Interpregnancy weight change: associations with severe maternal morbidity and neonatal outcomes

Abstract

BACKGROUND:

Prepregnancy body mass index and gestational weight gain have been linked with severe maternal morbidity, suggesting that weight change between pregnancies may also play a role, as it does for neonatal outcomes.

OBJECTIVE:

This study assessed the association of changes in prepregnancy body mass index between 2 consecutive singleton pregnancies with the outcomes of severe maternal morbidity, stillbirth, and small- and large-for-gestational-age infants in the subsequent pregnancy.

STUDY DESIGN:

This observational study was based on birth records from 1,111,032 consecutive pregnancies linked to hospital discharge records in California (2007–2017). Interpregnancy body mass index change between the beginning of an index pregnancy and the beginning of the subsequent pregnancy was calculated from self-reported weight and height. Severe maternal morbidity was defined based on the Centers for Disease Control and Prevention index, including and excluding transfusion-only cases. We used multivariable log-binomial regression models to estimate adjusted risks, overall and stratified by prepregnancy body mass index at index birth.

RESULTS:

Substantial interpregnancy body mass index gain (≥4 kg/m²) was associated with severe maternal morbidity in crude but not adjusted analyses. Substantial interpregnancy body mass index loss (>2 kg/m²) was associated with increased risk of severe maternal morbidity (adjusted relative risk, 1.13; 95% confidence interval (1.07–1.19), and both substantial loss (adjusted relative risk, 1.11 [1.02–1.19]) and gain (≥4 kg/m²; adjusted relative risk, 1.09 [1.02–1.17]) were associated with nontransfusion severe maternal morbidity. Substantial loss (adjusted relative risk, 1.17 [1.05–1.31]) and gain (1.26 [1.14–1.40]) were associated with stillbirth. Body mass index gain was positively associated with large-for-gestational-age infants and inversely associated with small-for-gestational-age infants.

CONCLUSION:

Substantial interpregnancy body mass index changes were associated with modestly increased risk of severe maternal morbidity, stillbirth, and small- and large-for-gestational-age infants.

Introduction

Severe maternal morbidity (SMM) includes unexpected potentially life-threatening complications related to pregnancy and childbirth such as eclampsia, sepsis, or acute heart failure. SMM identifies individuals most at risk of maternal mortality, which occurs in 17.4 per 100,000 births.¹ When assessed using the SMM Index developed by the US Centers for Disease Control and Prevention (CDC), the overall rate of SMM doubled between 1993 and 2014. Currently, SMM affects approximately 50,000 or 1.4% of all US births annually.²

Both high and low prepregnancy body mass index (BMI) are established risk factors for adverse infant outcomes. Infants born to individuals with obesity are at increased risks for various adverse outcomes, such as stillbirth, preterm birth, infant mortality, being large for gestational age (LGA), child obesity, and neurocognitive problems.^3–5 Infants born to individuals with low prepregnancy BMI are also at risk of adverse outcomes, including preterm birth, low birthweight, being small for gestational age (SGA), delayed neurodevelopment in early life, and increased hospital costs.^4–7 Recent studies have expanded to consider maternal health, with reported linkage of high and low prepregnancy BMI to SMM,^8–11 and 1 study reporting no association.¹²

Gestational weight change, postpartum health and weight retention, and age-related health and weight changes all affect interpregnancy weight change, which may in turn contribute to prepregnancy obesity or underweight at the beginning of a subsequent pregnancy. Increases in BMI between pregnancies have been associated with hypertensive disorders, gestational diabetes mellitus, stillbirth, and LGA infants in the subsequent pregnancy, and decreases in BMI between pregnancies have been associated with SGA infants and preterm birth.^13–16 Thus, this study investigated whether BMI gain or loss between pregnancies contributes to the risk of SMM, which may have similar underlying causes to those of previously studied outcomes. We hypothesized a U-shaped association between extremes of interpregnancy BMI change and SMM, and that this association may differ by prepregnancy BMI. We also sought to confirm previously reported associations of interpregnancy BMI change with stillbirth, SGA, and LGA in this study population.

