ABSTRACT
Background
High and low prepregnancy BMI are risk factors for severe maternal morbidity (SMM), but the contribution of gestational weight gain (GWG) is not well understood.
Objectives
We evaluated associations between GWG and SMM by prepregnancy BMI group.
Methods
We analyzed administrative records from 2,483,684 Californian births (2007–2012), utilizing z score charts to standardize GWG for gestational duration. We fit the z scores nonlinearly and categorized GWG as above, within, or below the Institute of Medicine (IOM) recommendations after predicting equivalent GWG at term from the z score charts. SMM was defined using a validated index. Associations were estimated using multivariable logistic regression models.
Results
We found generally shallow U-shaped relations between GWG z score and SMM in all BMI groups, except class 3 obesity (≥40 kg/m2), for which risk was lowest with weight loss. The weight gain amount associated with the lowest risk of SMM was within the IOM recommendations for underweight and class 2 obesity, but above the IOM recommendations for normal weight, overweight, and class 1 obesity. The adjusted risk ratios (RRs) and 95% CIs for GWG below the IOM recommendations, compared with GWG within the recommendations, were the following for underweight, normal weight, overweight, class 1 obesity, class 2 obesity, and class 3 obesity: 1.13 (0.99, 1.29), 1.09 (1.04, 1.14), 1.10 (1.01, 1.19), 1.07 (0.95, 1.21), 1.03 (0.88, 1.22), and 0.89 (0.73, 1.08), respectively. For GWG above the recommendations, the corresponding RRs and 95% CIs were 0.99 (0.84, 1.15), 1.04 (0.99, 1.08), 0.98 (0.92, 1.04), 1.03 (0.95, 1.13), 1.07 (0.94, 1.23), and 1.08 (0.91, 1.30), respectively.
Conclusions
High and low GWG may be modestly associated with increased risk of SMM across BMI groups, except in women with class 3 obesity, for whom low weight gain and weight loss may be associated with decreased risk of SMM.
Keywords: maternal health, pregnancy, weight gain, obesity, weight loss, population health, pregnancy complications, maternal mortality
Introduction
Serious complications of pregnancy and childbirth affect an increasing number of women in the United States (1, 2). Approximately 700 women die due to pregnancy complications each year (2 per 10,000 live births) (1), and >50,000 women experience severe maternal morbidity (SMM)—life-threatening conditions such as hemorrhage, embolism, and stroke (144 per 10,000 live births) (2). A better understanding of contributors to these tragic, multifactorial events is urgently needed. Recently, high and low prepregnancy BMI have been identified as moderate risk factors for SMM (3, 4). High prepregnancy BMI is prevalent, and may increase the risk of SMM by increasing the risk of comorbidities and cesarean delivery (4). Low prepregnancy BMI is less common, but also confers an increased risk of SMM for reasons not well understood (4).
Weight gain during pregnancy may further affect the risk of SMM, and may be more modifiable than prepregnancy weight (5). Evaluating this association, however, is particularly challenging. First, data sets rarely contain both an adequate sample size of women with SMM and data on maternal weight gain. Second, analyses must untangle weight gain from gestational age, because women gain more weight as the duration of pregnancy increases, and SMM—like many perinatal outcomes—is closely tied to gestational duration (6, 7). Third, weight gain can both affect and be affected by gestational hypertensive disorders and heart failure, which contribute to SMM. Finally, the relation between weight gain and SMM risk is likely nonlinear and modified by prepregnancy BMI. Such barriers have prevented the study of weight gain and SMM using traditional methods in available data sets. Overcoming these barriers and understanding the relation between weight gain and serious maternal health complications is needed to inform efforts to reduce the risk of SMM and optimize weight gain recommendations for maternal and child health.
The objectives of this study were 1) to characterize the relation between gestational weight gain (GWG)—standardized for gestational age—and SMM by prepregnancy BMI group and 2) to estimate the association between the Institute of Medicine (IOM) recommended ranges for GWG and SMM.
Methods
This population-based cohort study was drawn from 3,124,784 recorded live births in California during 2007–2012. The California Office of Statewide Health Planning and Development previously linked birth certificate data and patient discharge data from antepartum, delivery, and postpartum hospitalizations in the state for these years. California adopted the revised US birth certificate in 2007, which collects maternal weight and height data. Eligibility criteria for our analytical sample included singleton live births, linkage of the birth certificate and maternal hospitalization record, and absence of gestational hypertensive disorders or heart failure because these complications would be expected to affect total weight gain (n = 2,807,194) (5, 8). [Women with gestational hypertensive disorders or heart failure (n = 181,682) were included in a sensitivity analysis.] Implausible gestational age, weight, height, and BMI values were identified following the CDC recommendations and excluded from analyses (9). We also excluded GWGs below and above the 1st and 99th percentiles, respectively, after converting absolute weight gain to z scores. Supplemental Figure 1 shows sample selection details. The State of California Committee for the Protection of Human Subjects and the Stanford University Research Compliance Office provided ethics approval.
