Abstract
Objective
There is no clear evidence on the risk of gestational weight loss (GWL) for individuals with obesity. Our study aimed to assess the association between GWL and adverse perinatal outcomes among individuals with obesity.
Methods
This population‐based retrospective cohort study examined individuals with prepregnancy BMI ≥ 30 kg/m2 who had a singleton pregnancy, using Ontario, Canada, birth registry data from 2012 to 2020. The primary outcome was a composite of adverse outcomes, including perinatal death and neonatal morbidity. The association between GWL and risk of adverse perinatal outcomes was estimated using generalized estimating equation models and restricted cubic spline regression analysis. Stratified analysis was conducted by obesity class.
Results
Of the 157,205 individuals with obesity, 6.1% experienced GWL. Compared with adequate gestational weight gain, GWL was associated with an increased risk of a composite of adverse perinatal outcomes (adjusted risk ratio: 1.31; 95% CI: 1.22–1.39). Similar results were observed in the stratified analysis. Restricted cubic spline regression analysis revealed that average weekly gestational weight changes displayed a nonlinear U‐shaped association, with a higher risk of a composite of adverse perinatal outcomes noted in the extremities, particularly toward GWL and excessive weight gain.
Conclusions
Our findings suggest that GWL may increase the risk of adverse perinatal outcomes across all obesity classes.
Study Importance.
What is already known?
There is no tailoring of gestational weight‐gain recommendations by severity of obesity.
Gestational weight‐management guidelines lack information regarding risks related to weight loss among pregnant individuals with obesity.
What does this study add?
Individuals with obesity lose weight during pregnancy regardless of recommendations.
Gestational weight loss is associated with an increased risk of a composite of adverse perinatal outcomes across all obesity classes.
How might these results change the direction of research or the focus of clinical practice?
The results of this study support the following: tailoring gestational weight changes based on the severity of obesity; informing future gestational weight‐management guidelines with a focus on the pregnant population with obesity; and recognizing that the consequences of weight loss during pregnancy warrant serious consideration.
INTRODUCTION
With the increasing prevalence of individuals with obesity and the need to improve perinatal outcomes in this population [1, 2, 3], gestational weight loss (GWL) as opposed to weight gain has been considered an option by pregnant individuals and their health care providers [4, 5]. Although GWL is not recommended by current guidelines [6], ~8% of individuals with obesity in the United States have reported trying to lose weight during pregnancy, i.e., 2% to 5% with obesity class I, 4% to 9% with class II, and 9% to 16% with class III or higher (III+) [4].
Few studies have explored the association between GWL in individuals with obesity and adverse perinatal outcomes [7, 8, 9, 10], and findings have been mixed and complex. Relying on a single perinatal outcome such as fetal growth or preterm birth may be insufficient to generate a clear and straightforward conclusion on the risks related to GWL. Therefore, using a comprehensive measurement that combines a wide range of perinatal outcomes into a single composite may be the best and most balanced way to assess potential risks related to GWL among individuals with obesity [11].
Previous research on the association between GWL and adverse perinatal outcomes has commonly reported important limitations. Available studies have been based on small sample sizes; did not take into consideration important confounding factors such as race and ethnicity and socioeconomic status; lacked uniformity in defining obesity classes; or lacked information on maternal health problems that could cause unintended weight loss (e.g., maternal preexisting health conditions, pregnancy complications, psychosocial factors, etc.) [12, 13, 14, 15]. Moreover, it has been suggested that actual units of weight change (kilograms or pounds) may not be the most practical approach to address weight management, and the percentage of body weight change is recommended for use in the general population clinical practice [16, 17].
There is no tailoring of gestational weight‐gain recommendations by severity of obesity, and guidelines lack any information regarding risks related to GWL among individuals with obesity [6]. Consequently, the Institute of Medicine (IOM) together with health care providers has called for research to review gestational weight‐management recommendations for the population with obesity [5]. Because clinical trials of GWL are not ethical due to potential perinatal risks, a high‐quality, population‐based observational study is urgently needed to explore risks related to GWL. To address this gap, our study aimed to assess the impact of GWL on a composite measure of severe adverse perinatal outcomes [18] and has adopted both approaches of weight‐change measurement (percentage and total) to examine the associations using a world‐leading birth registry and perinatal database.
METHODS
Study design and population
This is a population‐based retrospective cohort study using Better Outcomes Registry and Network (BORN) data in Ontario, Canada (https://www.bornontario.ca/en/about-born/) from April 1, 2012, to March 31, 2020. Detailed information regarding BORN data [19, 20, 21] and methodology of gestational weight‐gain measurement [7] adopted in this study has been described in previous publications. This study was approved by the Children's Hospital of Eastern Ontario (19/09PE) and Ottawa Health Science Network (20190704‐01K) Research Ethics Boards.
Singleton pregnancies resulting in an Ontario birth between 2012 and 2020 with gestational age ≥ 22 weeks or birth weight ≥ 500 g were included in this study. The study cohort was limited to pregnant individuals with prepregnancy obesity (body mass index [BMI] ≥ 30 kg/m2).
Data sources and data linkage
BORN records were linked with maternal obstetrical and stillbirth/newborn discharge abstracts from the Canadian Institute for Health Information's (CIHI) Discharge Abstract Database (DAD) to identify severe adverse neonatal health conditions and improve ascertainment of some maternal health conditions, including those potentially causing significant unintended weight loss during pregnancy that are not collected by BORN. BORN‐CIHI DAD linkage tables have been created using a combination of deterministic (using Ontario Health Insurance Plan [OHIP] numbers) and probabilistic approaches [22, 23, 24]. An OHIP number is a unique six‐digit number that tells the Canadian Ministry of Health and Long‐Term Care whom claims are coming from. Additional linkage to 2011 Canadian Census data using Statistics Canada's postal code conversion file has been completed through the maternal residential postal code, which allowed us to obtain neighborhood income level and education status.
