Cardiometabolic Implications of Postpartum Weight Changes in the First Year After Delivery

Pregnancy is the only normal physiologic setting in which body weight increases by 20% or more during a 9-month period. After delivery, maternal capacity for restoring normal weight regulation may be further disrupted by lifestyle factors, including lack of time for exercise, smoking cessation, and limited sleep duration (1–3). As such, pregnancy and subsequent postpartum weight retention can significantly alter a woman’s long-term weight gain trajectory (4). Indeed, weight at 1 year postpartum is a stronger predictor of the likelihood of being overweight 15 years later than the weight gained during the pregnancy itself (5). In this context, it has been postulated that the cumulative effect of postpartum weight retention from each pregnancy in a women’s life ultimately contributes to her risk of cardiometabolic disease, including metabolic syndrome, type 2 diabetes, and vascular disease (6–9). Despite its conceptual appeal, however, there has been little direct evidence to support this hypothesis, owing to a paucity of studies assessing both antepartum/postpartum weight trajectory and cardiovascular/metabolic risk factors. Thus our objective in this study was to evaluate the cardiometabolic implications of patterns of postpartum weight change and the time course thereof in the first year after pregnancy.

This analysis was conducted in the setting of an observational study in which a cohort of women is undergoing serial cardiometabolic characterization on three occasions: at the time of recruitment in late pregnancy, at 3 months postpartum, and at 12 months postpartum (10). At our institution, all pregnant women undergo universal screening for gestational diabetes mellitus (GDM) at 24–28 weeks’ gestation by 50 g glucose challenge test (GCT), followed by referral for a diagnostic oral glucose tolerance test (OGTT) if the GCT result is abnormal. In this study, healthy pregnant women were recruited either prior to or just after their GCT (10). Recruitment of women following an abnormal GCT served to enrich the study population for those likely to have GDM. Regardless of their GCT result, all study participants underwent a 3-h 100-g OGTT to ascertain their glucose tolerance status in pregnancy, which ranged from normal glucose tolerance to GDM (diagnosed by National Diabetes Data Group criteria [11]). At 3 and 12 months postpartum, participants returned to the clinical investigation unit to undergo repeat cardiometabolic assessment, including evaluation of glucose tolerance by 2-h 75-g OGTT. The current analysis involved 305 women with singleton pregnancies who had complete cardiometabolic characterization at all three OGTTs. The study protocol has been approved by the Mount Sinai Hospital Research Ethics Board, and all participants have provided written informed consent.

On the morning of the OGTT in pregnancy, interviewer-administered questionnaires were completed pertaining to medical, obstetrical, and family history. Pregravid physical activity in the year before pregnancy was assessed using the Baecke questionnaire, an established instrument that has been extensively validated in several populations, including women of childbearing age (12,13). This questionnaire was completed during the OGTT (i.e., prior to knowing their gestational glucose tolerance status) (14,15). The Baecke questionnaire measures total physical activity and its three component domains: occupation-associated activity (work index), sport-related physical activity (sport index), and nonsport leisure-time activity (leisure-time index). The work index quantifies the exertion related to occupational activities such as sitting, standing, lifting, and walking, as well as associated effects on the individual (e.g., fatigue, perspiration). The sport index characterizes vigorous/sports activity with respect to intensity (using the updated compendium of physical activities) (16), duration, and frequency. The leisure-time index quantifies exertion associated with nonsport recreational activities (e.g., walking, television viewing). Prepregnancy BMI was calculated from participants’ self-reported pregravid weight and their measured height at the time of recruitment in pregnancy.