BMI and GWG
Maternal height, prepregnancy weight, and delivery weight were reported on the birth certificate. These anthropometric variables were either self-reported by the mother or measured during prenatal care or delivery admission. The birth certificate does not indicate the source of the information. We categorized prepregnancy BMI (in kg/m2) as underweight (<18.5), normal weight (18.5–24.9), overweight (25–29.9), obesity class 1 (30–34.9), obesity class 2 (35–39.9), and obesity class 3 (≥40). Weight gain during pregnancy was the difference between delivery weight and prepregnancy weight. We converted absolute weight gain to z scores using established charts specific to the 6 prepregnancy BMI groups and gestational age at delivery (10, 11). The z score was calculated as:
 |
(1) |
where x is absolute weight gain (kg), c is a constant to ensure nonnegative values, and µ and σ are the log-transformed mean and SD of weight gain at a given week of gestation in the same prepregnancy BMI group, respectively. These weight gain z scores are analogous to fetal growth z scores and enable estimation of associations between total weight gain and birth outcomes that are not confounded by duration of pregnancy (6, 12). Shorter duration of pregnancy is strongly associated with SMM and lower GWG (13, 14). The weight gain z scores also analytically allow ongoing pregnancies to be included in the denominator for estimating the risk of SMM, which is not possible with statistical adjustment for gestational age in regression models (6, 15).
We studied GWG as a continuous measurement and categorized it following the IOM recommendations. Because the IOM recommended ranges were created for full-term pregnancies, we used each person's weight gain z score to predict absolute weight gain at 40 weeks of gestation based on the z score charts. We then categorized that amount as low (below), adequate (within), or high (above) following the IOM recommended ranges, which are 12.5–18 kg if underweight, 11.5–16 kg if normal weight, 7–11.5 kg if overweight, and 5–9 kg if obese. [The IOM recommended a single range for all obesity classes because of insufficient evidence to consider separate recommendations for higher obesity classes (5).]
SMM
SMM events occurring from delivery hospitalization to 42 d postpartum were identified using the International Classification of Disease Clinical Modification 9th Revision diagnosis and procedure codes shown in Supplemental Table 1 (16). These codes form the basis of an index developed by the CDC and its partners to identify SMM in administrative data. The most common indicators of SMM reported previously in a national sample were blood transfusion, disseminated intravascular coagulation, and hysterectomy (17). Inclusion of transfusion is important for capturing hemorrhage, which is the most common cause of SMM; however, transfusion alone accounts for half of SMM cases and may capture cases with nonsevere morbidities (13, 16). For these reasons, we also studied as a secondary outcome in this analysis an alternative definition of SMM that excluded cases for which the only indicator was a blood transfusion. This latter outcome generally includes more severe conditions and is hereafter referred to as nontransfusion SMM.
Confounders
Confounders were selected based on directed acyclic graphs, prior evidence, and available data (3, 4, 13, 16, 18–22). They included maternal age at birth (y), height (m), obstetric history (nulliparous, multiparous without prior cesarean delivery, multiparous with prior cesarean delivery), education (less than high school degree, high school degree or equivalent, some college, college degree), race/ethnicity (US-born Hispanic/Latina, foreign-born Hispanic/Latina, non-Hispanic white, Asian/Pacific Islander, non-Hispanic black, other), expected method of payment (private insurance, Medi-Cal, other), and any of the medical comorbidities derived from a validated obstetric comorbidity index (23, 24): pre-existing diabetes or hypertension, asthma, hematologic disorder, chronic renal disease, neuromuscular disorder, lupus or autoimmune disease, HIV/AIDS, congenital or ischemic cardiac disease, pulmonary hypertension, and substance use or alcohol use disorder.
Statistical analysis
All statistical analyses were stratified by prepregnancy BMI group because optimal weight gain varies by BMI (5). We first studied GWG as a continuous variable after converting it to z scores. We fit the z scores with restricted cubic splines to capture nonlinear relations between weight gain and SMM. Akaike and Bayesian information criteria were used to determine the number of knots using Harrell's default placements (25). The continuous covariates, age and height, were also fit with restricted cubic splines using the same method. We then used multivariable logistic regression to model the association between weight gain z score and SMM. We used stabilized inverse probability weights to account for missing covariate data in all models. A logit link was used in regression modeling to ensure the predicted probability of the outcome was within [0,1] bounds, given the rarity of SMM (<2%). ORs approximated risk ratios (RRs) because of the rare outcome. We hereafter use the term RR to aid in interpretability of the results.
We used the regression models to predict the marginal probability of SMM across weight gain z scores in each BMI group. The probabilities were multiplied by 10,000 for SMM risk per 10,000 births. For each model, we identified the nadir of risk and the current weight gain ranges recommended by the IOM.
We also studied GWG categorized as low, adequate, and high following the IOM recommended ranges for full-term pregnancies. Before categorization, we predicted absolute weight gain at 40 weeks of gestation using the z score charts. We used multivariable logistic regression models to estimate RRs (approximated from ORs), risk differences, and corresponding 95% CIs for high and low weight gain relative to the recommended ranges. All analyses were replicated for the secondary outcome of nontransfusion SMM. We assessed the robustness of our results to analytical decisions by replicating analyses after 1) including women with gestational hypertensive disorders or heart failure, and 2) excluding women with gestational diabetes mellitus. Bidirectional relations between these complications and GWG likely exist, which cannot be disentangled when serial weight measurements are not available. Stata 15 (StataCorp LLC) was used to conduct the statistical analysis.
Results
Among the final study sample of 2,483,684 women, 21% gained weight below the IOM recommendations, 31% gained weight within the recommendations, and 48% gained weight above the recommendations. SMM occurred in 1.06% of births and nontransfusion SMM occurred in 0.43% of births. The incidence of SMM was highest in women with low weight gain (1.14%) (Table 1). Prepregnancy underweight and obesity class 2 or 3 were most common in women with low weight gain, whereas overweight and obesity class 1 were most common in women with high weight gain.
TABLE 1.