Measures
The main outcome of the study is a composite of adverse perinatal outcomes, including perinatal death and neonatal morbidity. Perinatal death includes stillbirth and neonatal death within 28 days of delivery. Neonatal morbidity includes any of the following: gestational age at delivery < 32 weeks, birth weight < 1500 g, respiratory distress syndrome, seizure, intraventricular hemorrhage (grades 2, 3, or 4), cerebral infarction, periventricular leukomalacia, severe birth trauma, hypoxic–ischemic encephalopathy, necrotizing enterocolitis, bronchopulmonary dysplasia, sepsis/septicemia, pneumonia, and other complications such as primary atelectasis, respiratory failure, resuscitation, ventilatory support (mechanical ventilation or continuous positive airway pressure), central venous or arterial catheter, transfusion of blood or blood products, and pneumothorax requiring an intercostal catheter. The indicator of neonatal morbidity has been described and analyzed in previous studies [18, 25].
Average weekly weight change during the second and third trimesters of pregnancy (in kilograms per week and percentage per week) is the exposure variable of interest and was modeled as both a continuous and categorical variable in our analyses. Total gestational weight change is derived from the difference between maternal weight at delivery and prepregnancy weight, as recorded in the BORN database. Prepregnancy weight was estimated by weight measured at the first trimester visit minus 2 kg or was self‐reported during pregnancy. Maternal weight at delivery was measured or self‐reported at delivery, and, if not available, the weight at the last prenatal visit was used. This weight‐change calculation has been reported in other studies that have used BORN data [7, 26], and a description of the weight‐change measure is provided in online Supporting Information.
When coding the gestational weight change as a categorical variable, adequate gestational weight gain based on the 2009 US IOM guidelines (0.17–0.27 kg/week for individuals with obesity [6]) is the reference group, and the comparison groups include GWL (≤0 kg/week), inadequate weight gain (>0 to <0.17 kg/week), and excessive weight gain (>0.27 kg/week) [7, 26]. Confounders and covariates were selected based on the literature, review of a directed acyclic graph, and data availability. Potential confounders and covariates include prepregnancy BMI, maternal age at delivery, parity, urban/rural residence, maternal race, median household income quintile, neighborhood education quintile (percentage of university degrees among adults ages 25–64 years), smoking during pregnancy (self‐reported any smoking at the first prenatal visit or at the time of labor/admission for delivery in patient health records), drug use during pregnancy, conception type, antenatal health care provider, prenatal classes, composite preexisting maternal health conditions, anxiety, depression, domestic/intimate partner violence, gestational diabetes, and hypertensive disorders of pregnancy. Data from BORN records are from patient health records and have been generated through conversations between health care providers and patients during their appointments.
We hypothesized that prepregnancy obesity class may act as an effect modifier of the association between weight change during pregnancy and risk of a composite of severe adverse perinatal outcomes. Therefore, we conducted stratified analyses by obesity class: class I (BMI 30.0–34.9 kg/m2); class II (BMI 35.0–39.9 kg/m2); class III (BMI 40.0–44.9 kg/m2); class IV (BMI 45.0–49.9 kg/m2); and class V (BMI ≥ 50 kg/m2). Owing to small sample sizes, obesity classes III, IV, and V were grouped as obesity class III+ for most of the stratified analyses.
Statistical analysis
Descriptive analysis
Summary statistics describe pregnant individuals with obesity based on gestational weight change: GWL; inadequate weight gain; adequate weight gain (reference); and excessive weight gain. Continuous variables were described by mean ± standard deviation (SD). Categorical variables were described by count and percentage. Prevalence of overall GWL was examined among individuals with obesity overall (BMI ≥ 30 kg/m2) and by obesity classes (I–V).
Association between GWL and risk of a composite of severe adverse perinatal outcomes
Bivariate regression analyses assessed the association between individual covariates and the composite perinatal outcomes. Multivariable modified Poisson [27] regression models were fit to estimate adjusted risk ratio (aRR) and 95% confidence interval (CI) for the association between gestational weight change and risk of a composite of adverse perinatal outcomes. The multivariable models adjusted for prespecified covariates and potential confounders and accounted for repeated pregnancies during the 2012 to 2020 period using generalized estimation equation methods to adjust for variance. Multiple imputation methods accounted for the missing data of covariates and confounders in the regression models. Five complete data sets were imputed using fully conditional specification (FCS) methods by using the FCS statement in PROC MI in SAS software version 9.4 (SAS Institute, Inc., Cary, North Carolina). Conception type was imputed using a generalized logit model. Maternal age at birth, parity, urban/rural residence, median household income quintile, neighborhood education quintile, smoking during pregnancy, drug use during pregnancy, antenatal health care provider, and prenatal classes were imputed using logistic regression models. The results of the analyses from the five imputed data sets were combined per Rubin's rules to account for uncertainty due to imputation, using PROC MIANALYZE capabilities within SAS software.
Stratified analyses were conducted to explore the association between gestational weight change and a composite of severe adverse perinatal outcomes by obesity class.
Nonlinear dose–response relationship between GWL and risk of a composite of adverse perinatal outcomes
The nonlinear associations of gestational weight change and the risk of a composite of adverse perinatal outcomes were examined by using modified Poisson regression models with restricted cubic spline terms to represent weight change in pregnancy. This analysis was conducted using the first imputed data set. Weekly average gestational weight change was modeled using restricted cubic splines, with five knots placed at quantiles recommended by Harrell: 0.05; 0.275; 0.50; 0.725; and 0.950 [28]. These regression models were adjusted for the confounders specified in the categorical analysis. RR for the composite of adverse perinatal outcomes were calculated as the predicted probability of outcome in the regression model divided by predicted probability at a chosen reference value, which was set to the IOM‐recommended value for weekly average weight change (i.e., 0.22 kg/week). 95% CI for the RR were calculated using the bootstrap percentile method with 1000 replicates [29].