Materials and Methods

Sample selection

We examined vital statistics records linked with maternal and infant hospital discharge records from antepartum, delivery, and postpartum hospitalizations, on the basis of files available from the California Department of Health Care Access and Information. There were 5,496,609 live and stillborn births between 2007 (when California began to collect data allowing calculation of BMI) and 2017 (the most current data available). We limited the analytical cohort to individuals with ≥2 consecutive singleton births with complete and plausible data on maternal body size, study outcomes, and covariates, as described in detail in Figure S1. Of 1,341,956 pairs of consecutive singleton births, 130,255 pairs (9.7%) were excluded because of missing or implausible maternal height, prepregnancy weight, or prepregnancy BMI.¹⁷ After calculating interpregnancy BMI change, we trimmed outliers (30,563 pairs with values beyond the 1st and 99th percentile). Of the remaining 1,181,139 pairs, 70,106 (5.5%) had missing covariates and were excluded. Our final analytical cohort consisted of 1,111,032 pairs of singleton–singleton births; within each pair, we identified the first and second birth as the index and subsequent birth, respectively. Individuals excluded due to missing or implausible data, had high association with Black race/ethnicity, young age, low education, private insurance, prepregnancy obesity, high interpregnancy weight loss or gain, and SMM and stillbirth in the subsequent pregnancy (Table S1).

Study outcomes

SMM was defined using the CDC SMM Index (a composite of 21 indicators identified from International Classification of Diseases codes, described in Table S2) from hospital discharge records from birth through 42 days postpartum.² Half of all SMM occurrences classified by this index are based solely on the code for blood transfusion as a proxy for hemorrhage, and without additional information on number of units of blood, this increases false positives by including individuals with potentially less severe morbidity.¹⁸ We therefore also examined nontransfusion SMM (SMM with an indicator other than transfusion alone), which the CDC now commonly reports. We designated maternal deaths as SMM cases only if they had diagnostic or procedure codes that identified them as having SMM. Infant outcomes included stillbirth (birth of a fetus that died in utero and was born at ≥20 weeks’ gestation) and fetal size for gestational age as LGA (birthweight above the 90th percentile for infants of the same sex and weeks of gestational age at birth) and SGA (birthweight <10th percentile) based on the distribution of all 5,321,133 singleton California births from 2007 to 2017 that had nonmissing data on gestational age and birthweight.

Maternal interpregnancy body mass index change

Maternal height and prepregnancy weight were recorded in the vital record and collected either by self-report or measurement (the vital record does not provide information on the source of the data). We calculated BMI (weight [kg]/height [m²]) and categorized it as underweight (<18.5), normal weight (18.5–24.9), overweight (25–29.9), obesity I (30–34.9), obesity II (35–39.9), and obesity III (≥40). Interpregnancy BMI change was defined as the difference in prepregnancy BMI between the index and subsequent birth and categorized into 6 levels: <−2, −2 to <−1, −1 to <1, 1 to <2, 2 to <4, and ≥4 kg/m².¹⁹ Henceforth, we refer to the most extreme category of interpregnancy BMI gain and loss as “substantial” and indicate values for other categories.

Covariates

We used a directed acyclic graph to select confounders based on known associations with maternal weight and perinatal health outcomes.^16,20 We adjusted for interpregnancy interval and the following variables from the index pregnancy: maternal race or ethnicity, age (continuous), education, height (continuous), prepregnancy BMI, principal source of payment for delivery, trimester of prenatal care initiation, gestational age at delivery (continuous), mode of birth, parity, smoking during pregnancy, birth year (continuous), SMM, stillbirth, and obstetrical comorbidity score. Interpregnancy interval for each pair of pregnancies was defined as the difference between the date of birth of the subsequent birth and index birth, minus the gestational age of the subsequent birth, and converted to months. The obstetrical comorbidity score weights the contribution of 26 specific maternal conditions to SMM and nontransfusion SMM.^21,22 We did not adjust for conditions occurring during the second pregnancy of each pair because they are on the causal pathway and could be mediators.¹⁶ Variables involving diagnostic or procedure codes (ie, SMM indicators and most comorbidities) were obtained from hospital discharge data; other variables (eg, age, parity, smoking) were obtained from vital records. We used data on hypertensive disorders and gestational diabetes mellitus reported in both data sources.

Statistical analysis

We used generalized estimating equation models with an exchangeable correlation structure to estimate crude and adjusted relative risks (aRR) for the association of interpregnancy BMI change with each outcome while taking into account the correlation between repeat pairs of pregnancies in the same individual. The models adjusted for all covariates in Table 1. A priori, we expected possible differences by prepregnancy BMI category at the index pregnancy; therefore, we also performed unadjusted and adjusted analyses stratified by “lower” (<25 kg/m²) vs “higher” (≥25 kg/m²) BMI.^19,23 Analyses were conducted in SAS version 9.4 (SAS Institute, Cary, NC). The State of California Committee for the Protection of Human Subjects, the Stanford University Research Compliance Office and the University of California, Berkeley Committee for Protection of Human Subjects approved this study.