Participants returned to the clinical investigation unit at both 3 and 12 months postpartum. On both occasions, interviewer-administered questionnaires were completed and physical examination was performed, including measurement of blood pressure, weight, and waist circumference. At both visits, participants underwent a 2-h 75-g OGTT, on which glucose tolerance status was classified according to Canadian Diabetes Association guidelines (17) into the following categories: 1) diabetes, defined as fasting glucose ≥7.0 mmol/L or 2-h glucose ≥11.1 mmol/L; 2) impaired glucose tolerance (IGT), defined by fasting glucose <6.1 mmol/L and 2-h glucose between 7.8 and 11.0 mmol/L inclusive; 3) impaired fasting glucose (IFG), defined as fasting glucose between 6.1 and 6.9 mmol/L inclusive and 2-h glucose <7.8 mmol/L; 4) combined IFG/IGT, defined as fasting glucose between 6.1 and 6.9 mmol/L inclusive and 2-h glucose between 7.8 and 11.0 mmol/L inclusive; and 5) normal glucose tolerance, defined as fasting glucose <6.1 mmol/L and 2-h glucose <7.8 mmol/L. Prediabetes refers to IGT, IFG, or combined IFG/IGT. At 3 months postpartum, the Baecke questionnaire assessed only sport activity and leisure-time activity over the preceding 3 months (as many women would not be working at their usual occupation during that time). At 12 months postpartum, the Baecke questionnaire assessed all three physical activity component domains over the preceding year (since delivery).

All OGTTs were performed in the morning after overnight fast, with venous blood samples drawn for measurement of glucose and insulin at fasting and at 30, 60, and 120 min (and 180 min in pregnancy) after ingestion of the glucose load. Glycemia was assessed by glucose tolerance status and by the area under the glucose curve (AUC_gluc) during the OGTT (calculated by trapezoidal rule). Insulin resistance was assessed with homeostasis model assessment of insulin resistance (HOMA-IR) (18). Whole-body insulin sensitivity was assessed by the Matsuda index (19). Total cholesterol, HDL cholesterol, and triglycerides were measured from fasting serum using the Roche Cobas 6000 c 501 analyzer (Roche Diagnostics, Laval, Quebec, Canada). LDL cholesterol was determined by Friedewald formula. Apolipoprotein-B (apoB) and apolipoprotein-A1 (apoA1) were measured using the Siemens Healthcare Diagnostics BN ProSpec (Siemens Healthcare Diagnostics, Mississauga, Ontario, Canada). Total adiponectin was measured by ELISA (Millipore, Linco Research, St. Charles, MO). High-sensitivity C-reactive protein (CRP) was measured by end point nephelometry using the Dade Behring BN ProSpec and N High Sensitivity CRP reagent (Dade Behring, Mississauga, Ontario, Canada).

All analyses were conducted using SAS.9.2 (SAS Institute, Cary, NC). Continuous variables were tested for normality of distribution, and natural log transformations of skewed variables were used, where necessary, in subsequent analyses. Study participants were categorized based on their weight change between prepregnancy and 3 months postpartum (as either loss or gain) and their weight change between 3 and 12 months postpartum (as either loss or gain). Based on their respective weight changes over these two periods of time, participants were stratified into the following four groups: 1) loss/loss, 2) gain/loss, 3) loss/gain, and 4) gain/gain (Fig. 1). In Table 1, continuous variables were compared across these groups by ANOVA, and categorical variables were compared by χ² or Fisher exact test. For each study group, continuous variables are presented as mean ± SD (if normally distributed) or median with interquartile range (if skewed), and categorical variables are presented as percentages. Adjusted mean levels of systolic blood pressure, diastolic blood pressure, HOMA-IR, adiponectin, LDL, and apoB were compared across the four groups by ANCOVA, after adjustment for age, ethnicity, family history of diabetes, parity, prepregnancy BMI, months postpartum at assessment, and duration of breastfeeding (Fig. 2). Sensitivity analyses were conducted in which these adjusted mean levels were further adjusted for GDM, prediabetes/diabetes at 3 months postpartum, and total physical activity in the first year, respectively. Logistic regression analysis was performed to determine whether specific pregravid/antepartum and postpartum factors predicted weight gain between 3 and 12 months postpartum, after adjustment for age, ethnicity, prepregnancy BMI, and family history of diabetes (Fig. 3).

The patterns of weight change in the first year postpartum are shown in Fig. 1. The majority of women (81.0%) had higher weight at 3 months postpartum as compared with prepregnancy (gained weight). Between 3 and 12 months, most women (74.4%) lost weight. Based on their weight change between prepregnancy and 3 months postpartum (either loss or gain) and between 3 and 12 months postpartum (either loss or gain), we stratified the participants into the following four groups: loss/loss (n = 25), gain/loss (n = 202), loss/gain (n = 33), and gain/gain (n = 45) (Fig. 1).