Sensitivity analysis
A series of sensitivity analyses were conducted to assess the impact of our analytical methods on the findings. First, complete case analyses assessed the effectiveness of the multiple imputation approach. In addition, we used a percentage‐based approach to quantify weight change during pregnancy and its correlation with risk of a composite of adverse perinatal outcomes. This involved calculating weight change as a percentage of baseline prepregnancy weight, which was then treated as a continuous exposure variable in the cubic spline analysis. Details related to the weight‐change measure for this sensitivity analysis are included in online Supporting Information.
To address the potential confounding effect of maternal health conditions on the relationship between GWL and adverse perinatal outcomes, individuals experiencing health conditions associated with unintended weight loss were excluded from the sensitivity analysis cohort. These conditions included tuberculosis, hepatitis, HIV, cancer, hypertension, autoimmune disorders, diabetes, hypothyroidism, hyperthyroidism, other endocrine disorders, anxiety, depression, eating disorders, gastrointestinal diseases, and hyperemesis gravidarum requiring hospital admission. Details regarding BORN and International Classification of Diseases, Tenth Revision (ICD‐10) codes are provided in Table S1.
SAS software version 9.4 (SAS Institute Inc.) was used to perform all other analyses. We used two‐tailed tests of statistical significance with a threshold of p < 0.05.
RESULTS
Among individuals with obesity, a total of 157,205 births were identified between 2012 and 2020. Sociodemographic data categorized by weight‐change groups, i.e., GWL (n = 9584; 6.1%), inadequate weight gain (n = 19,770; 12.6%), adequate weight gain (n = 19,186; 12.2%), and excessive weight gain (n = 108,665; 69.1%), are described in Table 1. The prevalence of GWL (median = −2.5 kg; interquartile [IQ]1 = −6.1, IQ3 = −0.7) increased with the severity of obesity, with 3.9% of individuals in obesity class I, 7.3% in class II, 11.3% in class III, 14.7% in class IV, and 15.6% in class V.
TABLE 1.
Demographic information of individuals with obesity and uncomplicated pregnancies resulting in a hospital birth in Ontario between 2012 and 2020 by gestational weight change (n = 157,205)
| GWL | Inadequate weight gain | Adequate weight gain | Excessive weight gain | Total | p value | ||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| n = 9584 | n = 19,770 | n = 19,186 | n = 108,665 | n = 157,205 | |||||||
| n | % (col) | n | % (col) | n | % (col) | n | % (col) | n | % (col) | ||
| Prepregnancy BMI (mean ± SD) | (38.2 ± 6.4) | (36.6 ± 5.4) | (35.8 ± 5.0) | (34.9 ± 4.9) | (35.4 ± 5.2) | <0.0001 | |||||
| 30.0–34.9 kg/m2 (obesity class I) | 3587 | 37.4 | 9345 | 47.3 | 10,404 | 54.2 | 69,358 | 63.8 | 92,694 | 59.0 | |
| 35.0–39.9 kg/m2 (obesity class II) | 2894 | 30.2 | 5924 | 30 | 5458 | 28.4 | 25,623 | 23.6 | 39,899 | 25.4 | |
| 40.0–44.9 kg/m2 (obesity class III) | 1772 | 18.5 | 2906 | 14.7 | 2222 | 11.6 | 8848 | 8.1 | 15,748 | 10.0 | |
| 45.0–49.9 kg/m2 (obesity class IV) | 832 | 8.7 | 1092 | 5.5 | 762 | 4 | 2970 | 2.7 | 5656 | 3.6 | |
| ≥50 kg/m2 (obesity class V) | 499 | 5.2 | 503 | 2.5 | 340 | 1.8 | 1866 | 1.7 | 3208 | 2.0 | |
| Maternal age at delivery, y (mean ± SD) | (30.6 ± 5.4) | (31.0 ± 5.3) | (31.2 ± 5.2) | (30.8 ± 5.2) | (30.9 ± 5.2) | <0.0001 | |||||
| ≤19 | 113 | 1.2 | 186 | 0.9 | 192 | 1.0 | 1517 | 1.4 | 2008 | 1.3 | |
| 20–34 | 7194 | 75.1 | 14,369 | 72.7 | 13,898 | 72.5 | 80,280 | 73.9 | 115,741 | 73.7 | |
| 35–39 | 1783 | 18.6 | 4151 | 21.0 | 4018 | 21.0 | 21,722 | 20.0 | 31,674 | 20.2 | |
| ≥40 | 485 | 5.1 | 1056 | 5.3 | 1065 | 5.6 | 5109 | 4.7 | 7715 | 4.9 | |
| Missing | 9 | 0.1 | 8 | 0.0 | 13 | 0.1 | 37 | 0.0 | 67 | 0.0 | |
| Gestational age at delivery, wk (mean ± SD) | (38.7 ± 2.5) | (39.0 ± 1.9) | (39.1 ± 1.8) | (39.2 ± 1.8) | (39.1 ± 1.9) | <0.0001 | |||||
| <37 | 1003 | 10.5 | 1468 | 7.4 | 1179 | 6.1 | 7446 | 6.9 | 11,096 | 7.1 | |
| ≥37 | 8581 | 89.5 | 18,302 | 92.6 | 18,007 | 93.9 | 101,219 | 93.1 | 146,109 | 92.9 | |
| Parity | <0.0001 | ||||||||||
| Nulliparous | 2755 | 28.9 | 5763 | 29.3 | 5972 | 31.3 | 43,922 | 40.6 | 58,412 | 37.3 | |