Study population characteristics by gestational weight gain status, California, 2007–20121
| Characteristic | All individuals (n = 2,483,684), % | Low weight gain (n = 511,962), % | Adequate weight gain (n = 769,717), % | High weight gain (n = 1,202,005), % |
|---|---|---|---|---|
| SMM | 1.06 | 1.14 | 1.02 | 1.06 |
| Nontransfusion SMM | 0.43 | 0.46 | 0.41 | 0.43 |
| Prepregnancy BMI | ||||
| Underweight | 4.2 | 6.5 | 5.9 | 2.1 |
| Normal weight | 50.5 | 59.3 | 58.7 | 41.3 |
| Overweight | 25.7 | 15.9 | 21.4 | 32.7 |
| Obese class 1 | 12.1 | 8.7 | 8.4 | 16.0 |
| Obese class 2 | 4.8 | 5.3 | 3.7 | 5.4 |
| Obese class 3 | 2.7 | 4.3 | 2.0 | 2.5 |
| Height, cm | ||||
| <157 | 22.1 | 27.1 | 23.9 | 18.9 |
| ≥157 | 77.9 | 72.9 | 76.1 | 81.1 |
| Medical comorbidity score | ||||
| 0 | 93.2 | 93.7 | 94.1 | 92.3 |
| 1–2 | 6.2 | 5.7 | 5.3 | 7.0 |
| ≥3 | 0.6 | 0.6 | 0.6 | 0.7 |
| Age, y | ||||
| <25 | 29.8 | 30.9 | 27.6 | 30.7 |
| 25–34 | 52.5 | 50.4 | 53.3 | 52.9 |
| ≥35 | 17.7 | 18.7 | 19.1 | 16.4 |
| Obstetric history | ||||
| Nulliparous | 38.7 | 33.9 | 37.6 | 41.5 |
| Multiparous without prior cesarean birth | 44.2 | 49.6 | 46.1 | 40.7 |
| Multiparous with prior cesarean birth | 17.1 | 16.5 | 16.3 | 17.8 |
| Race/ethnicity | ||||
| US-born Hispanic/Latina | 24.2 | 22.7 | 21.7 | 26.4 |
| Foreign-born Hispanic/Latina | 28.2 | 34.8 | 29.6 | 24.5 |
| Non-Hispanic white | 26.1 | 19.9 | 25.6 | 29.0 |
| Asian/Pacific Islander | 12.4 | 14.4 | 15.1 | 9.9 |
| Non-Hispanic black | 4.8 | 4.6 | 3.9 | 5.5 |
| Other | 4.3 | 3.5 | 4.1 | 4.7 |
| Educational attainment | ||||
| Less than high school completion | 24.3 | 30.1 | 23.9 | 22.0 |
| High school degree or equivalent | 26.1 | 26.1 | 24.2 | 27.3 |
| Some college | 23.9 | 21.2 | 22.4 | 25.9 |
| College degree or higher | 25.8 | 22.6 | 29.5 | 24.7 |
| Expected delivery payment method | ||||
| Private insurance | 47.7 | 42.2 | 49.9 | 48.6 |
| Medicaid | 47.1 | 52.3 | 45.0 | 46.3 |
| Other | 5.2 | 5.5 | 5.0 | 5.1 |
| Preterm birth (<37 wk) | 5.9 | 6.8 | 5.5 | 5.8 |
Categorized following Institute of Medicine recommendations after using z scores to predict total weight gain at 40 weeks of gestation. SMM, severe maternal morbidity.
Analysis of GWG z score as a continuous variable revealed nonlinear associations with SMM (Figure 1) that varied by prepregnancy BMI group. The lowest predicted risks of SMM were observed at the following full-term equivalent weight gain amounts: 16.2 kg in women with underweight, 15.3 kg in women with normal weight, 13.5 kg in women with overweight, 11.4 kg in women with obesity class 1, and 7.6 kg in women with obesity class 2. Risk continually decreased with decreasing weight change across the observed range—including weight loss—in women with obesity class 3. The weight gain amount associated with the lowest risk of SMM was within the range of IOM recommendations among women with underweight BMI or class 2 obesity, but above the IOM range among women with normal weight, overweight, or obesity class 1. Analysis of nontransfusion SMM showed minimal changes in risk across weight gain amounts, except for decreasing risk with decreasing weight change in women with obesity class 3 (Figure 2).
FIGURE 1.
Adjusted predicted risk of severe maternal morbidity across GWG z scores in women with prepregnancy underweight (A; n = 104,324), normal weight (B; n = 1,253,344), overweight (C; n = 639,093), obesity class 1 (D; n = 301,258), obesity class 2 (E; n = 119,571), and obesity class 3 (F; n = 66,094), California, 2007–2012. Dashed lines represent 95% CIs. Vertical lines indicate the current Institute of Medicine guidelines for GWG. GWG, gestational weight gain.
FIGURE 2.
Adjusted predicted risk of nontransfusion severe maternal morbidity across GWG z scores in women with prepregnancy underweight (A; n = 104,324), normal weight (B; n = 1,253,344), overweight (C; n = 639,093), obesity class 1 (D; n = 301,258), obesity class 2 (E; n = 119,571), and obesity class 3 (F; n = 66,094), California, 2007–2012. Vertical lines indicate the current Institute of Medicine guidelines for GWG. GWG, gestational weight gain.