TABLE 1.

Maternal characteristics in index pregnancy, interpregnancy body mass index change, and study outcomes in the subsequent pregnancy for the California birth cohort with 2 singleton consecutive births (n=1,111,032)

Characteristics	n (%)
Maternal characteristics in index pregnancy
Maternal race/ethnicity
US-born Hispanic	380,513 (34.2)
Foreign-born Hispanic	124,550 (11.2)
Non-Hispanic White	377,784 (34.0)
Asian/Pacific Islander	151,656 (13.7)
African-American	69,396 (6.2)
American Indian or other specified	7133(0.6)
Age, y
<20	118,874 (10.7)
20–24	291,631 (26.2)
25–29	337,051 (30.3)
30–34	269,171 (24.2)
35–39	86,329 (7.8)
≥40	7976 (0.7)
Education
Less than high school graduate	174,632 (15.7)
High school graduate	306,140 (27.6)
Some college	307,637 (27.7)
Undergraduate degree	212,496 (19.1)
Postgraduate degree	110,127(9.9)
Height, cm
≤159	366,216(33.0)
160–164	313,886 (28.3)
165–169	238,363 (21.5)
≥170	192,567 (17.3)
Prepregnancy BMI
Underweight	49,448 (4.5)
Normal weight	578,545 (52.1)
Overweight	270,003 (24.3)
Obesity I	129,513(11.7)
Obesity II	54,059 (4.9)
Obesity III	29,464 (2.7)
Principal source of payment for delivery
Medicaid	601,878(54.2)
Commercial insurance	483,552 (43.5)
Self-pay or other	25,602 (2.3)
Prenatal care began
First trimester	922,423 (83.0)
Second trimester	143,564 (12.9)
Third trimester or none	31,292 (2.8)
Unknown	13,753 (1.2)
Birth status and gestational age at delivery
Live birth, <37 wk	73,091 (6.6)
Live birth, ≥37 wk	1,032,152 (92.9)
Stillbirth	5789 (0.5)
Cesarean delivery
No	805,823 (72.5)
Yes	305,209 (27.5)
Parity (number of previous live births)
0	619,520(55.8)
1–4	475,982 (42.8)
≥5	15,530 (1.4)
Smoking during pregnancy
No	1,081,369(97.3)
Yes	29,663 (2.7)
Birth year
2007	152,347 (13.7)
2008	148,893 (13.4)
2009	146,421 (13.2)
2010	139,163 (12.5)
2011	131,733 (11.9)
2012	122,721 (11.0)
2013	107,812(9.7)
2014	88,740 (8.0)
2015	56,416(5.1)
2016	16,600 (1.5)
2017	186 (<0.1)
SMMa
No SMM	1,097,179(98.8)
SMM (total)	13,853 (1.2)
Nontransfusion SMM	5657 (0.5)
Comorbidity score for predicting SMM	Median (Q1-Q3): 0(0–11) Mean (SD): 7.0 (13.4)
Comorbidity score for predicting nontransfusion SMM	Median (Q1-Q3): 0 (0–6) Mean (SD): 3.6 (7.6)
Interpregnancy characteristics
Interpregnancy BMI change (kg/m2)
<−2	109,855 (9.9)
−2 to <−1	83,356 (7.5)
−1 to <1	409,226 (36.8)
1 to <2	158,903 (14.3)
2 to <4	188,286 (16.9)
≥4	161,406 (14.5)
Interpregnancy interval
<1 y	252,997 (22.8)
1 to <2 y	358,727 (32.3)
2 to ≥3 y	212,105 (19.1)
>3 y	287,203 (25.9)
Study outcomes in subsequent pregnancy
SMMa
No SMM	1,094,470(98.5)
SMM (total)	16,562 (1.5)
Nontransfusion SMM	7866 (0.7)
No	1,107,340(99.7)
Yes	3692 (0.3)
Small for gestational age
No	1,025,449(92.3)
Yes	85,496 (7.7)
Large for gestational age
No	978,317(88.1)
Yes	132,628 (11.9)

BMI, body mass index; SD, standard deviation; SMM, severe maternal morbidity; Q1-Q3, first quartile to third quartile.

^aNontransfusion SMM refers to the subset of cases for which transfusion was not the only qualifying SMM indicator. Abrams. Interpregnancy weight change and severe maternal morbidity. Am J Obstet Gynecol MFM 2022.

