Several large national and international prospective studies have linked obesity to incident chronic disease.1–5 Risk of type 2 diabetes, hypertension, dyslipidemia, stroke, heart disease, osteoarthritis, gallbladder disease, and certain cancers has been shown to be higher with increasing level of obesity. Several studies have also shown a relationship between obesity and all-cause or specific-cause mortality.6,7 Several more recent studies have failed to show an increase in mortality for the overweight category but do demonstrate higher mortality for the obese or morbidly obese category.8,9
The prevalence of overweight and obese adults has increased steadily since the early 1980s.10 In 1999–2000, the age-adjusted prevalence of obesity, defined as a body mass index (BMI) equal to or greater than 30 kg/m2, was 31%, compared with 23% in 1988–1994.10 The prevalence of overweight (BMI > 25) increased from 56% to 65% during this same time. Freedman et al11 reported a national trend toward an increase in the prevalence of class 3 obesity (BMI > 40) from 0.78% to 2.2% from 1990 to 2000. Although increases were similar across all age and gender groups, earlier studies suggested a significantly higher risk of developing significant obesity for women in middle age, particularly overweight women.12,13 Mokdad et al13 reported that the greatest increase in obesity from 1991 to 1998 was in 18- to 29-year-old women and in those with some college education.
Young adulthood may be a time when significant weight changes may occur. For some, pregnancy may represent the advent of weight control issues. Lifestyle and body image changes may encourage weight gain, setting a pattern for the next several decades. Many obese women attribute their adult weight gain to pregnancy.14,15 At no other time in a woman’s life is she encouraged to gain 25 pounds. Given the significant increase in obesity rates and the connection of obesity to morbidity and mortality (especially among the morbidly obese), we set out to estimate what role pregnancy-related weight gain and factors surrounding pregnancy may have on the development of obesity in midlife. We also set out to estimate what relationship this weight gain had to the development of diabetes and heart disease later in life.
MATERIALS AND METHODS
The original study was conducted at the Gundersen Lutheran Medical Center in La Crosse, Wisconsin, from April 12, 1988, to October 12, 1990, and had full institutional review board approval. Methods of this study are described elsewhere and summarized briefly below.16,17 All women having an uncomplicated pregnancy were included in the study. Weight measurements were obtained at the first prenatal visit (average of 10.3 weeks of gestation), at 20 weeks, and at term. In the postpartum, a weight was measured at the second day postpartum and at 2, 4, 6, 8, 12, and 24 weeks (6 months postpartum). Weights during pregnancy were measured on scales in the obstetrics or family practice departments and, in the postpartum period, in the family practice or pediatrics department. An additional survey was completed at each measurement, inquiring about certain behaviors, including breastfeeding, use of alcohol or tobacco, exercise type and frequency, sexual activity and contraceptive use, and resumption of work.
At approximately 4 years (unpublished) and 10 years17 after the initial study, a medical record review was conducted to record the last known weight. In 2004, 15 years later, a similar review was conducted, gathering information on last known weight. New to this follow-up, we recorded any obesity-related diseases documented in the medical record. The diseases recorded were diabetes (type 2), metabolic syndrome, heart disease, dyslipidemia, hypertension, asthma, depression, anxiety, gallbladder disease, thyroid conditions, headache, arthritis and other joint conditions, back conditions, sleep apnea, gastric conditions, and all forms of cancer.
The original cohort included 795 women. Excluded from the current follow-up study were 7 mothers who gave birth to twins, 1 mother who was missing all follow-up weight measures, and 6 mothers who were deceased. Of the remaining 781 mothers, 297 were missing weight measurements at 2 of the 3 follow-up periods (4, 10, or 15 years) and were excluded from this analysis. The remaining 484 (61% of the original cohort) had at least 1 of the weight measurements at 4 or 10 years, as well as a weight measurement beyond 10 years. In the event one measurement was missing, the weight at the previous follow-up was substituted. For example, if the 4-year follow-up weight was missing, the 6-month follow-up weight was used in the analysis (n < 15). Multivariable modeling did not differ substantially between those with all weight measurements and the full cohort, including those with substituted weight.
The major dependent variables for this study were weight gain from first prenatal visit to last known weight, current BMI (at last known follow-up), and presence of an obesity-related illness. Body mass index was calculated as weight in kilograms, divided by a squared measure of height in meters. Two obesity-related diseases or conditions examined were diabetes/prediabetes and heart disease/pre–heart disease. Diabetes/prediabetes (DM/pre-DM) was classified as having either a diagnosis of diabetes or metabolic syndrome or 3 or more of the 5 risk factors for metabolic syndrome (using BMI > 30 instead of waist circumference, triglycerides > 150 mg/dL, high-density lipoprotein cholesterol < 50 mg/dL, blood pressure > 130/> 85 mm Hg, and fasting glucose > 110 mg/dL).18 Patients diagnosed with heart disease, hypertension, or dyslipidemia were classified as having heart disease/pre–heart disease (CHD/pre-CHD). We chose to combine these categories to increase our ability to develop a stronger model, despite knowing that not all who have hypertension or dyslipidemia will develop heart disease.