GWG was further assessed as low, adequate, or high based on the IOM recommended ranges after predicting absolute weight gain at 40 weeks of gestation using the z score charts (Tables 2 and 3). Women with prepregnancy obesity class 3 and high weight gain experienced the highest incidence of SMM (147 per 10,000 births) and nontransfusion SMM (71 per 10,000 births). Women with prepregnancy underweight and low weight gain experienced the second highest incidence of SMM (128 per 10,000 births), but the incidence of nontransfusion SMM was higher among women with obesity than underweight. In crude regression models, low weight gain was significantly associated with SMM in women with underweight, normal weight, or overweight; the CIs excluded 1.0 (Table 2). The risk of SMM was elevated with high weight gain compared with adequate weight gain in women with obesity class 2 or class 3, but the CIs included 1.0. After adjustment for confounders, the measures of association for low weight gain in women with underweight, normal weight, or overweight and for high weight gain in women with obesity classes 2 and 3 were attenuated. For the secondary outcome of nontransfusion SMM, relative measures of association were marginally larger than for SMM overall (Table 3). All CIs included 1.0, except for low weight gain in women with normal weight or overweight.
TABLE 2.
Associations of low and high gestational weight gain, relative to IOM recommended ranges, with severe maternal morbidity by prepregnancy BMI category, California, 2007–20121
| Gestational weight gain | Number of cases | Incidence per 10,000 births | Crude RR (95% CI) | Adjusted RR (95% CI) | Crude RD (95% CI) per 10,000 births | Adjusted RD (95% CI) per 10,000 births |
|---|---|---|---|---|---|---|
| Underweight | ||||||
| Low | 426 | 128 | 1.20 (1.05, 1.36) | 1.13 (0.99, 1.29) | 20.8 (5.4, 36.2) | 14.2 (−1.6, 29.9) |
| Adequate | 492 | 108 | Reference | Reference | Reference | Reference |
| High | 281 | 110 | 1.03 (0.88, 1.19) | 0.99 (0.84, 1.15) | 2.7 (−13.2, 18.6) | −1.6 (−18.4, 15.1) |
| Normal weight | ||||||
| Low | 3407 | 112 | 1.14 (1.09, 1.19) | 1.09 (1.04, 1.14) | 13.2 (8.5, 17.9) | 8.8 (4.1, 13.6) |
| Adequate | 4479 | 99 | Reference | Reference | Reference | Reference |
| High | 5050 | 102 | 1.03 (0.99, 1.07) | 1.04 (0.99, 1.08) | 3.0 (−1.0, 7.0) | 3.6 (−0.6, 7.8) |
| Overweight | ||||||
| Low | 960 | 118 | 1.11 (1.02, 1.20) | 1.10 (1.01, 1.19) | 11.4 (2.6, 20.4) | 10.4 (1.4, 19.3) |
| Adequate | 1745 | 106 | Reference | Reference | Reference | Reference |
| High | 4031 | 103 | 0.96 (0.91, 1.02) | 0.98 (0.92, 1.04) | −3.7 (−9.6, 2.2) | −2.3 (−8.3, 3.7) |
| Obesity class 1 | ||||||
| Low | 502 | 113 | 1.08 (0.96, 1.21) | 1.07 (0.95, 1.21) | 8.0 (−4.6, 20.7) | 7.7 (−5.2, 20.5) |
| Adequate | 677 | 105 | Reference | Reference | Reference | Reference |
| High | 2116 | 110 | 1.04 (0.95, 1.14) | 1.03 (0.95, 1.13) | 4.3 (−4.2, 13.5) | -3.7 (−5.8, 1.3) |
| Obesity class 2 | ||||||
| Low | 300 | 111 | 1.01 (0.86, 1.19) | 1.03 (0.88, 1.22) | 0.13 (−16.1, 18.7) | 3.8 (−14.1, 21.7) |
| Adequate | 306 | 109 | Reference | Reference | Reference | Reference |
| High | 775 | 120 | 1.10 (0.96, 1.26) | 1.07 (0.94, 1.23) | 10.9 (−3.9, 25.0) | 8.1 (−7.1, 23.2) |
| Obesity class 3 | ||||||
| Low | 241 | 112 | 0.89 (0.74, 1.08) | 0.89 (0.73, 1.08) | −13.3 (−35.8, 9.2) | −14.2 (−37.7, 9.0) |
| Adequate | 186 | 125 | Reference | Reference | Reference | Reference |
| High | 436 | 147 | 1.17 (0.99, 1.39) | 1.08 (0.91, 1.30) | 21.4 (−0.8, 43.7) | 10.9 (−11.9, 33.8) |
n = 2,483,684. Logistic regression used to model associations. Multivariable models adjusted for maternal age, height, obstetric history, educational attainment, race/ethnicity, payment method, and prepregnancy medical comorbidity. Weight gain categorized following IOM recommended ranges after using z scores to predict total weight gain at 40 weeks of gestation. Women with gestational hypertensive conditions and heart failure were excluded. IOM, Institute of Medicine; RD, risk difference; RR, risk ratio.
TABLE 3.