| Multiparous | 6777 | 71.1 | 13,932 | 70.7 | 13,129 | 68.7 | 64,274 | 59.4 | 98,112 | 62.7 | |
| Missing | 52 | 0.5 | 75 | 0.4 | 85 | 0.4 | 469 | 0.4 | 681 | 0.4 | |
| Urban/rural residence | 0.6436 | ||||||||||
| Urban | 7919 | 83.6 | 16,216 | 83.1 | 15,743 | 83.2 | 89,405 | 83.4 | 129,283 | 83.3 | |
| Rural | 1556 | 16.4 | 3306 | 16.9 | 3171 | 16.8 | 17,830 | 16.6 | 25,863 | 16.7 | |
| Missing | 109 | 1.1 | 248 | 1.3 | 272 | 1.4 | 1430 | 1.3 | 2059 | 1.3 | |
| Maternal race | <0.0001 | ||||||||||
| White | 3272 | 63.8 | 7905 | 65.2 | 8103 | 66.5 | 51,316 | 71.7 | 70,596 | 69.9 | |
| Asian | 701 | 13.7 | 1811 | 14.9 | 1832 | 15.0 | 9168 | 12.8 | 13,512 | 13.4 | |
| Black | 797 | 15.5 | 1556 | 12.8 | 1367 | 11.2 | 6297 | 8.8 | 10,017 | 9.9 | |
| Other | 361 | 7.0 | 843 | 7.0 | 874 | 7.2 | 4780 | 6.7 | 6858 | 6.8 | |
| Missing | 4453 | 46.5 | 7655 | 38.7 | 7010 | 36.5 | 37,104 | 34.1 | 56,222 | 35.8 | |
| Neighborhood household median income quintile | <0.0001 | ||||||||||
| Quintile 1 (lowest) | 2301 | 24.7 | 4492 | 23.3 | 4114 | 22.0 | 21,415 | 20.2 | 32,322 | 21.1 | |
| Quintile 2 | 1679 | 18.0 | 3436 | 17.9 | 3345 | 17.9 | 18,188 | 17.2 | 26,648 | 17.4 | |
| Quintile 3 | 1845 | 19.8 | 3728 | 19.4 | 3740 | 20.0 | 20,923 | 19.8 | 30,236 | 19.8 | |
| Quintile 4 | 1976 | 21.2 | 4303 | 22.4 | 4151 | 22.2 | 24,645 | 23.3 | 35,075 | 22.9 | |
| Quintile 5 (highest) | 1515 | 16.3 | 3279 | 17.0 | 3311 | 17.7 | 20,650 | 19.5 | 28,755 | 18.8 | |
| Missing | 268 | 2.8 | 532 | 2.7 | 525 | 2.7 | 2844 | 2.6 | 4169 | 2.7 | |
| Neighborhood education quintile a | <0.0001 | ||||||||||
| Quintile 1 (lowest) | 1782 | 18.9 | 3674 | 18.9 | 3347 | 17.8 | 18,205 | 17.0 | 27,008 | 17.5 | |
| Quintile 2 | 2062 | 21.9 | 4041 | 20.8 | 3843 | 20.4 | 21,926 | 20.5 | 31,872 | 20.6 | |
| Quintile 3 | 2168 | 23.0 | 4395 | 22.6 | 4354 | 23.1 | 24,218 | 22.6 | 35,135 | 22.7 | |
| Quintile 4 | 2144 | 22.8 | 4572 | 23.5 | 4551 | 24.1 | 25,928 | 24.2 | 37,195 | 24.1 | |
| Quintile 5 (highest) | 1250 | 13.3 | 2742 | 14.1 | 2753 | 14.6 | 16,680 | 15.6 | 23,425 | 15.1 | |
| Missing | 178 | 1.9 | 346 | 1.8 | 338 | 1.8 | 1708 | 1.6 | 2570 | 1.6 | |
| Smoking during pregnancy | <0.0001 | ||||||||||
| Yes | 1642 | 17.4 | 2685 | 13.8 | 2164 | 11.4 | 12,183 | 11.4 | 18,674 | 12.1 | |
| No | 7814 | 82.6 | 16,810 | 86.2 | 16,747 | 88.6 | 94,892 | 88.6 | 136,263 | 87.9 | |
| Missing | 128 | 1.3 | 275 | 1.4 | 275 | 1.4 | 1590 | 1.5 | 2268 | 1.4 | |
| Alcohol exposure during pregnancy | 0.1825 | ||||||||||
| Yes | 235 | 2.5 | 437 | 2.2 | 397 | 2.1 | 2447 | 2.3 | 3516 | 2.3 | |
| No | 9166 | 97.5 | 18,991 | 97.8 | 18,489 | 97.9 | 104,353 | 97.7 | 150,999 | 97.7 | |
| Missing | 183 | 1.9 | 342 | 1.7 | 300 | 1.6 | 1865 | 1.7 | 2690 | 1.7 | |
| Drug exposure during pregnancy b | <0.0001 | ||||||||||
| Yes | 373 | 4.0 | 496 | 2.6 | 425 | 2.3 | 2528 | 2.4 | 3822 | 2.5 | |
| No | 9004 | 96.0 | 18,853 | 97.4 | 18,389 | 97.7 | 103,970 | 97.6 | 150,216 | 97.5 | |
| Missing | 207 | 2.2 | 421 | 2.1 | 372 | 1.9 | 2167 | 2.0 | 3167 | 2.0 | |
| Conception type | 0.4331 | ||||||||||
| In vitro fertilization | 202 | 2.2 | 369 | 1.9 | 377 | 2.0 | 2152 | 2.0 | 3100 | 2.0 | |
| Intrauterine insemination | 212 | 2.3 | 493 | 2.6 | 499 | 2.7 | 2726 | 2.6 | 3930 | 2.6 | |
| No assisted reproductive technology | 8854 | 95.5 | 18,244 | 95.5 | 17,733 | 95.3 | 100,438 | 95.4 | 145,269 | 95.4 | |
| Missing | 316 | 3.3 | 664 | 3.4 | 577 | 3.0 | 3349 | 3.1 | 4906 | 3.1 | |
| Antenatal health care provider | <0.0001 | ||||||||||
| Obstetrician | 7542 | 80.5 | 15,559 | 80.6 | 14,823 | 78.9 | 83,511 | 78.7 | 121,435 | 79.0 | |
| No obstetrician | 1829 | 19.5 | 3756 | 19.4 | 3965 | 21.1 | 22,643 | 21.3 | 32,193 | 21.0 | |
| Missing | 213 | 2.2 | 455 | 2.3 | 398 | 2.1 | 2511 | 2.3 | 3577 | 2.3 | |
| Prenatal education | <0.0001 | ||||||||||
| Yes | 1423 | 16.4 | 3227 | 18.0 | 3392 | 19.5 | 23,316 | 23.6 | 31,358 | 21.9 | |
| No | 7259 | 83.6 | 14,653 | 82.0 | 14,020 | 80.5 | 75,604 | 76.4 | 111,536 | 78.1 | |
| Missing | 902 | 9.4 | 1890 | 9.6 | 1774 | 9.2 | 9745 | 9.0 | 14,311 | 9.1 | |
| Preexisting maternal health conditions c | 0.003 | ||||||||||
| Yes | 1920 | 20.0 | 4054 | 20.5 | 3776 | 19.7 | 21,093 | 19.4 | 30,843 | 19.6 | |
