity and mortality composite (blood transfusion, postpartum hemorrhage, maternal sepsis, stroke, thromboembolism, intensive care unit admission, or death) during the delivery hospitalization. Factors were compared between those with and without the composite outcome to identify statistically significant variables. Those factors were then used to create a parsimonious multivariable logistic regression model with adjusted odds ratio (aOR) and 95% confidence intervals (CIs) using a backward strategy. We estimated area under the curve to quantify the predictive ability of these factors for the primary outcome.
Results
The analysis was based on 6660 patients, of whom 15.1% experienced the primary composite outcome. Factors associated with increased odds of the outcome included increasing maternal age, use of assisted reproductive technology, twin gestation, preeclampsia, placental abruption and cesarean delivery—aOR ranged from 1.02 [1.01–1.04] for maternal age (continuous) to 2.55 [2.08–3.14] for abruption. In contrast, nulliparity, gestational diabetes, preterm prelabor rupture of membranes (PPROM), and later birth gestational age were associated with decreased odds of the primary outcome—aOR ranging from 0.96 [0.95–0.98] for gestational age (continuous) to 0.54 [0.44–0.66] for PPROM. The model with these factors had low to moderate predictive ability (AUC 0.68 [0.66–0.69]).
Conclusion
Among individuals who experienced preterm birth before 35 weeks, multiple factors were associated with maternal morbidity and mortality. The factors together modestly discriminate maternal morbidity or mortality.
Keywords: maternal morbidity, maternal mortality, preterm birth
1. INTRODUCTION
Every day, over 200 patients who are hospitalized for delivery in the United States experience severe maternal morbidity [1]. Maternal morbidity and mortality rates have been steadily increasing over the past two decades [2, 3]. The United States has the highest maternal mortality ratio among developed countries [4]. Maternal morbidity increases length of hospital stay and overall medical costs [5]. Moreover, morbidity is likely a precursor to maternal mortality [1].
In patients who give birth preterm, the rate of serious maternal complications is more than fourfold greater compared with those who deliver at term [6, 7, 8]. Data are limited on risk factors for maternal morbidity and mortality specifically in the setting of preterm birth [6]. Preterm birth is known to be associated with maternal morbidity, but specific associated risk factors remain unclear [9, 10, 11]. Improved understanding of risk factors for morbidity in patients with preterm birth can help optimize interventions to best support patients experiencing preterm birth.
Our aim was to assess and quantify any associations between maternal demographic and obstetric factors and maternal morbidities and mortality among the high‐risk group of patients who gave birth <35 weeks’ gestation. We also quantified the ability of the risk factors to predict the morbidity and mortality outcome.
2. MATERIALS AND METHODS
This was a secondary analysis of the Eunice Kennedy Shriver National Institute of Child Health and Human Development Maternal‐Fetal Medicine Units (MFMU) Network Prematurity Registry. The registry is a database of baseline and outcome information from all patients who gave birth from 16 0/7 through 34 6/7 weeks’ gestation from January 1 through August 31 2019 at 33 hospitals associated with the MFMU Network throughout the United States. All centers obtained institutional review board approval for the study and a waiver of informed consent. In this secondary analysis, all individuals who delivered live or stillborn infants in a singleton or twin gestation at the affiliated hospitals in the study period were included. For those with twin gestations, at least one fetus must have been delivered in the qualifying gestational age range. Exclusion criteria were individuals with a twin gestation with deliveries of infants more than 72 h apart or those with a prior pregnancy included during the study period. Demographic data, medical history, obstetrical history, and antenatal, labor and delivery, and postpartum data were collected. Certified research coordinators abstracted data using predefined search criteria and definitions from the delivery admission for all deliveries that met eligibility criteria.
The primary outcome of this analysis was a composite of maternal morbidity or mortality defined as at least one of: transfusion of one or more blood products, postpartum hemorrhage, maternal sepsis, stroke, intensive care unit (ICU) admission, or death occurring during delivery hospitalization. Individual components were reported for each patient. The primary outcome was a dichotomous variable with one or more instances of the components classified as one composite outcome. Postpartum hemorrhage was defined by a clinical diagnosis and medical record documentation of “hemorrhage,” “postpartum bleed,” or “atony” or “expressed fresh clots” with estimated blood loss consistent with hemorrhage per the primary registry collection criteria. A secondary analysis was performed to evaluate for association of perinatal risk factors with a composite maternal infection (defined as chorioamnionitis, postpartum endometritis, maternal sepsis, and surgical site infection by maternal discharge) and cesarean delivery (restricting to only live births delivering from 23 0/7 through 34 6/7 weeks’ gestation). For the composite maternal infection outcome, chorioamnionitis was excluded as a predictor and instead was a component of the composite infection.
Preselected maternal, obstetrical, and neonatal characteristics were compared between those with and without the primary outcome using chi‐square, Fisher's exact, or Wilcoxon rank‐sum tests as appropriate. The preselected exposure variables were those that were available within the registry that have been previously associated with maternal morbidity and mortality [6, 7, 8]. Race and ethnicity were included and interpreted as social constructs. Those significantly associated with each outcome (p < 0.05) in the univariable analysis were eligible for the multivariable regression. The multivariable logistic regression models were developed and internally validated using stratified 10‐fold cross‐validation with stepwise backward elimination, with a significance level of 0.05 required to be retained in the model. Among the parsimoniously developed models, the final model was the model with the highest c‐statistic (area under the receiver operating curve [AUC]) in which the factors in the final model appear in at least eight out of the 10 models. A sensitivity analysis excluding postpartum hemorrhage from the maternal composite outcome was performed to help characterize the most severe morbidity and mortality outcomes, as postpartum hemorrhage composed a large component of the composite outcome and is clinically less severe than the other factors contributing to the composite. A second sensitivity analysis was performed with spontaneous versus iatrogenic (indicated) preterm birth separated to assess for differences in maternal morbidity and mortality factors. Iatrogenic preterm birth was classified as “no labor,” “induced,” or “dilation and evacuation” labor types; spontaneous preterm birth included “spontaneous” and “spontaneous augmented” labor types.
No adjustments were made for multiple comparisons, as secondary outcome analyses were considered exploratory. Analyses were completed using SAS 9.4.