Results

Among the 1,111,032 people included in the analyses, most were Hispanic or non-Hispanic White, 20 to 29 years old, had greater than a high school education, had a normal prepregnancy BMI, Medicaid insurance, started prenatal care in the first trimester, and were nulliparous at the index pregnancy (Table 1). More than half had interpregnancy intervals of <24 months between the index and subsequent pregnancies. A stable BMI (−1 to +<1 kg/m² change) between births was noted in 37% of participants; 10% had substantial BMI loss (>2 kg/m²) and 14% had substantial BMI gain (≥4 kg/m²). In the subsequent pregnancy, the prevalence of outcomes overall was 1.5% for SMM, 0.7% for nontransfusion SMM, 0.3% for stillbirth, 12% for LGA, and 8% for SGA.

Table 2 shows associations between interpregnancy BMI change and the study outcomes, using the stable BMI group as reference. Crude analyses suggested modest U-shaped risks for both definitions of SMM, which were attenuated in adjusted models. Substantial weight loss was associated with both SMM (aRR, 1.13; 95% confidence interval [CI], 1.07–1.19) and nontransfusion SMM (aRR, 1.11; 95% CI, 1.02–1.19). Substantial BMI gain was also associated with nontransfusion SMM (aRR, 1.09; 95% CI, 1.02–1.17). For stillbirth, substantial BMI loss (aRR, 1.17; 95% CI, 1.05–1.31), substantial BMI gain (aRR, 1.26; 95% CI, 1.14–1.40), and BMI gain of 2 to <4 units (aRR, 1.18; 95% CI, 1.07–1.30) were associated with increased risk. For LGA, adjusted risks indicated increasing risk with increasing BMI gain, and for SGA the opposite was observed.

TABLE 2.

Association of interpregnancy body mass index change between 2 consecutive singleton pregnancies with outcomes in the second pregnancy

Variables BMI change	Cases	Noncases	Crude RR		Adjusted RRa
			cRR (95% CI)	P value	aRR (95% CI)	P value
SMMb	16,562 (1.49%)	1,094,470(98.51%)
<−2	1990 (12.0)	107,865(9.9)	1.33(1.26–1.40)	<.01	1.13(1.07–1.19)	<.01
−2 to <−1	1229(7.4)	82,127(7.5)	1.08(1.02–1.15)	.01	1.02(0.96–1.09)	.47
−1 to <1	5579 (33.7)	403,647 (36.9)	Ref	Ref	Ref	Ref
1 to <2	2261 (13.7)	156,642 (14.3)	1.04(0.99–1.10)	.08	1.00(0.95–1.05)	.92
2 to <4	2845 (17.2)	185,441 (16.9)	1.11 (1.06–1.16)	<.01	1.02(0.97–1.06)	.46
≥4	2658 (16.0)	158,748 (14.5)	1.21 (1.15–1.26)	<.01	1.04 (1.00–1.10)	.07
Nontransfusion SMMb	7866(0.71%)	1,094,470(99.29%)
<−2	930 (11.8)	107,865(9.9)	1.33(1.23–1.43)	<.01	1.11 (1.02–1.19)	.01
−2 to <−1	592 (7.5)	82,127(7.5)	1.11 (1.02–1.22)	.02	1.04(0.95–1.14)	.34
−1 to <1	2614(33.2)	403,647 (36.9)	Ref	Ref	Ref	Ref
1 to <2	1091 (13.9)	156,642 (14.3)	1.07(1.00–1.15)	.04	1.03(0.96–1.11)	.36
2 to <4	1353 (17.2)	185,441 (16.9)	1.13(1.05–1.20)	<.01	1.04(0.97–1.11)	.25
≥4	1286 (16.3)	158,748 (14.5)	1.25(1.17–1.33)	<.01	1.09 (1.02–1.17)	.02
Fetal outcomes
Stillbirth	3692 (0.33%)	1,107,340(99.67%)
<−2	458 (12.4)	109,397(9.9)	1.50(1.34–1.67)	<.01	1.17(1.05–1.31)	<.01
−2 to <−1	270 (7.3)	83,086 (7.5)	1.16(1.02–1.33)	.03	1.06(0.93–1.21)	.38
−1 to <1	1140(30.9)	408,086 (36.9)	Ref	Ref	Ref	Ref
1 to <2	467 (12.6)	158,436 (14.3)	1.05(0.95–1.17)	.33	1.01 (0.91–1.13)	.83
2 to <4	682 (18.5)	187,604 (16.9)	1.30(1.18–1.43)	<.01	1.18(1.07–1.30)	<.01
≥4	675 (18.3)	160,731 (14.5)	1.50(1.37–1.65)	<.01	1.26 (1.14–1.40)	<.01
LGA	132,628 (11.94%)	978,317(88.06%)
<−2	12,818(9.7)	97,024 (9.9)	1.08(1.06–1.10)	<.01	0.91 (0.90–0.93)	<.01
−2 to <−1	9308 (7.0)	74,041 (7.6)	1.03(1.01–1.05)	<.01	0.95 (0.93–0.97)	<.01
−1 to <1	44,377 (33.5)	364,825 (37.3)	Ref	Ref	Ref	Ref
1 to <2	18,906 (14.3)	139,989 (14.3)	1.10(1.08–1.11)	<.01	1.11 (1.09–1.13)	<.01
2 to <4	24,117(18.2)	164,149 (16.8)	1.18(1.16–1.20)	<.01	1.17(1.16–1.19)	<.01
≥4	23,102 (17.4)	138,289 (14.1)	1.32(1.30–1.34)	<.01	1.29 (1.27–1.31)	<.01
SGA	85,496 (7.70%)	1,025,449(92.30%)
<−2	9696 (11.3)	100,146(9.8)	1.13(1.10–1.15)	<.01	1.16(1.14–1.19)	<.01
−2 to <−1	6861 (8.0)	76,488 (7.5)	1.05(1.02–1.08)	<.01	1.08 (1.05–1.11)	<.01
−1 to <1	32,065 (37.5)	377,137(36.8)	Ref	Ref	Ref	Ref
1 to <2	11,855 (13.9)	147,040 (14.3)	0.95 (0.93–0.97)	<.01	0.92 (0.90–0.94)	<.01
2 to <4	13,749 (16.1)	174,517(17.0)	0.93(0.91–0.95)	<.01	0.88 (0.87–0.90)	<.01
≥4	11,270 (13.2)	150,121 (14.6)	0.89(0.87–0.91)	<.01	0.83(0.81–0.85)	<.01