| No | 7664 | 80.0 | 15,716 | 79.5 | 15,410 | 80.3 | 87,572 | 80.6 | 126,362 | 80.4 | |
| Anxiety | <0.0001 | ||||||||||
| Yes | 1531 | 16.0 | 2682 | 13.6 | 2550 | 13.3 | 14,533 | 13.4 | 21,296 | 13.5 | |
| No | 8053 | 84.0 | 17,088 | 86.4 | 16,636 | 86.7 | 94,132 | 86.6 | 135,909 | 86.5 | |
| Depression | <0.0001 | ||||||||||
| Yes | 1408 | 14.7 | 2467 | 12.5 | 2172 | 11.3 | 12,281 | 11.3 | 18,328 | 11.7 | |
| No | 8176 | 85.3 | 17,303 | 87.5 | 17,014 | 88.7 | 96,384 | 88.7 | 138,877 | 88.3 | |
| Domestic/intimate partner violence | <0.0001 | ||||||||||
| Disclosure | 299 | 4.3 | 479 | 3.2 | 417 | 2.9 | 2260 | 2.7 | 3455 | 2.9 | |
| No disclosure | 6710 | 95.7 | 14,306 | 96.8 | 14,046 | 97.1 | 80,154 | 97.3 | 115,216 | 97.1 | |
| Unable to ask | 2575 | 26.9 | 4985 | 25.2 | 4723 | 24.6 | 26,251 | 24.2 | 38,534 | 24.5 | |
| Gestational diabetes | |||||||||||
| Yes | 1666 | 17.4 | 3657 | 18.5 | 3276 | 17.1 | 12,658 | 11.6 | 21,257 | 13.5 | <0.0001 |
| No | 7918 | 82.6 | 16,113 | 81.5 | 15,910 | 82.9 | 96,007 | 88.4 | 135,948 | 86.5 | |
| Hypertensive disorders of pregnancy | |||||||||||
| Yes | 1011 | 10.5 | 2070 | 10.5 | 2128 | 11.1 | 14,897 | 13.7 | 20,106 | 12.8 | <0.0001 |
| No | 8573 | 89.5 | 17,700 | 89.5 | 17,058 | 88.9 | 93,768 | 86.3 | 137,099 | 87.2 | |
| Conditions that may lead to unintentional GWL d | <0.0001 | ||||||||||
| Yes | 4589 | 47.9 | 9244 | 46.8 | 8735 | 45.5 | 47,192 | 43.4 | 69,760 | 44.4 | |
| No | 4995 | 52.1 | 10,526 | 53.2 | 10,451 | 54.5 | 61,473 | 56.6 | 87,445 | 55.6 | |
Note: Data were extracted from the Better Outcomes Registry and Network (BORN) Information System on May 1, 2023. Total gestational weight‐gain recommendations from the IOM 2009 guidelines (which were adopted by Health Canada in 2010) were used to define inadequate, adequate, and excessive gestational weight gain. Because gestational weight gain is associated with gestational length, we accounted for the duration of gestation in our calculations of adequate weight gain. The range for adequate weight gain was calculated based on IOM recommendations for the amount of weight gain during the first trimester (0.5 kg for individuals with obesity) plus the amount of weight gain during the second and third trimester (between [gestational age − 13] × [0.17 kg/week] and [gestational age − 13] × [0.27 kg/week]). If the gestational weight change fell below this range but was greater than 0 kg, then the individual was considered to have had inadequate weight gain. If the gestational weight change was above this range, the individual was considered to have had excessive weight gain. Variables with missing data excluded missing from percentage calculations (i.e., n of numerator/[n of denominator − n of missing]). Records with missing values for any variables required to calculate the numerator were included in the number of missing, and the percentage was calculated by using the following formula: n of missing/n of denominator.
Abbreviations: GWL, gestational weight loss; IOM, Institute of Medicine.
Percentage of university degrees among adults ages 25 to 64 years.
Including cannabis.
Hypertension, heart disease, pulmonary disease, endocrine imbalance, hematologic disorders, cancers, or autoimmune disorders.
Tuberculosis, hepatitis, HIV, cancer, endocrine imbalance, preexisting hypertension, hypertensive disorders of pregnancy, preexisting diabetes, gestational diabetes, autoimmune disorder, anxiety, depression, fetal anomalies, and hyperemesis gravidarum requiring hospital admission.
The majority of those experiencing GWL reduced their weight by 0% to 5% (4.3% of overall individuals), with 2.6% (n = 2398) in obesity class I, 5.1% (n = 2051) in class II, 8.2% (n = 1295) in class III, 10.8% (n = 611) in class IV, and 11.1% (n = 356) in class V. The GWL prevalence, distribution across weight‐loss percentage groups, and summary statistics regarding the percentage of weight change by obesity class are reported in Figure 1. Corresponding results for absolute weight change (in kilograms) are presented in Figure S1.
FIGURE 1.
Prevalence of gestational weight loss (GWL) and percentage of weight change by obesity class, Better Outcomes Registry and Network (BORN) Ontario, 2012 to 2020 (n = 157,205).