3. RESULTS
All 6660 pregnant patients who gave birth from 16 0/7 through 34 6/7 weeks’ gestation in the registry were included. 1003 patients (15.1%) experienced the primary outcome of composite maternal morbidity or mortality. Frequency of the composite outcome differed by race/ethnicity, maternal age, parity, plurality, mode of conception, preeclampsia with severe features, magnesium sulfate use, PPROM, abruption, gestational age at delivery, and cesarean delivery (Table 1). Of those who experienced the primary outcome, 455 (6.8%) were transfused with one or more blood products, 638 (9.6%) experienced postpartum hemorrhage, 44 (0.7%) were diagnosed with maternal sepsis, 6 (0.1%) with stroke, 12 (0.2%) with thromboembolism, 257 (3.9%) were admitted to the ICU, and 5 (0.1%) died (Table 2). Excluding postpartum hemorrhage, the frequency of the composite outcome was 365 (5.4%).
TABLE 1.
Participant characteristics and obstetrical determinants for maternal morbidity and mortality outcomes for births between 16,0 and 34,6 weeks; data presented as number (percentage) unless otherwise noted.
| Demographics | No maternal morbidity composite (N = 5657) | Maternal morbidity composite (N = 1003) | p value |
|---|---|---|---|
| Maternal characteristics | |||
| Race/ethnicity | |||
| Non‐Hispanic Black | 1686 (29.8) | 350 (34.9) | 0.003 |
| Non‐Hispanic White | 2511 (44.4) | 434 (43.3) | |
| Hispanic | 947 (16.7) | 132 (13.2) | |
| Other | 513 (9.1) | 87 (8.7) | |
| Maternal Age (years) (mean ± SD) | 29.7 ± 6.2 | 30.9 ± 6.2 | <0.001 |
| Employment | |||
| Employed | 2718 (48.1) | 491 (49.0) | 0.806 |
| Unemployed | 1551 (27.4) | 275 (27.4) | |
| Not documented | 1388 (24.5) | 237 (23.6) | |
| Married/living with partner | 3309 (58.5) | 565 (56.3) | 0.201 |
| Government insurance | 2920 (51.6) | 506 (50.5) | 0.495 |
| ≥ 2 prenatal visits | 4631 (81.9) | 800 (79.8) | 0.114 |
| Type of conception | |||
| ART | 333 (5.9) | 118 (11.8) | <0.001 |
| Spontaneous/Not documented | 5324 (94.1) | 885 (88.2) | |
| Maternal BMI at delivery (kg/m2) a (mean ± SD) | 32.3 ± 8.1 | 33.0 ± 8.4 | 0.037 |
| Obstetrical determinants | |||
| Nulliparous | 2449 (43.3) | 393 (39.2) | 0.015 |
| Cesarean delivery | 2677 (47.3) | 639 (63.7) | <0.001 |
| Twin gestation | 692 (12.2) | 163 (16.3) | <0.001 |
| Hypertensive disorders | |||
| GHTN or PE/SIPE without severe features | 571 (10.1) | 116 (11.6) | 0.158 |
| SIPE with severe features/ PE with severe features/ HELLP/Eclampsia | 1560 (27.6) | 415 (41.4) | <0.001 |
| Gestational diabetes b | 509 (9.0) | 74 (7.4) | 0.092 |
| PPROMc | 1691 (29.9) | 151 (15.1) | <0.001 |
| Placental abruption | 394 (7.0) | 157 (15.7) | <0.001 |
| Chorioamnionitis | 226 (4.0) | 48 (4.8) | 0.245 |
| Antenatal steroids | 3722 (65.8) | 653 (65.1) | 0.671 |
| Magnesium sulfate | 2872 (50.8) | 601 (59.9) | <0.001 |
| Neonatal characteristics | |||
| Gestational age at delivery, weeks (first twin) (mean ± SD) | 30.2 ± 4.8 | 29.9 ± 4.9 | 0.018 |
| Estimated fetal weight (if twins, largest) | |||
| <5% | 345 (6.1) | 54 (5.4) | 0.302 |
| 5%–10% | 269 (4.8) | 57 (5.7) | |
| >10% | 3168 (56.0) | 579 (57.7) | |
| Not assessed | 1875 (33.1) | 313 (31.2) | |
| Oligohydramnios d | 221 (3.9) | 37 (3.7) | 0.742 |
Abbreviations: ART, assisted reproductive technology; BMI, body mass index; GHTN, gestational hypertension; HELLP, hemolysis elevated liver enzymes low platelet syndrome; PE, preeclampsia; PPROM, preterm prelabor rupture of membranes; SIPE; superimposed preeclampsia.
BMI missing on 196 of 6660.
A priori variable specified for backward regression, missing on n = 7.
PPROM only if reason for delivery admission.
Oligohydramnios considered if true for either twin in multifetal gestation in the absence of PPROM.
TABLE 2.
Prevalence of outcomes during hospitalization for preterm birth.
| Outcome |
N (%) (n = 6660) |
|---|---|
| Maternal morbidity and mortality composite | 1003 (15.1) |
| Transfusion of ≥1 blood products | 455 (6.8) |
| Postpartum hemorrhage | 638 (9.6) |
| Maternal sepsis | 44 (0.7) |
| Stroke | 6 (0.1) |
| Thromboembolism | 12 (0.2) |
| ICU admission | 257 (3.9) |
| Death | 5 (0.1) |
| Maternal infection composite | 383 (5.8) |
| Chorioamnionitis | 274 (4.1) |
| Postpartum endometritis | 79 (1.2) |
| Maternal sepsis | 44 (0.7) |
| Surgical site infection | 23 (0.4) |
| Cesarean delivery | 3261 (58.6) |
Abbreviation: ICU, intensive care unit.