Key independent variables included prepregnancy BMI, weight gain during pregnancy, and retained weight by 6 months postpartum. Weight gain categories were based on the Institute of Medicine recommendations that are determined by prepregnancy BMI.19 Women considered underweight (pre-BMI < 19.8) should gain 12.5–18.0 kg, women of normal weight (pre-BMI 19.8–26.0) should gain 11.5–16.0 kg, overweight women (pre-BMI 26.1–29.0) should gain 7.0–11.5 kg, and obese women (pre-BMI > 29.0) should gain no more than 6.8 kg. Weight gains were divided into 3 categories: gaining above the recommended amount, gaining the recommended amount, or gaining below the recommended amount. Weight at 6 months postpartum was divided into 2 categories: having lost all pregnancy weight gained or having any retained weight.
All analysis was conducted using SAS statistical software (SAS Institute Inc, Carey, NC). Univariate analysis included analysis of variance and Pearson and Mantel Haenszel χ2 test. Characteristics from the univariate analysis that were significantly related to current BMI or weight gain, at a P value of .10, were entered into the multivariable analyses. We used a stepwise multiple linear regression to predict current BMI and weight gain. Regression coefficients and their 95% confidence intervals are reported. Similarly, we used a stepwise logistic regression to develop the best model to predict DM/pre-DM and CHD/pre-CHD occurrence. Odds ratios and their 95% confidence intervals are reported. The slopes of the change in BMI or change in weight over time by baseline BMI or disease status were compared by using a repeated measures analysis of variance.
Overall, 484 women were available for analysis. Of those lost to follow-up, 29% moved out of the service area, 50% received primary care elsewhere, and 21% were missing weight measures at both the 4- and 10-year follow-ups. Minor differences were found between those available for follow-up and those lost to follow-up (n = 297). Those lost to follow-up tended to be younger (27.7 years versus 28.6 years), to present for prenatal care later (11.1 weeks versus 9.8 weeks), to be primiparas (44% versus 39%), and to have retained less weight at 6 months postpartum (1.0 kg versus 1.7 kg) (all P values < .05). There were no differences in average baseline BMI (24.7 versus 24.2) or amount of weight gained (12.8 kg versus 13.0 kg).
Average length of follow-up for those in the analysis was 14.7 years (range 10.1–16.3 years). The average age at follow-up was 42.68 (range 28.5–56.0 years). Ninety percent of the women were married at the beginning of the study; 78% of these women were married to the same spouse at follow-up. At the beginning of the study, 88% of the women had private medical insurance compared with 85% at the end of the study.
The average BMI increased significantly from the first prenatal visit to the 15-year follow-up and increased at a significantly higher rate for women who were obese at baseline than for women in the other baseline weight categories (Fig. 1). At 15-year follow-up, 30% of the cohort was obese, 27% were overweight, 39% were normal weight, and 4% were underweight. The average weight gained by 15-year follow-up was 9.69 kg (21 pounds), an average change of 1.4 pounds per year. Weight gain was more dramatic for women who were obese at baseline compared with those who were overweight, normal weight, or underweight. Obese women at baseline were, on average, 15.5 kg (34.1 pounds) heavier 15 years later, an average annual gain of 2.3 pounds, compared with average gains of 1.7 pounds per year for overweight women, 1.2 pounds per year for normal weight women, and 1 pound per year for underweight women.
The following variables were related univariately to current BMI: marital status at delivery, change in marital status, current parity, insurance status at delivery, current insurance status, baseline BMI, weight gain at index pregnancy, retained weight at 6 months postpartum, participation in postpartum aerobic exercise, and duration of breastfeeding. Variables not significantly related to BMI included age (current or at delivery); current marital status; parity at delivery or change in parity; participation in postpartum walking, running, swimming, or biking; smoking status (current or at delivery); type of work; or amount of time between delivery and return to work.
The final multivariable model is shown in Table 1. The model, including marital status at delivery, weight gain during pregnancy, weight loss by 6 months postpartum, duration of breastfeeding, postpartum aerobic exercise, and baseline BMI, explained 63.5% of the variation in current BMI. Because of the high correlation of BMI over time, baseline BMI was removed from the model, which then explained only 12.6% of the variability in BMI. Although marital status was not significant in the overall model, we felt it was an important confounder to include. Women who gained more than recommended, retained pregnancy weight at 6 months postpartum, breastfed for a short duration or not at all, did not participate in aerobic exercise, or were obese at pregnancy had the highest BMI at follow-up.