aRR, adjusted relative risk; BMI, body mass index; CDC, Centers for Disease Control and Prevention; Cl, confidence interval; cRR, CRUDE relative; LGA, large for gestational age; Ref, reference interval; SGA, small for gestational age; SMM, severe maternal morbidity.

^aAdjusted for maternal race/ethnicity, age (continuous), education, height (continuous), prepregnancy BMI, principal source of payment for delivery, timing of prenatal care, gestational age at delivery (continuous), mode of birth, parity, smoking during pregnancy, birth year (continuous), SMM, comorbidity score (continuous), and stillbirth−all from the index birth, and interpregnancy interval (continuous);

^bSMM defined by CDC index. Nontransfusion SMM refers to cases for which transfusion was not the only indicator present.

Abrams. Interpregnancy weight change and severe maternal morbidity. Am J Obstet Gynecol MFM 2022.

The prevalence of SMM was 1.6% in the higher BMI group (≥25 kg/m²) and 1.4% in the lower BMI group (<25 kg/m²) (Table 3). Almost half of the lower BMI group maintained their BMI between pregnancies, whereas only 30% of the higher BMI group did so. Compared with the lower BMI group, a larger proportion of the higher BMI group had substantial interpregnancy BMI loss (17.2% vs 4.3%) (data not shown). The magnitude of association between substantial interpregnancy BMI loss and SMM overall was similar by prepregnancy BMI group but was more strongly related to risk of nontransfusion SMM in the lower BMI group than the higher BMI group. In addition, for the higher BMI group, a 1 to <2 kg/m² BMI gain was associated with increased risk of SMM and nontransfusion SMM (9% and 14% increased risks, respectively). For infant outcomes, the pattern of results was similar, but adjusted associations tended to be slightly stronger in the lower BMI group than the higher BMI group.

TABLE 3.

Association of interpregnancy body mass index change between 2 consecutive singleton pregnancies with outcomes in the second pregnancy, stratified by prepregnancy body mass index from the index pregnancy