Associations of low and high gestational weight gain, relative to IOM recommended ranges, with nontransfusion SMM by prepregnancy BMI category, California, 2007–20121
| Number of cases | Incidence per 10,000 births | Crude RR (95% CI) | Adjusted RR (95% CI) | Crude RD (95% CI) per 10,000 births | Adjusted RD (95% CI) per 10,000 births | |
|---|---|---|---|---|---|---|
| Underweight | ||||||
| Low | 146 | 44 | 1.24 (0.99, 1.56) | 1.22 (0.97, 1.54) | 8.6 (−0.4, 17.6) | 8.0 (−1.3, 17.2) |
| Adequate | 162 | 35 | Reference | Reference | Reference | Reference |
| High | 91 | 36 | 1.01 (0.78, 1.30) | 0.99 (0.76, 1.29) | 0.3 (−8.9, 9.4) | −0.3 (−1.0, 0.9) |
| Normal weight | ||||||
| Low | 1353 | 44 | 1.15 (1.07, 1.23) | 1.13 (1.05, 1.22) | 5.7 (2.7, 8.7) | 5.0 (2.0, 8.0) |
| Adequate | 1755 | 39 | Reference | Reference | Reference | Reference |
| High | 1899 | 38 | 0.99 (0.93, 1.05) | 0.98 (0.91, 1.05) | −0.1 (−3.0, 2.0) | −0.9 (−3.4, 1.7) |
| Overweight | ||||||
| Low | 410 | 50 | 1.16 (1.03, 1.31) | 1.16 (1.03, 1.32) | 6.9 (1.1, 12.7) | 6.9 (1.1, 12.8) |
| Adequate | 713 | 43 | Reference | Reference | Reference | Reference |
| High | 1683 | 43 | 0.99 (0.90, 1.08) | 1.00 (0.91, 1.10) | −0.6 (−4.4, 3.2) | 0.1 (−3.8, 3.9) |
| Obesity class 1 | ||||||
| Low | 198 | 45 | 1.00 (0.84, 1.20) | 1.00 (0.83, 1.20) | 0.2 (−7.9, 8.3) | 0.6 (−8.1, 8.3) |
| Adequate | 286 | 45 | Reference | Reference | Reference | Reference |
| High | 905 | 47 | 1.05 (0.92, 1.20) | 1.06 (0.92, 1.21) | 2.4 (−3.6, 8.4) | 2.5 (−3.7, 8.6) |
| Obesity class 2 | ||||||
| Low | 141 | 52 | 1.16 (0.92, 1.48) | 1.18 (0.93, 1.51) | 7.3 (−4.3, 18.9) | 8.3 (−3.8, 20.4) |
| Adequate | 125 | 45 | Reference | Reference | Reference | Reference |
| High | 325 | 50 | 1.13 (0.92, 1.39) | 1.06 (0.86, 1.31) | 5.7 (−3.8, 15.3) | 2.8 (−7.0, 12.6) |
| Obesity class 3 | ||||||
| Low | 106 | 49 | 0.85 (0.64, 1.13) | 0.83 (0.62, 1.11) | −8.6 (−23.8, 6.5) | −10.1 (−26.0, 5.9) |
| Adequate | 86 | 58 | Reference | Reference | Reference | Reference |
| High | 211 | 71 | 1.24 (0.97, 1.60) | 1.11 (0.86, 1.44) | 13.9 (−1.4, 29.2) | 6.8 (−9.0, 22.5) |
n = 2,483,684. Nontransfusion SMM defined as SMM for which blood transfusion was not the only indication. Logistic regression used to model associations. Multivariable models adjusted for maternal age, height, obstetric history, educational attainment, race/ethnicity, payment method, and prepregnancy medical comorbidity. Weight gain categorized following IOM recommended ranges after using z scores to predict total weight gain at 40 weeks of gestation. Women with gestational hypertensive conditions and heart failure were excluded. IOM, Institute of Medicine; RD, risk difference; RR, risk ratio; SMM, severe maternal morbidity.
In sensitivity analyses, we found notable increases in the risk of SMM with high weight gain in all BMI groups after including women with gestational hypertensive conditions or heart failure (Supplemental Figure 2, Supplemental Table 2). For example, in women with class 3 obesity and weight gain above the IOM recommendations, the adjusted RR for SMM was 1.08 (95% CI: 0.91, 1.30), nonsignificant, when women with gestational hypertension conditions or heart failure were excluded (Table 2); the corresponding adjusted RR was 1.17 (95% CI: 1.01, 1.34), therefore significant, when women with those conditions were included in analyses (Supplemental Table 2). In sensitivity analyses that excluded women with gestational diabetes mellitus from the study sample, results were unchanged from the main findings (results not shown).
Discussion
We found generally shallow U-shaped relations between GWG z score and SMM in women with BMI < 35. In women with class 3 obesity (BMI ≥ 35), the risk of SMM decreased with decreasing weight change—including weight loss. When GWG was expressed relative to IOM recommended ranges, SMM was most common in 2 subgroups: women with class 3 obesity and excessive GWG and women with underweight and inadequate weight gain. The CIs for these associations, however, included the null and the IOM ranges did not include the weight gain amount corresponding to the lowest risk of SMM, except in women with underweight or class 2 obesity. The pattern of results was overall similar for the secondary outcome of nontransfusion SMM, although effect sizes were attenuated. The relation between high weight gain and SMM was much stronger in sensitivity analyses that included women with gestational hypertensive conditions or heart failure, which likely have a bidirectional relation with GWG.
The current study extends previous work on the association between prepregnancy BMI and SMM to consider in addition weight gain during pregnancy. In combination with previous evidence (3, 4, 26), our study suggests a stronger role of prepregnancy BMI than GWG in affecting women's health. We previously reported that the risk of SMM was significantly elevated in women with underweight or obesity in the California birth cohort (adjusted RR range: 1.04–1.28; adjusted risk difference range: 5–39 per 10,000 births) (4). We further found that mediation by comorbidities and cesarean delivery explained the observed association between high prepregnancy BMI and SMM. Similar pathways may explain an increased risk of SMM associated with high GWG in women with high prepregnancy BMI. High total weight gain has a well-established association with cesarean delivery, and researchers have also recently reported associations between high gain before the diagnosis of pre-eclampsia and gestational diabetes mellitus, 2 conditions that can contribute to SMM (5, 16, 27–29). The association of low BMI or low weight gain with SMM is less understood. Anemia has been proposed as 1 contributor (3), but we did not find anemia to have an effect in a previous study on BMI and SMM (4). Examination of potential pathways is limited by the scarcity of large data sets with multiple weight gain measurements during pregnancy, which are needed to ensure temporality (6).