Factors associated with a higher rate of the composite maternal morbidity and mortality in the multivariable model included maternal age (per year increase) (adjusted odds ratio [aOR] 1.02; 95% confidence interval [CI], [1.01–1.04]), use of assisted reproductive technology (ART) (aOR, 1.84; [1.44–2.35]), twin gestation (aOR, 1.37; [1.13–1.67]), preeclampsia with severe features (aOR, 1.66; [1.41–1.94]), placental abruption (aOR, 2.55; [2.08–3.14]), and cesarean delivery (aOR, 1.65; [1.42–1.92]) (Table 3). Factors associated with a lower rate of the composite outcome were nulliparity (aOR, 0.82; [0.71–0.96]), gestational diabetes mellitus (aOR, 0.75; [0.57–0.97]), preterm prelabor rupture of membranes (PPROM) at admission (aOR, 0.54; [0.44–0.66]), and later birth gestational age (per week increase) (aOR, 0.96; [0.95–0.98]) (Table 3). The adjusted model with these factors modestly discriminated maternal morbidity, with an AUC of 0.68 (95% CI; 0.66–0.69).
TABLE 3.
Obstetrical factors and association with composite maternal morbidity and mortality outcome (maternal death, intensive care unit [ICU] admission, transfusion of one or more blood products, postpartum hemorrhage, maternal sepsis, stroke, or thromboembolism).
| Risk factors | Unadjusted OR or unit increase | Adjusted OR or unit increase a | OR without hemorrhage in composite | aOR without hemorrhage in composite b |
|---|---|---|---|---|
| Maternal characteristics | ||||
| Race/ethnicity | ||||
| Non‐Hispanic Black | 1.21 (1.03, 1.41) | g | 1.26 (1.05, 1.51) | 1.51 (0.95, 1.40) |
| Non‐Hispanic White | Ref | g | Ref | Ref |
| Hispanic | 0.81 (0.66, 1.00) | g | 0.74 (0.57, 0.95) | 0.75 (0.57, 0.97) |
| Other | 0.98 (0.77, 1.26) | g | 0.81 (0.59, 1.11) | 0.84 (0.61, 1.17) |
| Maternal age (each increasing year) | 1.03 (1.02, 1.04) | 1.02 (1.01, 1.04) | 1.02 (1.01, 1.04) | 1.03 (1.02, 1.05) |
| Employed | 1.04 (0.90, 1.18) | g | 0.92 (0.78, 1.08) | g |
| Married/living with partner | 0.92 (0.80, 1.05) | g | 0.76 (0.65, 0.89) | 0.74 (0.61, 0.89) |
| Government insurance | 0.96 (0.84, 1.09) | g | 1.20 (1.02, 1.41) | g |
| ≥2 prenatal visits | 0.87 (0.73, 1.03) | g | 0.76 (0.63, 0.93) | g |
| Type of conception | ||||
| ART | 2.13 (1.71, 2.66) | 1.84 (1.44, 2.35) | 1.36 (1.02, 1.82) | g |
| Spontaneous/ Not documented | Ref | Ref | Ref | g |
| Maternal BMI at delivery (each increasing kg/m2) c | 1.01 (1.00, 1.02) | g | 1.01 (1.00, 1.02) | g |
| Obstetrical determinants | ||||
| Nulliparity | 0.84 (0.74, 0.97) | 0.82 (0.71, 0.96) | 0.80 (0.68, 0.95) | g |
| Cesarean delivery | 1.95 (1.70, 2.25) | 1.65 (1.42, 1.92) | 2.99 (2.50, 3.58) | 2.40 (1.99, 2.90) |
| Twin gestation | 1.39 (1.16, 1.68) | 1.37 (1.13, 1.67) | 1.00 (0.79, 1.27) | g |
| SIPE with severe features/ PE with severe features/ HELLP/ Eclampsia | 1.85 (1.61, 2.13) | 1.66 (1.41, 1.94) | 2.25 (1.91, 2.65) | 1.53 (1.28, 1.82) |
| Gestational diabetes mellitus d | 0.81 (0.62, 1.04) | 0.75 (0.57, 0.97) | 0.65 (0.47, 0.91) | 0.61 (0.43, 0.86) |
| PPROM e | 0.42 (0.35, 0.50) | 0.54 (0.44, 0.66) | 0.30 (0.23, 0.39) | 0.41 (0.31, 0.53) |
| Placental abruption | 2.48 (2.03, 3.02) | 2.55 (2.08, 3.14) | 2.35 (1.86, 2.96) | 2.34 (1.84, 2.98) |
| Chorioamnionitis | 1.21 (0.88, 1.66) | g | 1.22 (0.84, 1.79) | g |
| Antenatal steroids | 0.97 (0.84, 1.12) | g | 1.06 (0.89, 1.26) | g |
| Magnesium sulfate | 1.45 (1.27, 1.66) | g | 1.82 (1.54, 2.15) | g |
| Neonatal characteristics | ||||
| Gestational age (each increasing completed week) | 0.99 (0.97, 1.00) | 0.96 (0.95, 0.98) | 0.99 (0.98, 1.01) | g |
| Estimated fetal weight | ||||
| <5% | 0.86 (0.63, 1.16) | g | 0.89 (0.62, 1.27) | g |
| 5%–10% | 1.16 (0.86, 1.56) | g | 1.41 (1.01, 1.98) | g |
| >10% | Ref | g | Ref | g |
| Not assessed | 0.91 (0.79, 1.06) | g | 0.84 (0.70, 1.00) | g |
| Oligohydramnios f | 0.78 (0.60, 1.02) | g | 0.91 (0.67, 1.23) | |
| Adjusted AUC (95% CI) | 0.68 (0.66, 0.69) | 0.71 (0.69, 0.73) |
Abbreviations: aOR, adjusted odds ratio; ART, assisted reproductive technology; AUC, area under the curve; BMI, body mass index; HELLP, hemolysis elevated liver enzymes low platelet syndrome; OR, odds ratio; PE, preeclampsia; PPROM, preterm prelabor rupture of membranes; SIPE, superimposed preeclampsia.
Adjusted for placental abruption, PPROM, ART, PE with severe features, cesarean delivery, twin gestation, gestational diabetes mellitus, multiparity, maternal age, gestational age.
Adjusted for race/ethnicity, cesarean delivery, maternal age, marital status, placental abruption, gestational diabetes mellitus, PE with severe features, PPROM.
BMI missing on 196 of 6660.
A priori variable specified for backward regression, missing on n = 7.