Table 1 shows the pregravid, antepartum, and postpartum characteristics of the four study groups. At the antepartum OGTT, the groups differed with respect to ethnicity (P = 0.03), prepregnancy BMI (P < 0.0001), gestational weight gain up to the OGTT (P < 0.0001), 1-h glucose on the OGTT (P = 0.0006), 2-h glucose on the OGTT (P = 0.002), AUC_gluc (P = 0.003), GDM (P = 0.0002), Matsuda index (P < 0.0001), HOMA-IR (P < 0.0001), adiponectin (P = 0.026), and CRP (P = 0.001). The groups were otherwise similar in age, family history of diabetes, parity, smoking exposure, pregravid physical activity, fasting glucose, and lipid profile.

At 3 months postpartum, overall differences were noted between the groups in BMI (P = 0.017), Matsuda index (P = 0.009), HOMA-IR (P = 0.028), total cholesterol (P = 0.006), LDL cholesterol (P = 0.011), and apoB (P = 0.016). However, these differences were modest and did not exhibit a clear pattern in their distribution. In contrast, at 12 months postpartum, profound differences were apparent between the groups, reflecting a distinct pattern. Specifically, there were robust gradients in BMI (P < 0.0001), waist (P < 0.0001), total physical activity (P = 0.005), sport index (P = 0.021), fasting glucose (P = 0.031), Matsuda index (P < 0.0001), HOMA-IR (P < 0.0001), adiponectin (P < 0.0001), CRP (P = 0.005), total cholesterol (P < 0.0001), LDL (P < 0.0001), triglycerides (P = 0.001), and apoB (P < 0.0001). These overall differences were driven by lower physical activity and greater obesity, insulin resistance, hypoadiponectinemia, inflammation, and dyslipidemia in the two groups of women who gained weight between 3 and 12 months postpartum (loss/gain and gain/gain).

To further evaluate the emergent differences in cardiometabolic profile between the loss/loss, gain/loss, loss/gain, and gain/gain groups, we compared their adjusted mean values of vascular and metabolic risk factors at both 3 and 12 months postpartum, after adjusting for the potential confounders age, ethnicity, family history of diabetes, parity, prepregnancy BMI, months postpartum at assessment, and duration of breastfeeding (Fig. 2). As shown in Fig. 2A, there was no difference in mean adjusted systolic blood pressure between the groups at 3 months postpartum (P = 0.52). By 12 months, however, a near-significant gradient had emerged (P = 0.05), with systolic blood pressure progressively increasing from the loss/loss group to gain/loss to loss/gain to gain/gain. Similarly, in Fig. 2B, mean adjusted diastolic blood pressure was similar across the groups at 3 months (P = 0.48) but showed a gradient at 12 months (P = 0.02), with higher levels in women that gained weight between 3 and 12 months (loss/gain and gain/gain) than in their peers. The same pattern was seen with HOMA-IR (Fig. 2C), which did not differ at 3 months (P = 0.29) but then showed significant differences between the groups at 12 months (P = 0.0003). This pattern was mirrored by that for mean adjusted adiponectin (Fig. 2D), which was similar across the groups at 3 months (P = 0.89) but showed a stepwise decrease at 12 months from the loss/loss group to gain/loss to loss/gain to gain/gain (P = 0.003). Similar findings were seen with the Matsuda index (3 months P = 0.67; 12 months P = 0.02; data not shown). Finally, with both LDL cholesterol and apoB (Fig. 2E and F, respectively), there were modest differences across the groups at 3 months postpartum (P = 0.01 and P = 0.02) that then became clear gradients from loss/loss to gain/loss to loss/gain to gain/gain at 12 months (LDL P = 0.001; apoB P < 0.0001).

Having demonstrated that marked differences in cardiometabolic profile emerged between the study groups by 12 months postpartum, we next performed a series of sensitivity analyses to evaluate the robustness of these findings. These analyses showed that the findings in Fig. 2 were unchanged with further adjustment for each of the following: 1) GDM, 2) the diagnosis of either prediabetes or diabetes on the OGTT at 3 months postpartum (except for attenuation of diastolic blood pressure to P = 0.07), and 3) total physical activity in the first year postpartum (data not shown). Finally, the adverse cardiometabolic profile that emerged by 12 months postpartum in relation to the pattern of weight change appeared to be comprised of those components shown in Fig. 2 (blood pressure, insulin resistance, adiponectin, LDL, and apoB). Indeed, after adjustment for covariates, other cardiometabolic factors did not differ between the groups at 12 months postpartum (including fasting glucose, 2-h glucose, AUC_gluc, CRP, HDL, and triglycerides).