For women without the most severe level of obesity, we observed generally shallow U-shaped associations between weight gain and SMM, which have also been found in studies of GWG and preterm birth and infant mortality (12, 30, 31). In women with class 3 obesity, however, risk of SMM was lowest with weight loss. This finding adds to a growing body of evidence suggesting that weight change below the IOM guidelines, including weight loss, in women with severe obesity is associated with better health outcomes for the woman and her infant, with the notable exception of a higher risk of infants being small-for-gestational age (27, 28, 32). The current IOM recommendations for weight gain during pregnancy, published in 2009, did not have sufficient evidence to consider separate recommendations for women with class 2 or 3 obesity (5). The IOM recommended that all women with BMI ≥ 30 gain 5–9 kg during a full-term pregnancy (5). In conjunction with other recent studies (27, 28, 33), our findings are informative for future revisions of GWG guidelines that optimize outcomes for mothers and their infants and consider separate guidelines for women with very high BMI. Prepregnancy class 3 obesity has been increasing more rapidly than less severe obesity in the United States, and affects ∼5% of pregnant women (34).
Our study addressed sources of potential bias that have previously limited the study of GWG and SMM. For one, gestational duration is closely tied to both GWG and SMM. Women generally gain weight as pregnancy continues, and the IOM recommended weight gain ranges were developed for full-term pregnancies (5). In addition, a substantial proportion of women with SMM give birth preterm; a study in California reported preterm birth rates of 41% in women with confirmed SMM and 8% in other women (7). Our study used established z score charts to account for the dependency of weight gain on gestational duration (10, 11). This approach enabled us to include preterm births and predict each woman's absolute weight gain at 40 weeks of gestation (term) to apply the IOM guidelines consistently across all women. A second potential source of bias is a likely bidirectional relation between GWG at delivery and gestational hypertensive conditions and heart failure because these conditions cause weight gain from water retention (5, 8). Two recent studies with data on weight gain measured early in pregnancy report an association with gestational hypertensive conditions (29, 35), but studies with data only on total weight gain at delivery may lead to spurious conclusions. We found that the association between high weight gain at delivery and SMM was notably larger when we included women with these conditions in our analyses, but without weight gain measured before disease onset, we cannot disentangle the temporality of these conditions and weight gain.
These sources of potential major bias likely contributed to differences between our results and those recently reported in a study conducted among full-term births in a single metropolitan area (36); the study reported that weight gain above the IOM guidelines, but not below, was associated with increased risk of SMM in women with underweight, normal weight, or class 3 obesity (37). We encourage observational studies of GWG to consider available approaches to reduce bias as well as include preterm births in their analyses (6, 36).
This observational study should be interpreted in light of its limitations. Maternal weight and height measurements recorded in vital records are often self-reported, which could cause exposure misclassification. Individuals tend to under-report prepregnancy weight by 0.3–3 kg and delivery weight by 0–1.3 kg (38), although underweight women tend to over-report their weight (39). In a systematic review, these magnitudes of error were not found to bias associations with birth outcomes (38). Valid use of the weight gain-for-gestational age z scores requires the pattern of GWG in the study population to be similar to that in the z score standard population. We believe this assumption was reasonable based on similar weight gain patterns in other study populations and a previous assessment of the generalizability of the z scores charts to California (5, 30, 40). In addition, very rare maternal complications are substantially underreported in patient discharge data (41). We used both patient discharge and vital records data to increase sensitivity (41). The index that we used to identify SMM cases has been found to have a sensitivity of 0.77 and a specificity of 0.99 compared with medical records in California, with a sensitivity of 0.53 for nontransfusion cases (42). The index is a composite intended to capture events that could have resulted in death, but studies are also needed to study specific complications. Our main analyses excluded women with gestational hypertensive conditions or heart failure because these conditions can affect total weight gain through water retention, but these conditions are important contributors to SMM. Data sets with serial weight gain measurements are required to allow studies to overcome this problem. Although California is a diverse state representing 1 in 8 births in the United States, generalizability of our results to other settings is unstudied.
In conclusion, modest increases in risk of SMM were observed for low and high weight gain during pregnancy. In women with class 3 obesity, the risk of SMM decreased with weight loss. Current IOM recommended weight gain ranges did not include the weight gain amounts at which the risk of SMM was lowest, except in women with underweight or class 2 obesity. Our findings demonstrate the need for additional research on weight gain and health outcomes in pregnant individuals, and are informative for efforts to optimize weight gain recommendations for all BMI groups and reduce the risk of serious maternal health complications.
Supplementary Material
ACKNOWLEDGEMENTS
The authors’ responsibilities were as follows—SAL, BA, and SLC: designed the study, SAL: performed the data analysis, wrote the manuscript, and had primary responsibility for the final content; BA, SLC, EKM, and DJL: provided study oversight; and all authors: read and approved the final manuscript. The authors report no conflicts of interest.
Notes
Supported by National Institute of Nursing Research and Office of Research on Women's Health grant R01 NR017020 (to SLC); Eunice Kennedy Shriver National Institute of Child Health and Human Development grant F32 HD091945 (to SAL); and the Stanford Maternal and Child Health Research Institute (to SAL).
Supplemental Tables 1 and 2 and Supplemental Figures 1 and 2 are available from the “Supplementary data” link in the online posting of the article and from the same link in the online table of contents at https://academic.oup.com/ajcn/.
Data and the codebook described in the article are available upon request to the California Office of Statewide Health Planning and Development. Analytical code will be made available from the corresponding author upon request.
Abbreviations used: GWG, gestational weight gain; IOM, Institute of Medicine; RR, risk ratio; SMM, severe maternal morbidity.