PPROM only if reason for delivery admission.
Oligohydramnios considered if true for either twin in multifetal gestation in the absence of PPROM.
Not part of parsimonious model.
In the sensitivity analysis which excluded postpartum hemorrhage from the composite morbidity and mortality outcome, a few differences were observed in the adjusted model (Table 3): Hispanic ethnicity and a relationship status of being married or living with a partner were associated with lower odds of the outcome; ART, nulliparity, twin gestation, and gestational age were not significantly associated with the outcome. Cesarean delivery was more strongly associated with composite morbidity and mortality excluding hemorrhage (aOR, 2.40; [1.99–2.90]) compared to composite morbidity and mortality that included hemorrhage (aOR, 1.65; [1.42–1.92]). The associations with other factors remained consistent. The adjusted model excluding hemorrhage moderately discriminated the composite maternal morbidity with an AUC of 0.71 (95% CI, 0.69–0.73).
In the sensitivity analysis by iatrogenic versus spontaneous preterm birth, several differences were observed in the adjusted models (Table 4). Black race was associated with maternal morbidity for both iatrogenic (aOR, 1.22; 95% CI, [1.01–1.47]) and spontaneous preterm birth (OR, 1.43; 95% CI, [1.06–1.94]). Maternal age and cesarean delivery were associated with maternal morbidity in iatrogenic preterm birth (age: aOR, 1.03; 95% CI, [1.01–1.04], cesarean delivery: aOR, 1.45; 95% CI, [1.21–1.74]), but the finding was attenuated with spontaneous preterm birth (age: aOR, 1.01; 95% CI, [0.99–1.04], cesarean delivery: aOR, 1.19; 95% CI, [0.90–1.59]). Gestational age was less associated with maternal morbidity in spontaneous preterm birth (OR, 0.93; 95% CI, [0.91–0.96]), but the finding was not significant for iatrogenic preterm birth (OR, 1.00; 95% CI, [0.99–1.02]). The stratified adjusted models for the association between risk factors and composite morbidity has comparable discrimination to each other and to the overall cohort: AUC of 0.65 (95% CI, 0.61–0.66) for iatrogenic and 0.64 (95% CI, 0.60–0.68) for spontaneous preterm birth.
TABLE 4.
Obstetrical factors and association with composite maternal morbidity and mortality outcome (maternal death, ICU admission, transfusion of one or more blood products, postpartum hemorrhage, maternal sepsis, stroke, or thromboembolism) based on etiology for preterm birth.
| Iatrogenic preterm birth (N = 4026) | Spontaneous preterm birth (N = 2634) | |||
|---|---|---|---|---|
| Risk factors | Unadjusted OR or unit increase | Adjusted OR or unit increase a | Unadjusted OR or unit increase | Adjusted OR or unit increase a |
| Maternal characteristics | ||||
| Race/ethnicity | ||||
| Non‐Hispanic Black | 1.31 (0.94, 1.35) | 1.22 (1.01, 1.47) | 1.43 (1.06, 1.94) | g |
| Non‐Hispanic White | Ref | Ref | Ref | g |
| Hispanic | 0.75 (0.58, 0.96) | 0.81 (0.63, 1.04) | 1.04 (0.71, 1.54) | g |
| Other | 0.93 (0.69, 1.25) | 0.91 (0.67, 1.23) | 1.20 (0.75, 1.92) | g |
| Maternal age (each increasing year) | 1.03 (1.02, 1.04) | 1.03 (1.01, 1.04) | 1.02 (1.00, 1.04) | 1.01 (0.99, 1.04) |
| Employed | 0.99 (0.84, 1.56) | g | 1.08 (0.83, 1.40) | g |
| Married/living with partner | 0.89 (0.76, 1.05) | g | 0.93 (0.72, 1.21) | g |
| Government insurance | 1.00 (0.85, 1.17) | g | 0.94 (0.72, 1.22) | g |
| ≥2 prenatal visits | 0.96 (0.78, 1.18) | g | 0.61 (0.46, 0.82) | g |
| Type of conception | ||||
| ART | 1.94 (1.50, 2.52) | 1.61 (1.21, 2.13) | 2.36 (1.52, 3.66) | 1.88 (1.17, 3.02) |
| Spontaneous/Not documented | Ref | Ref | ||
| Maternal BMI at delivery (each increasing kg/m2) c | 1.00 (0.99, 1.01) | g | 1.02 (1.00, 1.04) | g |
| Obstetrical determinants | ||||
| Nulliparity | 0.86 (0.73, 1.01) | g | 0.79 (0.60, 1.03) | g |
| Cesarean delivery | 1.73 (1.45, 2.06) | 1.45 (1.21, 1.74) | 1.38 (1.05, 1.82) | 1.19 (0.90, 1.59) |
| Twin gestation | 1.47 (1.16, 1.87) | 1.43 (1.12, 1.83) | 1.69 (1.23, 2.30) | 1.56 (1.09, 2.16) |
| SIPE with severe features/ PE with severe features/ HELLP/ Eclampsia | 1.39 (1.19, 1.63) | 1.40 (1.19, 1.65) | 1.94 (1.28, 2.94) | 1.88 (1.23, 2.87) |
| Gestational diabetes mellitus d | 0.91 (0.69, 1.21) | g | 0.48 (0.26, 0.88) | g |
| PPROM e | g | g | 0.54 (0.40, 0.72) | g |
| Placental abruption | 2.62 (2.04, 3.38) | 2.70 (2.08, 3.51) | 2.82 (2.01, 3.96) | 2.96 (2.10, 4.17) |
| Chorioamnionitis | 0.93 (0.62, 1.40) | g | 2.08 (1.25, 3.48) | g |
| Antenatal steroids | 1.09 (0.92, 1.28) | g | 0.80 (0.61, 1.04) | g |
| Magnesium sulfate | 1.36 (1.16, 1.60) | g | 1.20 (0.93, 1.56) | g |
| Neonatal characteristics | ||||
| Gestational age (each increasing completed week) | 1.00 (0.99, 1.02) | g | 0.93 (0.91, 0.96) | g |
| Estimated fetal weight | ||||
| <5% | 0.64 (0.47, 0.89) | g | 1.41 (0.59, 3.35) | g |
| 5%–10% | 0.91 (0.66, 1.26) | g | 1.54 (0.68, 3.47) | g |
| >10% | Ref | g | 1.27 (0.97, 1.66) | g |
| Not assessed | 0.86 (0.72, 1.04) | g | g | |
| Oligohydramnios f | 0.65 (0.44, 0.99) | g | 2.13 (1.06, 4.27) | g |
| Adjusted AUC (95% CI) | 0.65 (0.61, 0.66) | 0.64 (0.60, 0.68) | ||
Bolded values indicate statistical significance.