Having demonstrated that weight gain from 3 to 12 months postpartum identified women with an adverse cardiometabolic profile (i.e., loss/gain and gain/gain groups), we next sought to identify predictors of weight gain during this 9-month window. Specifically, we constructed a series of logistic regression models that individually evaluated the relationship between pregravid, antepartum, and postpartum factors and (dependent variable) weight gain between 3 and 12 months, after adjustment for age, ethnicity, prepregnancy BMI, and family history of diabetes. As shown in Fig. 3, among pregravid and antepartum factors, only antepartum insulin sensitivity (Matsuda index) was independently and inversely associated with the likelihood of gaining weight (odds ratio [OR] = 0.859; 95% CI 0.749–0.985). Among postpartum factors, total physical activity in the first year postpartum (OR = 0.766; 95% CI 0.615–0.954) and specifically sport index (OR = 0.615; 95% CI 0.412–0.918) were inverse predictors of weight gain between 3 and 12 months postpartum. Of note, prepregnancy BMI was a consistent positive predictor in all of the models (data not shown). Thus, while pregravid adiposity is a determinant of weight gain between 3 and 12 months postpartum, its effects may be partially mitigated by maternal physical activity and particularly sport activity in the first year after delivery.

In this study, we explored the effect of patterns of postpartum weight change on cardiometabolic health. First, gain/loss was the most common weight pattern in the study population, indicating that the majority of women did not return to their prepregnancy weight at 3 months postpartum but did lose weight in the 9 months thereafter (though not necessarily returning to their prepregnancy weight). Importantly, we further demonstrate that not losing, but rather gaining, weight between 3 and 12 months postpartum was associated with the development of an adverse cardiometabolic risk factor profile. Specifically, women who gained weight between 3 and 12 months postpartum had higher blood pressure, greater insulin resistance, lower adiponectin, higher LDL cholesterol, and higher apoB levels than their peers. Finally, greater physical activity, particularly sport activity, was associated with a lesser likelihood of gaining weight between 3 and 12 months postpartum. Taken together, these data suggest that weight change during this time may provide a means of identifying women at risk for cardiometabolic disease, and physical activity may offer an approach for modifying this risk.

It has been previously suggested that the cumulative effect of postpartum weight retention from each pregnancy in a woman’s life contributes to her ultimate risk of metabolic and vascular disease (6–9). The current study provides evidence in support of this concept. We show that the majority of women exhibit some degree of gestational weight retention at 3 months after delivery, consistent with other studies in the early postpartum (20–22). An earlier study noted that excess adipose tissue remaining at 6 and 12 months postpartum tends to be localized centrally (23), suggestive of visceral fat accumulation, which is more metabolically deleterious than subcutaneous fat. In this context, we show that, while most women lose some of this pregnancy-acquired weight between 3 and 12 months postpartum, those who do not lose weight in this period indeed develop an adverse cardiometabolic profile that was not present at 3 months. This profile includes insulin resistance and hypoadiponectinemia, both of which are potentially consistent with visceral fat accumulation (24,25). These data thus suggest that failure to lose pregnancy-acquired weight has negative cardiometabolic implications, and these adverse sequelae emerge quite early after delivery.

There are several points to recognize about these cardiometabolic sequelae. First, the between-group differences in blood pressure, insulin resistance, adiponectin, LDL, and apoB were independent of adjustment for covariates, including prepregnancy BMI, age, ethnicity, family history, parity, and breastfeeding. Second, these differences were generally unchanged by adjustment for either GDM (which is associated with chronic metabolic dysfunction) (26) or the diagnosis of prediabetes/diabetes at 3 months postpartum (diagnoses that may influence lifestyle practices). Third, it is important to note that the absolute levels of the clinical cardiometabolic risk factors in question (blood pressure, LDL, apoB) were not necessarily striking from a practitioner perspective. However, long-term exposure to the gradients that emerge after stratification of women by their patterns of postpartum weight change may ultimately lead to differential risks of diabetes and cardiovascular disease, as previously suggested (27).