References
- 1. CDC. Pregnancy-Related Deaths. [Internet] Atlanta, GA: CDC; 2019; [cited 12 September, 2019]. Available from: https://www.cdc.gov/reproductivehealth/maternalinfanthealth/pregnancy-relatedmortality.htm. [Google Scholar]
- 2. CDC. Severe Maternal Morbidity in the United States. [Internet] Atlanta, GA: CDC; 2017; [cited 12 September, 2019]. Available from: https://www.cdc.gov/reproductivehealth/maternalinfanthealth/severematernalmorbidity.html#anchor_SMM. [Google Scholar]
- 3. Lisonkova S, Muraca GM, Potts J, Liauw J, Chan WS, Skoll A, Lim KI. Association between prepregnancy body mass index and severe maternal morbidity. JAMA. 2017;318(18):1777–86. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Leonard SA, Carmichael SL, Main EK, Lyell DJ, Abrams B. Risk of severe maternal morbidity in relation to prepregnancy body mass index: roles of maternal co-morbidities and caesarean birth. Paediatr Perinat Epidemiol. 2019; [Epub ahead of print]. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Institute of Medicine. Weight Gain During Pregnancy: Reexamining the Guidelines. Rasmussen KM, Yaktine AL,editors. Washington (DC): The National Academies Press; 2009. [PubMed] [Google Scholar]
- 6. Hutcheon JA, Bodnar LM. Good practices for observational studies of maternal weight and weight gain in pregnancy. Paediatr Perinat Epidemiol. 2018;32(2):152–60. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Kilpatrick SJ, Abreo A, Gould J, Greene N, Main EK. Confirmed severe maternal morbidity is associated with high rate of preterm delivery. Am J Obstet Gynecol. 2016;215(2):233.e1–7. [DOI] [PubMed] [Google Scholar]
- 8. Remme WJ, Swedberg K; Task Force for the Diagnosis and Treatment of Chronic Heart Failure, European Society of Cardiology . Guidelines for the diagnosis and treatment of chronic heart failure. Eur Heart J. 2001;22(23):1527–60. [DOI] [PubMed] [Google Scholar]
- 9. CDC. User Guide to the 2016 Natality Public Use File. [Internet] Hyattsville, MD: US Department of Health and Human Services; 2016; [cited 12 September, 2019]. Available from: ftp://ftp.cdc.gov/pub/Health_Statistics/NCHS/Dataset_Documentation/DVS/natality/UserGuide2016.pdf. [Google Scholar]
- 10. Hutcheon JA, Platt RW, Abrams B, Himes KP, Simhan HN, Bodnar LM. A weight-gain-for-gestational-age z score chart for the assessment of maternal weight gain in pregnancy. Am J Clin Nutr. 2013;97(5):1062–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Hutcheon JA, Platt RW, Abrams B, Himes KP, Simhan HN, Bodnar LM. Pregnancy weight gain charts for obese and overweight women. Obesity (Silver Spring). 2015;23(3):532–5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Bodnar LM, Hutcheon JA, Parisi SM, Pugh SJ, Abrams B. Comparison of gestational weight gain z-scores and traditional weight gain measures in relation to perinatal outcomes. Paediatr Perinat Epidemiol. 2015;29(1):11–21. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Grobman WA, Bailit JL, Rice MM, Wapner RJ, Reddy UM, Varner MW, Thorp JM Jr, Leveno KJ, Caritis SN, Iams JD et al.. Frequency of and factors associated with severe maternal morbidity. Obstet Gynecol. 2014;123(4):804–10. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Hutcheon JA, Bodnar LM, Joseph KS, Abrams B, Simhan HN, Platt RW. The bias in current measures of gestational weight gain. Paediatr Perinat Epidemiol. 2012;26(2):109–16. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Caughey AB, Stotland NE, Escobar GJ. What is the best measure of maternal complications of term pregnancy: ongoing pregnancies or pregnancies delivered?. Am J Obstet Gynecol. 2003;189:1047–52. [DOI] [PubMed] [Google Scholar]
- 16. Callaghan WM, Creanga AA, Kuklina EV. Severe maternal morbidity among delivery and postpartum hospitalizations in the United States. Obstet Gynecol. 2012;120(5):1029–36. [DOI] [PubMed] [Google Scholar]
- 17. Fingar KR, Hambrick MM, Heslin KC, Moore JE. Trends and Disparities in Delivery Hospitalizations Involving Severe Maternal Morbidity, 2006–2015. Rockville, MD: Agency for Healthcare Research and Quality; 2018. [PubMed] [Google Scholar]
- 18. Butwick AJ, Bentley J, Leonard SA, Carmichael SL, El-Sayed YY, Stephansson O, Guo N. Prepregnancy maternal body mass index and venous thromboembolism: a population based cohort study. BJOG. 2019;126(5):581–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Schummers L, Hutcheon JA, Bodnar LM, Lieberman E, Himes KP. Risk of adverse pregnancy outcomes by prepregnancy body mass index: a population-based study to inform prepregnancy weight loss counseling. Obstet Gynecol. 2015;125(1):133–43. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Lyndon A, Lee HC, Gilbert WM, Gould JB, Lee KA. Maternal morbidity during childbirth hospitalization in California. J Matern Fetal Neonatal Med. 2012;25(12):2529–35. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Lisonkova S, Potts J, Muraca GM, Razaz N, Sabr Y, Chan WS, Kramer MS. Maternal age and severe maternal morbidity: a population-based retrospective cohort study. PLoS Med. 2017;14(5):e1002307. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Sebire NJ, Jolly M, Harris JP, Wadsworth J, Joffe M, Beard RW, Regan L, Robinson S. Maternal obesity and pregnancy outcome: a study of 287,213 pregnancies in London. Int J Obes. 2001;25(8):1175–82. [DOI] [PubMed] [Google Scholar]