Abbreviations: OR, odds ratio; ART, assisted reproductive technology; BMI, body mass index; SIPE, superimposed preeclampsia; PE, preeclampsia; HELLP, hemolysis elevated liver enzymes low platelet syndrome; PPROM, preterm prelabor rupture of membranes.
Adjusted for placental abruption, PPROM, ART, PE with severe features, cesarean delivery, twin gestation, gestational diabetes mellitus, multiparity, maternal age, and gestational age.
Adjusted for race/ethnicity, cesarean delivery, maternal age, marital status, placental abruption, gestational diabetes mellitus, PE with severe features, and PPROM.
BMI missing on 196 of 6660.
A priori variable specified for backward regression, missing on n = 7.
PPROM only if reason for delivery admission.
Oligohydramnios considered if true for either twin in multifetal gestation in the absence of PPROM.
Not part of parsimonious model due to non‐significance as a univariable and/or excluded per k‐fold regression.
For the secondary outcome, the maternal infection composite occurred in 383 patients (5.8%). Chorioamnionitis occurred in 274 patients (4.1%), postpartum endometritis in 79 patients (1.2%), maternal sepsis in 44 patients (0.7%), and surgical site infection in 23 patients (0.4%) (Table 2). PPROM was associated with higher odds of composite maternal infection outcome (aOR, 2.57; [2.08–3.18]), while later gestational age at birth was associated with lower odds (aOR, 0.90; [0.88–0.91]) (Supplemental Table 2). The adjusted model discriminated the composite infection outcome with an AUC of 0.70 (95% CI, 0.68–0.73).
Of 5561 patients with live births after 23 weeks, 3261 (58.6%) underwent a cesarean delivery (Table 2). Greater maternal age (aOR, 1.05; [1.04–1.06]), greater delivery body mass index (BMI) (aOR, 1.02; [1.01–1.03]), twin gestation (aOR, 2.02; [1.70–2.41]), preeclampsia with severe features (aOR, 1.70; [1.47–1.95]), and placental abruption (aOR, 1.48; [1.19–1.83]) were associated with increased odds of cesarean delivery (Supplemental Table 3). Lower estimated fetal weight (EFW) was associated with higher odds of cesarean birth compared with EFW ≥10% (EFW, 5%–10%; aOR, 1.97; [1.47–2.63]; EFW < 5%; aOR, 2.90; [2.12–3.97]). PPROM (aOR, 0.45; [0.39–0.51]) and increasing gestational age in weeks (aOR, 0.91; [0.89–0.93]) were associated with lower risk of cesarean delivery rate (Supplemental Table 3). The adjusted model with these factors discriminated for cesarean delivery with an AUC of 0.72 (95% CI, 0.70–0.73).
4. DISCUSSION
In this large and diverse cohort of people with preterm births in the United States, a total of 10 out of 21 factors considered in this analysis were associated with the composite maternal morbidity and mortality outcome. Increasing maternal age, use of ART, twin gestation, preeclampsia with severe features, placental abruption, and cesarean delivery were independently associated with an increased risk for composite maternal morbidity and mortality. In contrast, nulliparity, gestational diabetes mellitus, PPROM, and later gestational age were independently associated with a lower risk. When hemorrhage was excluded from the composite outcome, being married or living with a partner was associated with a lower risk, whereas ART and twin gestation were no longer associated with the outcome. In the sensitivity analysis for the etiology of preterm birth, there was an association of maternal morbidity with ART, twin gestation, preeclampsia and placental abruption for both types of preterm birth. For iatrogenic preterm birth only, there is an association with maternal morbidity and non‐Hispanic Black race, maternal age, and cesarean delivery.
PPROM was associated with increased infectious risk while increased gestational age was associated with lower risk. Greater maternal age and delivery BMI, twin gestation, preeclampsia with severe features, placental abruption, and lower EFW were associated with an increased risk of cesarean delivery, while PPROM and a later gestational age at delivery were associated with a lower risk of cesarean delivery. Despite these identified associations, the parsimonious models demonstrated only modest discrimination for maternal outcomes.
Few studies have evaluated associations with maternal morbidity and mortality following preterm birth. Most previous studies evaluated maternal morbidity without distinguishing between preterm and term births [6, 7, 8, 11, 12, 13, 14, 15, 16, 17]. Overlapping with our findings, in deliveries of all gestational ages, maternal morbidity and mortality has been associated with: extremes of maternal age, minority race (including Black race), single marital status, government insurance, obesity, nulliparity, multifetal gestation, hypertensive disorders of pregnancy, gestational diabetes mellitus, earlier gestational age at time of delivery, and cesarean delivery [6, 7, 8, 11, 12, 13, 14, 15, 16, 17]. Our multicenter, well‐characterized study cohort of preterm birth mirrors findings of established risk factors among individuals who predominantly gave birth at term.
In this study, when hemorrhage was excluded from the composite morbidity and mortality, cesarean delivery was more strongly associated with the composite outcome. Cesarean delivery has been associated with severe maternal morbidity and mortality [6, 18]. However, the model excluding hemorrhage continued to demonstrate modestly discrimination for maternal outcomes compared to the full composite model.
When we performed a sensitivity analysis evaluating outcomes based on preterm birth etiology, non‐Hispanic Black race, assessed as a social construct, was associated with an increased risk for maternal morbidity in iatrogenic but not in spontaneous preterm birth. Patients of racial and ethnic minorities have an increased risk for medical comorbidities affecting pregnancy compared to patients of White race [19]. It is possible that these underlying medical comorbidities contribute to the racial differences, however, disparities in health care also contribute to disparities in health outcomes [19].