The composite of adverse perinatal outcomes was identified in 8.4% (n = 13,269) of the overall cohort cases described in this study. Examining weight changes during pregnancy, individuals who experienced GWL had a higher rate of composite adverse perinatal outcomes (10%) compared with those with inadequate weight gain (8.3%), adequate weight gain (7.3%), and excessive weight gain (8.5%). After adjusting for confounders and imputing missing data, GWL among individuals with obesity was associated with an increased risk of a composite of adverse perinatal outcomes (aRR: 1.31; 95% CI: 1.22–1.39) compared with those with adequate gestational weight gain. The percentage of missing data for covariates can be found in Table 1. The interaction term for obesity class and weight‐gain categories was significant (Wald test p < 0.01). Detailed information on obesity classes and weight‐gain patterns (i.e., inadequate and excessive) can be found in Figure 2. Similar results were obtained from the analysis using complete case data sets, as indicated in Table S2. Moreover, subgroup analysis conducted among individuals without health conditions linked to unintended weight loss also demonstrated consistent findings (Table S3).
FIGURE 2.
Association between weight change during pregnancy and risk of a composite of adverse perinatal outcomes, Better Outcomes Registry and Network (BORN) Ontario, 2012 to 2020 (n = 13,269). *Risk ratios were fitted using Poisson regression models adjusted for prepregnancy BMI, maternal age at birth, parity, urban/rural residence, maternal race, median household income quintile, neighborhood education quintile, smoking during pregnancy, drug use during pregnancy, conception type, antenatal health care provider, prenatal classes, composite preexisting maternal health conditions, anxiety, depression, domestic/intimate partner violence, gestational diabetes, and hypertensive disorders of pregnancy. aRR, adjusted risk ratio.
In exploring the nonlinear dose–response relationship between weekly gestational weight changes and the adopted composite of adverse perinatal outcomes, restricted cubic spline regression analysis revealed a U‐shaped curve. A higher composite of adverse perinatal outcomes risk was noted in the extremities, particularly toward GWL and excessive weight gain. This pattern remained consistent across all obesity classes (I, II, and III+) when examining weight change using total kilograms and percentage change (Figures 3 and 4). Sensitivity analysis conducted among individuals without health conditions linked to unintended weight loss demonstrated similar findings (Figure S2). A comparison of percentage of gestational weight change and weight‐change measures used in the main analysis is described in Figure S3.
FIGURE 3.
Dose–response relationship between weight change during pregnancy (in kilograms) and risk of a composite of adverse perinatal outcomes, Better Outcomes Registry and Network (BORN) Ontario, 2012 to 2020 (n = 13,269).
FIGURE 4.
Dose–response relationship between weight change during pregnancy (in percentage) and risk of a composite of adverse perinatal outcomes, Better Outcomes Registry and Network (BORN) Ontario, 2012 to 2020 (n = 157,205).
DISCUSSION
In this large population‐based retrospective study among 157,205 individuals with obesity, 6.1% (n = 9584) of the cohort experienced GWL. An increase in the GWL prevalence across obesity classes was identified, with 3.9% in class I, 7.3% in class II, 11.3% in class III, 14.7% in class IV, and 15.6% in class V. Our analysis revealed that GWL significantly increased the risk of a composite of adverse perinatal outcomes among individuals with obesity. This association remained statistically significant after stratification by prepregnancy obesity class. Moreover, dose–response relationship analysis revealed a U‐shaped curve, suggesting an elevated risk of a composite of adverse perinatal outcomes following GWL and excessive weight gain. Given the extensive research on the perinatal risks associated with excessive gestational weight gain, our study focused on the potential impact of GWL on perinatal outcomes for individuals with obesity.
Pregnancy is deemed a period that requires weight gain to accommodate the heightened energy demands of both the pregnant individual and the growing fetus [6, 14]. Among individuals with obesity, a minimal amount of gestational weight gain of 5 kg has been considered necessary to support optimal maternal and fetal outcomes [6, 15]. The gestational increase in energy requirements is mainly due to weight gain and higher metabolic rate associated with maternal cardiac output and fetal metabolism throughout mid and late pregnancy [30]. However, mobilization of maternal fat mass in pregnant individuals with obesity may compensate for the energy demand produced by the pregnancy while considering weight maintenance; however, GWL has not been fully studied [30]. Most et al. highlight that advocating weight maintenance for pregnant individuals with obesity is too premature until long‐term effects on offspring development have been determined [30]. Our study demonstrates that GWL is linked to severe adverse perinatal outcomes; therefore, the results warrant serious consideration. With the increased use of some second‐line noninsulin antidiabetic medications during the first trimester and the potential association of these medications with weight loss [31], research on the effects of GWL became vitally important.
Our findings are inconsistent with results from a recent Swedish population‐based cohort study, which reported no increased risk of a composite of adverse perinatal outcomes associated with GWL in pregnancies with class III obesity (aRR: 081; 95% CI: 0.71–0.89) [5]. However, disparities related to individual components of their composite should be noted. The composite of adverse perinatal outcomes analyzed by Johansson et al. [5] encompassed stillbirth; infant death; small for gestational age (SGA) and large for gestational age (LGA) neonates; preterm birth; unplanned cesarean delivery; gestational diabetes; preeclampsia; excess postpartum weight retention; and new‐onset, longer‐term maternal cardiometabolic disease after pregnancy [5]. Although this research excels in including some long‐term outcomes, previous studies have shown that GWL in individuals with obesity is associated with an increased risk of SGA neonates [32, 33] but decreased risk of LGA neonates [7, 13, 32]. The inverse association between GWL and fetal growth outcomes, i.e., SGA and LGA, may invalidate combining these outcomes into a unique composite. Bodnar et al. have previously explored the association between GWL and infant mortality and found an increased risk of infant death among individuals with class I and II obesity [10]. Nevertheless, a systematic review and meta‐analysis including 54 studies representing data from 30,245,946 pregnancies showed favorable results for weight gain lower than what is recommended in the 2009 US IOM recommendation [34]. The evidence gap in research looking into the association between GWL and multiple adverse outcomes remains.