Thus a key clinical implication of these data is that weight gain between 3 and 12 months postpartum may act as a marker for women at risk for cardiometabolic disease. It is currently unknown whether this group of women can be identified prior to this 9-month window. In the current study, the only consistent clinical predictor of future weight gain during this window was prepregnancy BMI (antepartum insulin sensitivity, though significant, is not measured in practice). The other clinical question that emerges is whether intervention can be implemented for those women who do not lose weight between 3 and 12 months in order to reduce their cardiometabolic risk. One potential option suggested by our data is that sport activity warrants investigation as a lifestyle intervention that may reduce the likelihood of weight retention in this setting. Of note, the moderating effect of sport activity observed in this study is consistent with an earlier report that showed that aerobic exercise in the first 6 months postpartum was associated with lesser weight gain at mean 8.5 years after delivery (28).

To our knowledge, our study is the first to investigate weight patterns during the first year postpartum and their relation to changes in cardiometabolic health. A strength of this study is its prospective systematic assessment of cardiometabolic function in a well-characterized cohort of women at three points in time during and after pregnancy. This design enabled demonstration of the emergence of cardiometabolic consequences by 12 months postpartum that were not present 9 months earlier. Because of the observational nature of this study, we cannot imply causality in the relationship between sport index and reduced risk of weight gain and cannot exclude the possibility that physical activity is a marker for some other unmeasured factor. The study is also limited by the use of self-reported prepregnancy weight, which is susceptible to recall bias. These limitations, however, do not undermine our demonstration of the incident emergence of cardiometabolic sequelae between 3 and 12 months after delivery in relation to directly measured patterns in weight change over that time.

Our findings highlight the cardiometabolic implications of postpartum weight change patterns and emphasize the importance of limiting gestational weight retention as early as 1 year postpartum. Future study is needed for evaluation of the cardiometabolic effects of weight retention in the subsequent years beyond the first year. Additionally, interventional studies should be considered to promote weight loss between 3 and 12 months postpartum. Lifestyle modification particularly targeting sport activity in the postpartum is an intervention that warrants consideration.

In summary, while most women exhibit some gestational weight retention 3 months after delivery, those who do not lose weight in the subsequent 9 months develop an adverse cardiometabolic profile by as early as 12 months postpartum. This profile consists of higher blood pressure, greater insulin resistance, lower adiponectin, higher LDL cholesterol, and higher apoB levels. As the observed risk factor gradients are not present at 3 months but are readily apparent at 12 months, these data highlight the time course linking postpartum weight changes with incident cardiometabolic sequelae. Furthermore, they implicate the period between 3 and 12 months postpartum as a critical window during which patient and practitioner attention to weight control may be important to long-term metabolic and vascular health.

Funding. This study was supported by operating grants from the Canadian Institutes of Health Research (MOP-84206), the Heart and Stroke Foundation of Ontario (NA6747), and the Canadian Diabetes Association (OG-3-10-3051-BZ and OG-3-11-3300-RR). A.J.H. holds a Tier II Canada Research Chair in Diabetes Epidemiology, B.Z. holds the Sam and Judy Pencer Family Chair in Diabetes Research at Mount Sinai Hospital and University of Toronto, and R.R. holds an Ontario Ministry of Research and Innovation Early Researcher Award.

Duality of Interest. No potential conflicts of interest relevant to this article were reported.

Author Contributions. S.K. researched the data, contributed to the analysis and interpretation of the data and critical revision of the manuscript, and wrote and approved the manuscript. C.Y. performed the statistical analyses, contributed to interpretation of the data and critical revision of the manuscript, and approved the manuscript. A.J.H., P.W.C., M.S., and B.Z. were involved in the design and implementation of the overall study, contributed to interpretation of the data and critical revision of the manuscript, and approved the manuscript. R.R. was involved in the design and implementation of the overall study, designed the analysis plan, contributed to interpretation of the data and critical revision of the manuscript, supervised the analysis and the manuscript, and approved the manuscript. R.R. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.