PPROM was associated with lower odds of composite maternal morbidity and mortality but higher odds of maternal infectious morbidity in our secondary analysis. Previous work across all gestational ages have identified an association of PPROM with an increased risk of infection and placental abruption [20, 21, 22]. The attenuated association of PPROM with the primary outcome in this cohort may reflect the greater contribution of other high‐risk conditions, including preeclampsia, to severe morbidity [23].
Similar to our findings, previous work in term births found an association between cesarean delivery and placental abruption [24], PPROM [25], preeclampsia [26], multiple gestation [27], magnesium sulfate administration [28], lower fetal weight [29], greater maternal age [17, 30], and greater maternal delivery BMI [28].
In previous work including individuals who delivered across all gestational ages, obstetrical factors associated with composite maternal morbidity have been shown to be only modestly predictive of maternal morbidity. Our estimates in this cohort for maternal morbidity and mortality following preterm birth were similar to previous work that included term and preterm births (AUC 0.64–0.66) [10, 11].
This study provides insight for health care workers caring for patients who give birth preterm. Preterm birth frequently co‐occurs with maternal morbidity and mortality. Even when preterm birth is not indicated due to maternal morbidity, those who experience preterm birth of any etiology remain at elevated risk for maternal morbidity and mortality. Recognizing patients who are at increased risk for morbidity within the preterm gestational age range, regardless of whether the morbidity occurs before or after the birth, is essential for guiding clinical decision making and optimizing resource allocation. Despite identification of risk factors for maternal morbidity and mortality, predictive models using these factors are insufficient for clinical practice.
Strengths of this study include multicenter data with a large sample, as part of the MFMU Network representing different regions of the United States. Data were abstracted by trained research coordinators, and the outcomes are accurately and thoroughly characterized. However, there are several limitations. First, although the cohort was large, the relatively low incidence of the primary outcome limited statistical power to detect associations with less common but clinically important risk factors, and the analysis was limited to variables available within the Prematurity Registry. Second, the rate of each outcome within the primary composite differs greatly, with receipt of a blood transfusion and hemorrhage as the leading diagnoses that contribute to the morbidity outcome. As such, blood transfusions and hemorrhage may be a driver for the associations with the composite morbidity and risk factors. To address this potential limitation, a sensitivity analysis was performed to quantify the impact of blood transfusions and hemorrhage. The only significant difference in the sensitivity analysis was that cesarean delivery was more strongly associated with maternal morbidity and mortality; there was no significant difference in the AUC. We therefore infer that the impact of the unequal prevalence of the components of the composite does not change the conclusions of this study. Third, this study is restricted to preterm births only. While the data were collected from 33 different centers within the MFMU Network, most centers within the MFMU Network are large regional centers—there may be underrepresentation of the community hospitals. Still, most preterm deliveries systematically occur at large regional centers [31].
Future studies focusing on earlier gestational ages at delivery may better characterize risk factors for maternal morbidity in this population, given the increased risk observed at earlier gestational ages in this study. Additional studies that stratify based on indication for preterm birth may further elucidate these findings.
5. CONCLUSION
This work highlights that patients who give birth preterm have a 15% risk of maternal morbidity and mortality. Multiple obstetric and demographic factors have been associated with maternal morbidity and mortality for those who give birth preterm. Later gestational age was associated with lower rates of maternal morbidity and mortality. However, the obstetrical factors together were only modestly predictive for maternal morbidity and mortality following preterm birth.
AUTHOR CONTRIBUTIONS
Marcia K. Des Jardin: Writing—original draft; writing—review and editing. Paula L. McGee: Writing—original draft; writing—review and editing; formal analysis; data curation. Baha M. Sibai: Writing—review and editing. Rebecca G. Clifton: Writing—review and editing. Maged M. Costantine: Writing—review and editing. Michael W. Varner: Writing—review and editing. Monica Longo: Writing—review and editing. Edward F. Bell: Writing—review and editing. Barbara J. Stoll: Writing—review and editing. Donna Allard: Writing—review and editing. Ashley Salazar: Writing—review and editing. John M. Thorp Jr.: Writing—review and editing. Uma M. Reddy: Writing—review and editing. Emily S. Miller: Writing—review and editing. Dwight J. Rouse: Writing—review and editing. Hyagriv N. Simhan: Writing—review and editing. Alan T. N. Tita: Conceptualization; methodology; writing—original draft; writing—review and editing. Jennifer L. Bailit: Writing—review and editing. Lorraine Dugoff: Writing—review and editing. George A. Macones: Writing—review and editing. George R. Saade: Writing—review and editing.
CONFLICT OF INTEREST STATEMENT
The authors declare no conflicts of interest.
Supporting information
Supporting Information File 1: pmf270497‐sup‐0001‐TableS1.docx
Supporting Information File 2: pmf270497‐sup‐0002‐TableS2.docx
Supporting Information File 3: pmf270497‐sup‐0003‐TableS3.docx
Supporting Information File 4: pmf270497‐sup‐0004‐TableS4.docx
ACKNOWLEDGMENTS
The authors thank Lida Fette, MS, for study and data management and Elizabeth A. Thom, PhD, for protocol development and oversight. We thank Lynn Yee, MD, and Catherine Eppes, MD, for their thoughtful review of this manuscript. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health. This work is funded by the Eunice Kennedy Shriver National Institute of Child Health and Human Development (UG1 HD087230, UG1 HD027869, UG1 HD027915, UG1 HD034208, UG1 HD040500, UG1 HD040485, UG1 HD053097, UG1 HD040544, UG1 HD040545, UG1 HD040560, UG1 HD040512, UG1 HD087192, UG1 HD053109, UG1 HD027851, and U24 HD036801).
See the Supporting Information Appendix for a full list of members of the Eunice Kennedy Shriver NICHD MFMU Network
This manuscript was presented in abstract form at the 2024 SMFM Annual Pregnancy Meeting, February 2024, National Harbor, MD.
DATA AVAILABILITY STATEMENT
Deidentified participant data from the study will be available at the NICHD Data and Specimen Hub (DASH): https://dash.nichd.nih.gov/. Data sharing will be accessed and governed according to the procedures and policies of DASH. A data dictionary will be provided. This data will be provided within 12 months of publication.