Our multiracial cohort of pregnant individuals with obesity increases the potential generalizability and overall relevance of our findings to other similarly structured regional or subcontinental populations. Data from over 150,000 individuals with obesity allowed us to outline results by obesity classes. Moreover, using a comprehensive measurement that combines a wide range of perinatal morbidity components into a single composite outcome seems the best approach to measure the safety of GWL [11]. To our knowledge, this is the first study to explore the association between GWL on a composite of severe adverse perinatal outcomes that included perinatal death and neonatal morbidity. Several confounders have been taken into consideration during our analysis, including prepregnancy BMI, maternal age at birth, parity, urban/rural residence, maternal race, median household income quintile, neighborhood education, smoking during pregnancy, drug use during pregnancy, conception type, antenatal health care provider, prenatal classes, composite preexisting maternal health conditions, anxiety, depression, domestic/intimate partner violence, gestational diabetes, and hypertensive disorders of pregnancy. Avant‐garde, we performed multivariable regression models with restricted cubic splines to precisely estimate the dose–response relationship between GWL and a composite of adverse perinatal outcomes among individuals with high BMI values overall and per obesity class.
This study is subject to certain limitations, primarily stemming from the use of maternal weight and height data sourced from administrative databases. The BORN registry contains self‐reported data on prepregnancy weight and weight at delivery, which may lead to an underestimation of the prevalence of obesity and gestational weight changes. Nonetheless, previous research has indicated that individuals with higher BMI values tend to underestimate their self‐reported weight [35, 36], and correlated data have been found between self‐reported and measured values in BMI and gestational weight change [37, 38, 39]. Using self‐reported prepregnancy obesity may slightly underestimate our outcome; however, the general trend of association between pregnancy weight and neonatal outcomes will likely not be affected [35]. To mitigate potential limitations related to use of retrospective data, multiple procedures have been taken to verify our data; however, potential for recall bias, misclassification, and residual confounding remains. Our results for individuals with obesity class III+ must be interpreted with caution due to the small sample size and increased risk of placental dysfunction in this population that could affect perinatal outcomes.
CONCLUSION
Our population‐based study, encompassing over 150,000 pregnant individuals with obesity and using high‐quality contemporary data along with a robust modeling strategy, contributes significantly to the limited research on the impact and safety of GWL. Given that GWL has been associated with an increased risk of a composite of adverse perinatal outcomes across all obesity classes, health care providers should exercise caution during gestational weight‐management counseling and alert their patients about potential perinatal risks associated with GWL. This study highlights the importance of updating the 2009 US IOM gestational weight‐gain guidelines to include tailored recommendations based on obesity classes.
CONFLICT OF INTEREST STATEMENT
The authors declared no conflict of interest.
Supporting information
DATA S1: Supplementary Information.
Guo Y, Souza SCS, Bruce L, et al. Risk related to gestational weight loss among individuals with obesity: a population‐based cohort study. Obesity (Silver Spring). 2024;32(12):2376‐2387. doi: 10.1002/oby.24143
REFERENCES
- 1. Fryar CD, Carroll MD, Afful J. Prevalence of overweight, obesity, and severe obesity among adults aged 20 and over: United States, 1960–1962 through 2017–2018. Health E‐Stats. National Center for Health Statistics; 2020. [Google Scholar]
- 2. Martínez‐Hortelano JA, Cavero‐Redondo I, Álvarez‐Bueno C, Garrido‐Miguel M, Soriano‐Cano A, Martínez‐Vizcaíno V. Monitoring gestational weight gain and prepregnancy BMI using the 2009 IOM guidelines in the global population: a systematic review and meta‐analysis. BMC Pregnancy Childbirth. 2020;20(1):649. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Wang MC, Freaney PM, Perak AM, et al. Trends in prepregnancy obesity and association with adverse pregnancy outcomes in the United States, 2013 to 2018. J Am Heart Assoc. 2021;10(17):e020717. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Bish CL, Chu SY, Shapiro‐Mendoza CK, Sharma AJ, Blanck HM. Trying to lose or maintain weight during pregnancy‐United States, 2003. Matern Child Health J. 2009;13(2):286‐292. [DOI] [PubMed] [Google Scholar]
- 5. Johansson K, Bodnar LM, Stephansson O, Abrams B, Hutcheon JA. Safety of low weight gain or weight loss in pregnancies with class 1, 2, and 3 obesity: a population‐based cohort study. Lancet. 2024;403(10435):1472‐1481. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Institute of Medicine (IOM) . Weight Gain During Pregnancy: Reexamining the Guidelines. National Academies Press; 2009. [PubMed] [Google Scholar]
- 7. Guo Y, Souza SCS, Bruce L, et al. Gestational weight loss and fetal growth in uncomplicated pregnancies among women with obesity: a population‐based retrospective cohort study. Int J Obes (Lond). 2023;47(12):1269‐1277. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Petersen JM, Hutcheon JA, Bodnar LM, Parker SE, Ahrens KA, Werler MM. Weight gain patterns among pregnancies with obesity and small‐ and large‐for‐gestational‐age births. Obesity (Silver Spring). 2023;31(4):1133‐1145. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. 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):1642‐1648. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Bodnar LM, Siminerio LL, Himes KP, et al. Maternal obesity and gestational weight gain are risk factors for infant death. Obesity (Silver Spring). 2016;24(2):490‐498. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Korst LM, Fridman M, Lu MC, et al. Monitoring childbirth morbidity using hospital discharge data: further development and application of a composite measure. Am J Obstet Gynecol. 2014;211(3):268.e1‐268.e16. [DOI] [PubMed] [Google Scholar]