- 23. Bateman BT, Mhyre JM, Hernandez-Diaz S, Huybrechts KF, Fischer MA, Creanga AA, Callaghan WM, Gagne JJ. Development of a comorbidity index for use in obstetric patients. Obstet Gynecol. 2013;122(5):957–65. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Easter SR, Sweeney V, Managanaro K, Lassey SC, Bateman BT, Robinson JN. 278: prospective clinical validation of the obstetric comorbidity index for maternal risk assessment. Am J Obstet Gynecol. 2019;220(1):S198–9. [Google Scholar]
- 25. Harrell FEJ, Lee KL, Pollock BG. Regression models in clinical studies: determining relationships between predictors and response. J Natl Cancer Inst. 1988;80:1198–202. [DOI] [PubMed] [Google Scholar]
- 26. Siddiqui A, Azria E, Howell EA, Deneux-Tharaux C. Associations between maternal obesity and severe maternal morbidity: findings from the French EPIMOMS population-based study. Paediatr Perinat Epidemiol. 2019;33(1):7–16. [DOI] [PubMed] [Google Scholar]
- 27. Kominiarek MA, Peaceman AM. Gestational weight gain. Am J Obstet Gynecol. 2017;217(6):642–51. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. Goldstein RF, Abell SK, Ranasinha S, Misso M, Boyle JA, Black MH, Li N, Hu G, Corrado F, Rode L et al.. Association of gestational weight gain with maternal and infant outcomes: a systematic review and meta-analysis. JAMA. 2017;317(21):2207–25. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29. Bodnar LM, Himes KP, Abrams B, Parisi SM, Hutcheon JA. Early-pregnancy weight gain and the risk of preeclampsia: a case-cohort study. Pregnancy Hypertension. 2018;14:205–12. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30. Leonard SA, Hutcheon JA, Bodnar LM, Petito LC, Abrams B. Gestational weight gain-for-gestational age z-score charts applied across U.S. populations. Paediatr Perinat Epidemiol. 2018;32(2):161–71. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31. Bodnar LM, Siminerio LL, Himes KP, Hutcheon JA, Lash TL, Parisi SM, Abrams B. Maternal obesity and gestational weight gain are risk factors for infant death. Obesity. 2016;24(2):490–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32. Bodnar LM, Pugh SJ, Lash TL, Hutcheon JA, Himes KP, Parisi SM, Abrams B. Low gestational weight gain and risk of adverse perinatal outcomes in obese and severely obese women. Epidemiology. 2016;27(6):894–902. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33. Bodnar LM, Siega-Riz AM, Simhan HN, Himes KP, Abrams B. Severe obesity, gestational weight gain, and adverse birth outcomes. Am J Clin Nutr. 2010;91(6):1642–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34. Deputy NP, Dub B, Sharma AJ. Prevalence and trends in prepregnancy normal weight — 48 states, New York City, and District of Columbia, 2011–2015. MMWR Morb Mortal Wkly Rep. 2018;66:1402–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35. Hutcheon JA, Stephansson O, Cnattingius S, Bodnar LM, Wikström A-K, Johansson K. Pregnancy weight gain before diagnosis and risk of preeclampsia: a population-based cohort study in nulliparous women. Hypertension. 2018;72(2):433–41. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36. Leonard SA, Carmichael SL, Abrams B. Letter to the Editor: gestational weight gain and severe maternal morbidity at delivery hospitalization. Obstet Gynecol. 2019;134(2):420. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37. Platner MH, Ackerman C, Howland RE, Xu X, Pettker CM, Illuzzi JL, Reddy UM, Chung S, Lipkind HS. Gestational weight gain and severe maternal morbidity. Obstet Gynecol. 2019;133(3):515–24. [DOI] [PubMed] [Google Scholar]
- 38. Headen I, Cohen AK, Mujahid M, Abrams B. The accuracy of self-reported pregnancy-related weight: a systematic review. Obes Rev. 2017;18(3):350–69. [DOI] [PubMed] [Google Scholar]
- 39. Han E, Abrams B, Sridhar S, Xu F, Hedderson M. Validity of self-reported pre-pregnancy weight and body mass index classification in an integrated health care delivery system. Paediatr Perinat Epidemiol. 2016;30(4):314–19. [DOI] [PubMed] [Google Scholar]
- 40. Cheikh Ismail L, Bishop DC, Pang R, Ohuma EO, Kac G, Abrams B, Rasmussen K, Barros FC, Hirst JE, Lambert A et al.. Gestational weight gain standards based on women enrolled in the Fetal Growth Longitudinal Study of the INTERGROWTH-21st project: a prospective longitudinal cohort study. BMJ. 2016;352:i555. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41. Lydon-Rochelle MT, Holt VL, Nelson JC, Cárdenas V, Gardella C, Easterling TR, Callaghan WM. Accuracy of reporting maternal in-hospital diagnoses and intrapartum procedures in Washington State linked birth records. Paediatr Perinat Epidemiol. 2005;19(6):460–71. [DOI] [PubMed] [Google Scholar]
- 42. Main EK, Abreo A, McNulty J, Gilbert W, McNally C, Poeltler D, Lanner-Cusin K, Fenton D, Gipps T, Melsop K et al.. Measuring severe maternal morbidity: validation of potential measures. Am J Obstet Gynecol. 2016;214(5):643.e1–e10. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.