	BMI <25			BMI >25
BMI change	Cases	cRR (95% CI)	aRR (95% CI)a	Cases	cRR (95% CI)	aRR (95% CI)a
SMMb	8735 (1.39%)			7827 (1.62%)
<−2	441 (5.0)	1.28 (1.16–1.41)	1.11 (1.01–1.23)	1549 (19.8)	1.23 (1.15–1.32)	1.13(1.06–1.21)
−2 to <−1	566 (6.5)	1.05(0.96–1.15)	1.01 (0.92–1.10)	663 (8.5)	1.05(0.96–1.15)	1.03(0.94–1.12)
−1 to <1	3525 (40.4)	Ref	Ref	2054 (26.2)	Ref	Ref
1 to <2	1301 (14.9)	1.00(0.94–1.07)	0.94(0.88–1.00)	960 (12.3)	1.09 (1.01–1.17)	1.09 (1.01–1.18)
2 to <4	1588 (18.2)	1.17(1.10–1.24)	1.03(0.97–1.10)	1257 (16.1)	1.01 (0.94–1.08)	0.99(0.93–1.07)
≥4	1314(15.0)	1.28 (1.20–1.36)	1.04(0.97–1.11)	1344 (17.2)	1.09 (1.02–1.16)	1.04(0.97–1.12)
Nontransfusion SMMb	3986 (0.64%)			3880(0.81%)
<−2	207 (5.2)	1.32 (1.14–1.52)	1.22 (1.05–1.41)	723 (18.6)	1.17(1.07–1.29)	1.09(0.99–1.19)
−2 to <−1	261 (6.5)	1.07(0.94–1.22)	1.05(0.92–1.19)	331 (8.5)	1.07(0.94–1.21)	1.05(0.93–1.18)
−1 to <1	1605(40.3)	Ref	Ref	1009(26.0)	Ref	Ref
1 to <2	595 (14.9)	1.01 (0.92–1.11)	0.95(0.87–1.05)	496 (12.8)	1.15(1.03–1.28)	1.14(1.03–1.27)
2 to <4	699 (17.5)	1.13(1.03–1.23)	1.03(0.94–1.13)	654 (16.9)	1.06(0.97–1.17)	1.05(0.95–1.16)
≥4	619(15.5)	1.32 (1.21–1.45)	1.14(1.03–1.26)	667 (17.2)	1.10(1.00–1.21)	1.05(0.95–1.16)
Fetal outcomes
Stillbirth	1773(0.28%)			1919(0.40%)
<−2	88 (5.0)	1.37 (1.09–1.71)	1.20(0.96–1.50)	370 (19.3)	1.25 (1.09–1.43)	1.16(1.02–1.33)
−2 to <−1	113(6.4)	1.13(0.92–1.38)	1.09(0.89–1.32)	157(8.2)	1.06(0.88–1.27)	1.04(0.87–1.25)
−1 to <1	657(37.1)	Ref	Ref	483 (25.2)	Ref	Ref
1 to <2	251 (14.2)	1.04(0.90–1.20)	0.99(0.86–1.15)	216(11.3)	1.04(0.89–1.22)	1.03(0.88–1.21)
2 to <4	354 (20.0)	1.39 (1.23–1.59)	1.27 (1.11–1.45)	328 (17.1)	1.12(0.97–1.28)	1.09(0.95–1.26)
≥4	310(17.5)	1.62 (1.41–1.85)	1.36 (1.18–1.57)	365 (19.0)	1.26 (1.10–1.44)	1.19(1.03–1.37)
LGA	57,735(9.19%)			74,893 (15.51%)
<−2	1849(3.2)	0.80 (0.76–0.84)	0.91 (0.87–0.96)	10,969 (14.6)	0.86 (0.84–0.88)	0.92 (0.90–0.94)
−2 to <−1	3188(5.5)	0.89 (0.86–0.92)	0.92 (0.89–0.95)	6120(8.2)	0.96 (0.93–0.98)	0.98(0.95–1.00)
−1 to <1	23,587 (40.9)	Ref	Ref	20,790 (27.8)	Ref	Ref
1 to <2	9618(16.7)	1.11 (1.08–1.13)	1.18(1.15–1.20)	9288 (12.4)	1.04 (1.02–1.06)	1.05 (1.02–1.07)
2 to <4	10,522 (18.2)	1.15(1.13–1.18)	1.29 (1.26–1.31)	13,595 (18.2)	1.07 (1.05–1.10)	1.09 (1.07–1.12)
≥4	8971 (15.5)	1.30 (1.27–1.34)	1.48 (1.44–1.51)	14,131 (18.9)	1.13(1.11–1.15)	1.19(1.17–1.21)
SGA	54,454 (8.67%)			31,042(6.43%)
<−2	3102(5.7)	1.34 (1.29–1.38)	1.13(1.09–1.17)	6594(21.2)	1.29(1.25–1.33)	1.16(1.12–1.20)
−2 to <−1	4108(7.5)	1.14(1.10–1.17)	1.08 (1.05–1.11)	2753 (8.9)	1.07 (1.03–1.12)	1.04 (1.00–1.09)
−1 to <1	23,712(43.5)	Ref	Ref	8353 (26.9)	Ref	Ref
1 to <2	8328 (15.3)	0.95 (0.93–0.98)	0.90 (0.88–0.92)	3527 (11.4)	0.98(0.95–1.02)	0.98(0.94–1.02)
2 to <4	8816(16.2)	0.96 (0.94–0.99)	0.86 (0.84–0.88)	4933 (15.9)	0.97(0.94–1.00)	0.94(0.91–0.98)
≥4	6388 (11.7)	0.92 (0.90–0.95)	0.79(0.77–0.81)	4882 (15.7)	0.97(0.94–1.01)	0.89 (0.86–0.92)