REFERENCES
- 1. Matas, J. L. , Mitchell L. E., Sharma S. V., Louis J. M., and Salemi J. L.. 2021. “Severe Maternal Morbidity at Delivery and Postpartum Readmission in the United States.” Paediatric Perinatal Epidemiology 35(6): 627–34. 10.1111/ppe.12762. [DOI] [PubMed] [Google Scholar]
- 2. Zeitlin, J. , Egorova N. N., Janevic T., Hebert P. L., Lebreton E., Balbierz A., and Howell E. A.. 2019. “The Impact of Severe Maternal Morbidity on Very Preterm Infant Outcomes.” Journal of Pediatrics 215: 56–63.e1. 10.1016/j.jpeds.2019.07.061. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Hoyert, D. L. . 2023. Maternal Mortality Rates in the United States. 2021. Centers for Disease Control and Prevention. Accessed September 4, 2024. https://www.cdc.gov/nchs/data/hestat/maternal‐mortality/2021/maternal‐mortality‐rates‐2021.htm [Google Scholar]
- 4.“Maternal Morbidity and Mortality,” National Institute of Child Health and Human Development. Accessed September 4, 2024. https://www.nichd.nih.gov/health/topics/maternal‐morbidity‐mortality. [Google Scholar]
- 5.“Severe Maternal Morbidity,” Centers for Disease Control and Prevention. Accessed September 4, 2024. https://www.cdc.gov/maternal‐infant‐health/php/severe‐maternal‐morbidity/?CDC_AAref_Val=https://www.cdc.gov/reproductivehealth/maternalinfanthealth/severematernalmorbidity.html. [Google Scholar]
- 6. Reddy, U. M. , Rice M. M., Grobman W. A., Bailit J. L., Wapner R. J., Varner M. W., Thorp J. M., et al. 2015. “Serious Maternal Complications After Early Preterm Delivery (24–33 Weeks' Gestation).” American Journal of Obstetrics and Gynecology 213(4): 538.e1–e9. 10.1016/j.ajog.2015.06.064. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Kilpatrick, S. J. , Abreo A., Gould J., Greene N., and Main E. K.. 2016. “Confirmed Severe Maternal Morbidity Is Associated With High Rate of Preterm Delivery.” American Journal of Obstetrics and Gynecology 215(2): 233.e1–e7. 10.1016/j.ajog.2016.02.026. [DOI] [PubMed] [Google Scholar]
- 8. Phillips, J. M. , Bodnar L. M., and Himes K. P.. 2023. “Association Between Gestational Age at Delivery and Indicator‐Specific Severe Maternal Morbidity.” Journal of Maternal‐Fetal and Neonatal Medicine 36(1): 2198633. 10.1080/14767058.2023.2198633 [DOI] [PubMed] [Google Scholar]
- 9. Korst, L. M. , Gregory K. D., Nicholas L. A., Saeb S., Reynen D. J., Troyan J. L., Greene N., and Fridman M.. 2021. “A Scoping Review of Severe Maternal Morbidity: Describing Risk Factors and Methodological Approaches to Inform Population‐Based Surveillance.” Maternal Health, Neonatology and Perinatology 7(1): 3. 10.1186/s40748-020-00123-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Rosenbloom, J. I. , Tuuli M. G., Stout M. J., Young O. M., Woolfolk C. L., Lopez J. D., Macones G. A., and Cahill A. G.. 2019. “A Prediction Model for Severe Maternal Morbidity in Laboring Patients at Term.” American Journal of Perinatology 36(1): 8–14. 10.1055/s-0038-1626716. [DOI] [PubMed] [Google Scholar]
- 11. Bateman, B. T. , Mhyre J. M., Hernandez‐Diaz S., Huybrechts K. F., Fischer M. A., Creanga A. A., Callaghan W. M., and Gagne J. J.. 2013. “Development of a Comorbidity Index for Use in Obstetric Patients.” Obstetrics and Gynecology, 122(5): 957–65. 10.1097/AOG.0b013e3182a603bb. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Gray, K. E. , Wallace E. R., Nelson K. R., Reed S. D., and Schiff M. A.. 2012. “Population‐Based Study of Risk Factors for Severe Maternal Morbidity.” Paediatric Perinatal Epidemiology 26(6): 506–14. 10.1111/ppe.12011. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Geller, S. E. , Rosenberg D., Cox S. M., Brown M. L., Simonson L., Driscoll C. A., and Kilpatrick S. J.. 2004. “The Continuum of Maternal Morbidity and Mortality: Factors Associated With Severity.” American Journal of Obstetrics and Gynecology 191(3): 939–44. 10.1016/j.ajog.2004.05.099. [DOI] [PubMed] [Google Scholar]
- 14. Joseph, K. S. , Liu S., Rouleau J., Kirby R. S., Kramer M. S., Sauve R., Fraser W. D., Young D. C., and Liston R. M.. 2010. “Severe Maternal Morbidity in Canada, 2003 to 2007: Surveillance Using Routine Hospitalization Data and ICD‐10CA Codes.” Journal of Obstetrics and Gynaecology Canada 32(: 837–46. 10.1016/S1701-2163(16)34655-2. [DOI] [PubMed] [Google Scholar]
- 15. Zwart, J. J. , Richters J. M., Ory F., de Vries J. I., Bloemenkamp K. W., and van Roosmalen J..2008. “Severe Maternal Morbidity During Pregnancy, Delivery and Puerperium in the Netherlands: A Nationwide Population‐Based Study of 371,000 Pregnancies.” British Journal of Obstetrics and Gynaecology 115(7): 842–50. 10.1111/j.1471-0528.2008.01713.x. [DOI] [PubMed] [Google Scholar]
- 16. Zhang, J. , Meikle S., and Trumble A.. 2003. “Severe Maternal Morbidity Associated With Hypertensive Disorders in Pregnancy in the United States.” Hypertension in Pregnancy 22(2): 203–12. 10.1081/PRG-120021066. [DOI] [PubMed] [Google Scholar]