Weight gain over 15 years was related to current age, marital status at delivery, change in marital status, insurance status at delivery, baseline BMI, weight gain at index pregnancy, retained weight at 6 months postpartum, participation in postpartum aerobic exercise, duration of breastfeeding, and current smoking status. Variables not significantly related to weight gain included age at delivery; current marital status; parity at delivery or change in parity; current insurance status; postpartum participation in walking, running, swimming, or biking; smoking status at delivery; type of work; or amount of time between delivery and return to work.
The final multivariable model to predict weight gain is shown in Table 1. The model was similar to the model to predict current BMI, but it explained less of the variability in weight gain (R2 = 0.167). When baseline BMI was removed from the model, the R2 was 0.130.
Rates of obesity-related diseases were examined overall and compared across BMI categories at 15-year follow-up. Overall, 1.2% of the women had developed diabetes, 11.8% had been diagnosed with, or had 3 or more of the 5 risk factors for, metabolic syndrome; and 13% overall had DM/pre-DM. The rate of DM/pre-DM increased significantly with increasing BMI category (0% of underweight women, 2% of normal weight women, 4% of overweight women, and 38% of obese women).
Overall, 4% of women were diagnosed with heart disease. Additionally, 14% had a diagnosis of dyslipidemia, and 19% had hypertension. When combined, 30% of women had CHD/pre-CHD. The rate of CHD/pre-CHD increased significantly with increasing BMI category (10% of underweight women, 17% of normal weight women, 30% of overweight women, and 51% of obese women).
Other conditions or diseases related to obesity in a linear fashion included sleep apnea (2%), gallbladder disease (9%), and chronic pain, including arthritis, back pain, carpal tunnel, and fibromyalgia (16%) (Mantel-Haenszel χ2 P < .05). Conditions not related to BMI were cancer (15%), asthma (12%), depression or anxiety (35%), thyroid conditions (12%), chronic headache (29%), and gastric conditions (14%).
In addition to the rate of DM/pre-DM varying by current BMI, the following variables were all related to the presence of DM/pre-DM: current and delivery marital status, baseline BMI, weight gain during pregnancy, and participation in postpartum aerobic exercise. The multivariable analysis is found in Table 2. Because no underweight women developed diabetes, the model failed to converge. The model excluding underweight women was similar to a model combining underweight and normal weight categories. Thus, the model combining the 2 categories is shown. Only baseline BMI and change in weight from baseline to present were significantly related to risk of DM/pre-DM. Women overweight at baseline were 3.3 times more likely to develop DM/pre-DM, and obese women were nearly 9 times more likely to have DM/pre-DM at 15-year follow-up. For every 1-kg gain in weight from baseline, there was a 9% increase in the risk of DM/pre-DM.
Weight gain and BMI at each follow-up measure were examined for women who developed DM/pre-DM, compared with women who did not (Fig. 2). The patterns of weight gain and BMI change over time were significantly different from each other (P < .001). Women who developed DM/pre-DM had a significantly higher average BMI at all time points, as well as a dramatic weight increase over the 15 years. At 15-year follow-up, women with DM/pre-DM were 21 kg (46 pounds) heavier than at baseline compared with those women who did not develop DM/pre-DM (8 kg or 17.5 pounds heavier).
Age at delivery and current age, baseline BMI, current smoking status, and participation in aerobic exercise in the postpartum were all related to the presence of CHD/pre-CHD. The multivariable analysis is found in Table 2. For every 1-year increase in age, the rate of CHD/pre-CHD increased 11%. For every 1-kg weight increase over 15 years, the rate of CHD/pre-CHD increased by 5%. Current smokers were 2.5 times more likely to have CHD/pre-CHD, and former smokers were 1.85 times more likely to have CHD/pre-CHD than women who never smoked. Finally, women who were overweight or obese at the index pregnancy were 1.7 or 4.9 times more likely to have CHD/pre-CHD than were women who were underweight or normal weight at baseline.
Weight gain and BMI at each follow-up measure for women who, at 15 years, had developed CHD/pre-CHD were compared with those of women who had not (Fig. 2). The patterns (slopes) of weight change and BMI over time were significantly different from each other (P < .001), showing significantly greater weight gain and higher BMIs at all follow-up measures for those women developing CHD/pre-CHD. At 15-year follow-up, women with CHD/pre-CHD were, on average, 14 kg (31 pounds) heavier than at baseline, whereas those women who did not develop CHD/pre-CHD were an average of 8 kg (17.5 pounds) heavier than at baseline.