aRR, adjusted relative risk; BMI, body mass index; CDC, Centers for Disease Control and Prevention; CI, confidence interval; cRR cumulative relative risk; LGA, large for gestational age; Ref, reference interval; SGA, small for gestational age; SMM, severe maternal morbidity.

^aAdjusted for maternal race/ethnicity, age (continuous), education, height (continuous), principal source of payment for delivery, timing of prenatal care, gestational age at delivery (continuous), mode of birth, parity, smoking during pregnancy, birth year (continuous), SMM, comorbidity score (continuous), and stillbirth−all from the index birth, and interpregnancy interval (continuous);

^bSMM defined by CDC index. Nontransfusion SMM refers to cases for which transfusion was not the only indicator present.

Abrams. Interpregnancy weight change and severe maternal morbidity. Am J Obstet Gynecol MFM 2022.

Discussion

Principal findings

This study investigated the relationship between interpregnancy BMI change and SMM. In this diverse sample of almost 900,000 individuals with at least 2 consecutive births over an 11-year period, multivariable analysis suggested that a substantial BMI loss (>2 kg/m²) between the index and subsequent pregnancies was associated with increased risks of 12% for SMM and 11% for nontransfusion SMM. Substantial BMI gain (≥4 kg/m²) was associated with a 9% increase in nontransfusion SMM, independent of confounding factors. We found a U-shaped relationship between BMI change and stillbirth risk, a positive relationship between BMI gain and LGA, and an inverse relationship between BMI gain and SGA. After stratifying by prepregnancy BMI, links between BMI change and maternal and infant outcomes were generally similar; however, associations with nontransfusion SMM and the infant outcomes appeared slightly stronger for the group who began pregnancy with a lower BMI (<25 kg/m²). Overall, these findings suggest that, in this population, substantial interpregnancy BMI gain and loss are modest risk factors for serious maternal complications and stillbirth, and affect infant size.

Results in the context of what’s known

We hypothesized that BMI gain between pregnancies would increase SMM risk on the basis of previous evidence linking such BMI gain with increased risk of related maternal complications like preeclampsia and cesarean delivery^13–16 and suggesting that prepregnancy obesity and excessive gestational weight gain increase the risk of SMM.^{8–11,20,24,25} In the current adjusted analysis, interpregnancy BMI gain was associated with a small increased risk of SMM only for nontransfusion SMM. Our previous study reported the association with obesity was stronger for nontransfusion SMM.⁹ The magnitude of association between prepregnancy obesity and SMM in previous studies was also small relative to other risk factors like pregnancy comorbidities and cesarean delivery.²⁶

Our finding of a modest increased risk for SMM with substantial BMI loss between pregnancies, regardless of BMI at the index pregnancy, is inconsistent with previous evidence that interpregnancy weight loss is protective against maternal complications, including gestational diabetes mellitus,¹⁶ hypertensive disorders,¹⁵ and cesarean delivery.^27,28 A relationship between BMI loss between pregnancies and SMM is not entirely unexpected given that previous evidence links both underweight BMI^8,9 and inadequate gestational weight gain^20,23 with SMM.

Our findings on interpregnancy BMI change and infant size are consistent with previous evidence that BMI gains are inversely associated with SGA and positively associated with LGA, with effect magnitudes that are similar, if smaller, than pooled estimates from 3 meta-analyses.^13,14,16 The U-shaped association of substantial interpregnancy loss and gain with increased risk of stillbirth is similar to that reported for primiparous people in Pennsylvania with prepregnancy BMI of 18.5 to <25 kg/m².²⁹ In that study, developing either underweight or obesity was associated with doubled stillbirth risk,²⁹ whereas in the current analysis, findings were more modest, even for the more comparable people with prepregnancy BMI <25 kg/m². Our 2 studies framed the question differently, but reached a similar conclusion. However, 3 other studies reported increased risk of stillbirth only with BMI gain, not BMI loss,^19,23,30 suggesting that studies in other populations are needed. We note that all previous studies included only people who were primiparous at the first birth, whereas our study included any 2 consecutive births and adjusted for parity.