- 12. Bogaerts A, Ameye L, Martens E, Devlieger R. Weight loss in obese pregnant women and risk for adverse perinatal outcomes. Obstet Gynecol. 2015;125(3):566‐575. [DOI] [PubMed] [Google Scholar]
- 13. Blomberg M. Maternal and neonatal outcomes among obese women with weight gain below the new Institute of Medicine recommendations. Obstet Gynecol. 2011;117(5):1065‐1070. [DOI] [PubMed] [Google Scholar]
- 14. Kiel DW, Dodson EA, Artal R, Boehmer TK, Leet TL. Gestational weight gain and pregnancy outcomes in obese women: how much is enough? Obstet Gynecol. 2007;110(4):752‐758. [DOI] [PubMed] [Google Scholar]
- 15. Catalano PM, Mele L, Landon MB, et al. Inadequate weight gain in overweight and obese pregnant women: what is the effect on fetal growth? Am J Obstet Gynecol. 2014;211(2):137.e1‐137.e7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Sharma AM, Karmali S, Birch DW. Reporting weight loss: is simple better? Obesity (Silver Spring). 2010;18(2):219. [DOI] [PubMed] [Google Scholar]
- 17. National Institute for Health and Care Excellence (NICE) . Clinical Guideline [CG189]. Obesity: Identification, Assessment and Management . Published November 27, 2014. Updated July 26, 2023. https://www.nice.org.uk/guidance/cg189. [PubMed]
- 18. Dimanlig‐Cruz S, Corsi DJ, Lanes A, et al. Perinatal and pediatric outcomes associated with the use of fertility treatment: a population‐based retrospective cohort study in Ontario, Canada. BMC Pregnancy Childbirth. 2023;23(1):121. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Murphy MSQ, Fell DB, Sprague AE, et al. Data resource profile: Better Outcomes Registry & Network (BORN) Ontario. Int J Epidemiol. 2021;50(5):1416‐1425. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Dunn S, Lanes A, Sprague AE, et al. Data accuracy in the Ontario birth registry: a chart re‐abstraction study. BMC Health Serv Res. 2019;19(1):1001. doi: 10.1186/s12913-019-4825-3 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Dunn S, Bottomley J, Ali A, Walker M. 2008 Niday perinatal database quality audit: report of a quality assurance project. Chronic Dis Inj Can. 2011;32(1):32‐42. [PubMed] [Google Scholar]
- 22. Miao Q, Fell DB, Dunn S, Sprague AE. Agreement assessment of key maternal and newborn data elements between birth registry and clinical administrative hospital databases in Ontario, Canada. Arch Gynecol Obstet. 2019;300:135‐143. [DOI] [PubMed] [Google Scholar]
- 23. Wen SW, Miao Q, Taljaard M, et al. Associations of assisted reproductive technology and twin pregnancy with risk of congenital heart defects. JAMA Pediatr. 2020;174(5):446‐454. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Fell DB, Park AL, Sprague AE, Islam N, Ray JG. A new record linkage for assessing infant mortality rates in Ontario, Canada. Can J Public Health. 2020;111:278‐285. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Ukah UV, Bayrampour H, Sabr Y, et al. Association between gestational weight gain and severe adverse birth outcomes in Washington state, US: a population‐based retrospective cohort study, 2004‐2013. PLoS Med. 2019;16(12):e1003009. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Guo Y, Miao Q, Huang T, et al. Racial/ethnic variations in gestational weight gain: a population‐based study in Ontario. Can J Public Health. 2019;110(5):657‐667. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. Zou G. A modified poisson regression approach to prospective studies with binary data. Am J Epidemiol. 2004;159(7):702‐706. [DOI] [PubMed] [Google Scholar]
- 28. Harrell FE Jr. Ordinal logistic regression. In: Regression Modeling Strategies: With Applications to Linear Models, Logistic and Ordinal Regression, and Survival Analysis. 2nd ed. Springer; 2015:311‐325. [Google Scholar]
- 29. Shao J, Tu D. The Jackknife and Bootstrap. Springer Science & Business Media; 2012. [Google Scholar]
- 30. Most J, Amant MS, Hsia DS, et al. Evidence‐based recommendations for energy intake in pregnant women with obesity. J Clin Invest. 2019;129(11):4682‐4690. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31. Cesta CE, Rotem R, Bateman BT, et al. Safety of GLP‐1 receptor agonists and other second‐line antidiabetics in early pregnancy. JAMA Intern Med. 2024;184(2):144‐152. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32. Kapadia MZ, Park CK, Beyene J, Giglia L, Maxwell C, McDonald SD. Weight loss instead of weight gain within the guidelines in obese women during pregnancy: a systematic review and meta‐analyses of maternal and infant outcomes. PLoS One. 2015;10(7):e0132650. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33. Roussel E, Touleimat S, Ollivier L, Verspyck E. Birthweight and pregnancy outcomes in obese class II women with low weight gain: a retrospective study. PLoS One. 2019;14(5):e0215833. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34. Mustafa HJ, Seif K, Javinani A, et al. Gestational weight gain below instead of within the guidelines per class of maternal obesity: a systematic review and meta‐analysis of obstetrical and neonatal outcomes. Am J Obstet Gynecol MFM. 2022;4(5):100682. [DOI] [PubMed] [Google Scholar]
- 35. Dzakpasu S, Duggan J, Fahey J, Kirby RS. Estimating bias in derived body mass index in the maternity experiences survey. Health Promot Chronic Dis Prev Can. 2016;36(9):185‐193. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36. Robinson EJOR. Overweight but unseen: a review of the underestimation of weight status and a visual normalization theory. Obes Rev. 2017;18(10):1200‐1209. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37. Natamba BK, Sanchez SE, Gelaye B, Williams MA. Concordance between self‐reported pre‐pregnancy body mass index (BMI) and BMI measured at the first prenatal study contact. BMC Pregnancy Childbirth. 2016;16:187. doi: 10.1186/s12884-016-0983-z [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38. Bannon AL, Waring ME, Leung K, et al. Comparison of self‐reported and measured pre‐pregnancy weight: implications for gestational weight gain counseling. Matern Child Health J. 2017;21:1469‐1478. [DOI] [PubMed] [Google Scholar]
- 39. Holland E, Moore Simas TA, Doyle Curiale DK, Liao X, Waring ME. Self‐reported pre‐pregnancy weight versus weight measured at first prenatal visit: effects on categorization of pre‐pregnancy body mass index. Matern Child Health J. 2013;17:1872‐1878. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
DATA S1: Supplementary Information.