- 17. Carr, R. C. , McKinney D. N., Cherry A. L., and Defranco E. A.. 2022. “Maternal Age‐Specific Drivers of Severe Maternal Morbidity.” American Journal of Obstetrics and Gynecology Maternal Fetal Medicine 4(2): 100529. 10.1016/j.ajogmf.2021.100529. [DOI] [PubMed] [Google Scholar]
- 18. Doherty, B. T. , Lynch S., Naavaal A., Li C., Cole K., MacPhee L., Banning L., et al. 2025. “Maternal and Infant Morbidity and Mortality in Relation to Delivery Mode in a Large U.S. Health Care Claims Database in 2019 and 2020.” American Journal of Perinatology 42(6): 758–67. 10.1055/a-2419-8916. [DOI] [PubMed] [Google Scholar]
- 19. Bryant, A. S. , Worjoloh A., Caughey A. B., and Washington A. E.. 2010. “Racial/Ethnic Disparities in Obstetric Outcomes and Care: Prevalence and Determinants.” American Journal of Obstetrics and Gynecology. 202(4): 335–43. 10.1016/j.ajog.2009.10.864. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Subramaniam, A. , Cliver S. S., Smeltzer S., Tita A. T. N., and Wetta L. L.. 2016. “Preterm Premature Rupture of Membranes (PPROM): Outcomes of Delivery at 32(0/7)‐33(6/7) Weeks After Confirmed Fetal Lung Maturity (FLM) Versus Expectant Management Until 34(0/7) Weeks.” Journal of Maternal‐Fetal and Neonatal Medicine, 29 12: 895–9. 10.3109/14767058.2015.1074996. [DOI] [PubMed] [Google Scholar]
- 21. Tran, S. H. , Cheng Y. W., Kaimal A. J., and Caughey A. B.. 2008. “Length of Rupture of Membranes in the Setting of Premature Rupture of Membranes at Term and Infectious Maternal Morbidity.” American Journal of Obstetrics and Gynecology 198(6): 700 e1–5. 10.1016/j.ajog.2008.03.031. [DOI] [PubMed] [Google Scholar]
- 22. Waters, T. P. , and Mercer B. M.. 2009. “The Management of Preterm Premature Rupture of the Membranes Near the Limit of Fetal Viability.” American Journal of Obstetrics and Gynecology 201(3): 30–40. 10.1016/j.ajog.2009.06.049. [DOI] [PubMed] [Google Scholar]
- 23. Tita, A. T. N. , Doherty L., Roberts J. M., Myatt L., Leveno K. J., Varner M. W., Wapner R. J., et al. 2018. “Adverse Maternal and Neonatal Outcomes in Indicated Compared With Spontaneous Preterm Birth in Healthy Nulliparas: A Secondary Analysis of a Randomized Trial.” American Journal of Perinatology 35(7): 24–631. 10.1055/s-0037-1608787. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Li, Y. , Tian Y., Liu N., Chen Y., and Wu F.. 2019. “Analysis of 62 Placental Abruption Cases: Risk Factors and Clinical outcomes.” Taiwan Journal of Obstetrics and Gynecology 58(2): 223–6. 10.1016/j.tjog.2019.01.010. [DOI] [PubMed] [Google Scholar]
- 25. Salman, L. , Aviram A., Holzman R., Hay‐Azogui H., Ashwal E., Hadar E., and Gabbay‐Benziv R.. 2020. “Predictors for Cesarean Delivery in Preterm Premature Rupture of Membranes.” Journal of Maternal‐Fetal and Neonatal Medicine 33(22): 3761–6. 10.1080/14767058.2019.1585422. [DOI] [PubMed] [Google Scholar]
- 26. Kim, L. H. , Cheng Y. W., Delaney S., Jelin A. C., and Caughey A. B.. 2010. “Is Preeclampsia Associated With an Increased Risk of Cesarean Delivery If Labor is Induced?” Journal of Maternal‐Fetal and Neonatal Medicine 23(5): 383–8. 10.3109/14767050903168432. [DOI] [PubMed] [Google Scholar]
- 27. Young, B. C. , and Wylie B. J.. 2012. “Effects of Twin Gestation on Maternal Morbidity.” Seminars in Perinatology 36(3): 162–8. 10.1053/j.semperi.2012.02.007. [DOI] [PubMed] [Google Scholar]
- 28. Seyb, S. T. , Berka R. J., Socol M. L., and Dooley S. L.. 1999. “Risk of Cesarean Delivery With Elective Induction of Labor at Term in Nulliparous Women.” Obstetrics and Gynecology 94(4): 600–7. 10.1016/s0029-7844(99)00377-4. [DOI] [PubMed] [Google Scholar]
- 29. Malloy, M. H. , Rhoads G. G., Schramm W., and Land G.. 1989. “Increasing Cesarean Section Rates in Very Low‐Birth Weight Infants. Effect on Outcome.” Journal of the American Medical Association 262(11): 1475–8. [PubMed] [Google Scholar]
- 30. Nwabuobi, C. , Gowda N., Schmitz J., Wood N., Pargas A., Bagiardi L., Odibo L., et al. 2020. “Risk Factors for Cesarean Delivery in Pregnancy With Small‐for‐Gestational‐Age Fetus Undergoing Induction of Labor.” Ultrasound in Obstetrics and Gynecology 55(6): 799–805. 10.1002/uog.20850. [DOI] [PubMed] [Google Scholar]
- 31. 2019. “Levels of Maternal Care: Obstetric Care Consensus No. 9.” Obstetrics and Gynecology. 134(2): e41–55. 10.1097/AOG.0000000000003383. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Supporting Information File 1: pmf270497‐sup‐0001‐TableS1.docx
Supporting Information File 2: pmf270497‐sup‐0002‐TableS2.docx
Supporting Information File 3: pmf270497‐sup‐0003‐TableS3.docx
Supporting Information File 4: pmf270497‐sup‐0004‐TableS4.docx
Data Availability Statement
Deidentified participant data from the study will be available at the NICHD Data and Specimen Hub (DASH): https://dash.nichd.nih.gov/. Data sharing will be accessed and governed according to the procedures and policies of DASH. A data dictionary will be provided. This data will be provided within 12 months of publication.