Clinical implications

We note that incident obesity owing to interpregnancy weight gain may reflect a shorter duration of exposure to mechanisms like inflammation or metabolic dysfunction than that experienced by individuals with established, long-term obesity.³¹ The mechanisms for associations between low BMI and SMM are unclear. In a previous mediation analysis, we reported that the association between prepregnancy obesity and SMM was explained by accounting for comorbidities and cesarean delivery, but these mediators did not affect the association between low BMI and SMM, suggesting different pathways.⁹ Because SMM is relatively rare, our large sample size was insufficient to assess relationships for an interpregnancy gain category greater than 4 units or to specifically assess associations in high-grade obesity subgroups.

Research implications

Replication in other populations, ideally with longitudinal observations of interpregnancy weight and body composition changes, is needed. There was no available information to assess whether maternal weight loss between pregnancies was intentional, and if so, whether it was achieved through healthy or unhealthy practices. Our models could not account for bariatric surgery between pregnancies (although trimming extreme values removed extreme weight losses) or causes of involuntary weight loss, such as serious physical or psychiatric illness, eating disorders, substance abuse, psychosocial stress, breastfeeding, or food insecurity.³² Future studies are required to confirm the association between interpregnancy BMI loss and SMM, ideally in datasets that include explanatory variables that can better characterize the reasons and methods by which individuals lose weight between 2 pregnancies, overall and by prepregnancy BMI category at the index pregnancy. Because prepregnancy underweight and weight loss between pregnancies are much less common than obesity and BMI gain, underweight individuals are often either excluded or combined into the reference group of individuals with normal BMI. Future studies with adequate sample sizes are needed to specifically focus on individuals with low BMI and those who lose weight between pregnancies.

Strengths and limitations

Our study had strengths and limitations. We used a well-accepted composite index to define SMM that was created by the CDC and validated in this population,¹⁸ but SMM conditions are heterogeneous. Large future studies are needed to investigate whether the contribution of interpregnancy weight change varies for specific but relatively infrequent subtypes of SMM (eg, sepsis or eclampsia). Vital records rely primarily on maternal self-reports of height and weight, which are subject to error, and although it is reassuring that a systematic review concluded that these errors did not substantially bias estimates of relationships between maternal body size and birth outcomes,³³ we are not aware of any validation studies with SMM as the outcome. However, our results for fetal outcomes are consistent with established findings, serving as a proof of concept that the analyzed data are robust. We did not adjust for gestational weight gain as a confounding variable because it may be on the causal pathway.¹⁶ Future mediation analyses of longitudinal measures that could differentiate the timing of BMI change (postpartum weight retention vs a gain or loss after delivery) are needed. We also recognize that other definitions of interpregnancy BMI change, such as a change in BMI category, would be informative. Our study’s large, diverse population-based sample was a strength, in that California represents 1 in 8 US births. There was possible selection bias, given the excluded 20% of the sample with missing data differed from those analyzed, but because most of the missing data were for the primary exposure and outcome variables, we did not conduct imputation. Finally, unmeasured confounding (eg, by diet, exercise, social determinants of health) is a limitation owing to the observational design, which also limits causal inference.

Conclusions

In this population-based birth cohort, about a quarter of individuals either lost >2 BMI units or gained ≥4 BMI units between an index and subsequent pregnancy. Similarly, to previous studies, interpregnancy weight change was associated with increased risk of several infant outcomes. Furthermore, substantial BMI loss (even in individuals with an index prepregnancy BMI ≥25 kg/m²) or gain between pregnancies were associated with modestly increased risk of SMM, raising questions about possible mechanisms explaining these findings. Our results add to growing evidence that weight before, during, and between pregnancies may play a role in serious maternal health complications related to childbearing.

AJOG MFM at a Glance.

Why was this study conducted?

Interpregnancy weight change increases risks for adverse pregnancy outcomes, but its role in severe maternal morbidity (SMM) has not been reported. We assessed the associations between changes in body mass index (BMI) and SMM and infant outcomes in >1 million pairs of consecutive California births.

Key findings

Substantial BMI gain (≥4 kg/m²) was associated with a 9% increased risk of nontransfusion SMM. Substantial BMI loss (<2 kg/m²) was associated with 12% and 10% increased risks of SMM and nontransfusion SMM, respectively. The associations were U-shaped for stillbirth and consistent with previous evidence for fetal size. The findings appeared stronger for BMI <25 at the index birth.

What does this add to what is known?

Interpregnancy weight change may play a modest role in SMM risk and fetal outcomes.