The intent of this study was to determine the long-term impact of pregnancy weight gain on obesity and the development of obesity-related conditions. Our study continues to demonstrate that women who gain excessively during pregnancy, fail to lose pregnancy weight in an appreciable time, or are overweight or obese before pregnancy are at a significant risk of excess weight gain and obesity during midlife. Our study also examined the impact of pregnancy-related factors on disease. Baseline BMI was the only independent factor that was significantly related to incident disease from those available from the index pregnancy. Change in weight over the 15 years was indirectly related to disease, in that women who gained excessively during pregnancy or failed to lose pregnancy weight by 6 months postpartum were at increased risk of excess weight gain 15 years later and, consequently, at a higher risk of developing DM/pre-DM or CHD/pre-CHD. Willett et al20 found that, after 14 years of follow-up, even modest weight gains increased the risk of coronary heart disease.
It is conceivable that, the longer the follow-up, the greater the possibility that the impact of pregnancy and birth on development of disease will be diluted by numerous other lifestyle factors. However, we found that the ability to control weight during pregnancy was correlated to the ability to control after pregnancy. Although we have no additional information about the lifestyles of the women in our study after 6-months postpartum, those women who were participating in aerobic exercise during the postpartum were less likely to have excess weight gain and be obese 15 years later. Participation in aerobic exercise (not running, swimming, biking, or walking) was likely an indicator of a fitness-oriented woman. Although smoking status at pregnancy was not related to disease risk, current smoking status was significantly related to the development of CHD/pre-CHD.
A 15-year follow-up to a pregnancy study in Sweden conducted by Linne et al15 in 2002 found that weight gain during pregnancy was a better predictor of long-term weight change than initial body weight at the beginning of their study. They suggest that pregnancy and weight development are enmeshed in a complex pattern that includes a change in lifestyle factors such as eating behavior, exercise, and smoking. These lifestyle factors are all generally accepted as risk factors for chronic diseases such as diabetes and heart disease. The Linne study included only 26% of the original cohort of 1,200 women. Our study included 61% of the original cohort and found both weight gain during pregnancy and initial BMI to be predictors of long-term weight change.
Several studies have shown a link between parity and obesity or obesity-related diseases.21,22 Weng et al21 found a 7% increase in the rate of obesity with each additional child, adjusting for age, race, income, work status, physical activity, and tobacco and alcohol use. Lawlor et al22 reported a 30% increased odds of CHD for every additional child in the home for women, but also found a 12% increase for men. Adjustment for obesity and metabolic risk factors wiped out the relationship, although the relationship between children and BMI and waist-to-hip ratio was significant. Our study failed to show a relationship between parity and BMI, change in weight, or obesity-related illnesses.
Sakurai23 reported on the findings of 13 studies that examined the effect of duration of obesity on diabetes development. Inconsistencies in data made a comparison difficult, but Sakurai concluded that the occurrence of diabetes is “probably associated with duration of obesity.” Although our study cannot address this association directly, all women who developed diabetes or prediabetes were disease-free at baseline. While date of development of diabetes or heart disease was not recorded for our study, Figure 2 illustrates the change in BMI over time and suggests that, as a group, those women with DM/pre-DM or CHD/pre-CHD have been obese for at least 10–13 years, and women obese at baseline were much more likely to develop diabetes and heart disease by 15-year follow-up.
Our study is limited, in that the role of ethnicity or socioeconomic status in the development of disease cannot be explored because of the small variability of the cohort. Our population included only white, predominately middle-class women. Many studies have demonstrated a higher risk of excess gestational weight gain and retained weight in minority groups, including black and Hispanic women and women of lower socioeconomic status.24,25
We classified obesity based on BMIs calculated from heights and weights measured in the physicians’ offices. Some research indicates that coronary heart disease or all-cause mortality may be more strongly related to abdominal obesity than to overall obesity.26,27 Waist circumference, however, is not typically measured in the physician office and was unavailable for our study. We also have no reliable information on physical activity or nutrition—either over time or at study’s end.
Our study found a strong relationship between baseline BMI and weight gain during pregnancy to current BMI. Baseline for our study was in 1988. Disconcerting are findings that prepregnancy BMI and weight gain during pregnancy are likely much higher today than 15 years ago.28,29 This could significantly amplify the findings from our study and also impact the prevalence of obesity-related conditions in the future.
An additional concern in interpreting our study is that our results are based on 61% of the original cohort. Those lost to follow-up may be at an even higher risk of obesity, given that they presented for prenatal care later in their pregnancy. Half of those lost to follow-up still live within our service area but are not receiving primary care from our facility. This could indicate less access to primary care overall and increased risk of poorer health. The implication for our findings would be that we may be underestimating the impact of pregnancy weight change on long-term obesity and disease risk.
The Institute of Medicine recommendations were developed based on optimizing fetal growth, as well as reducing prematurity and perinatal mortality.19 Little evidence exists of a relationship between lower maternal weight gain and adverse events. Thus, given the impact of excess weight gain on long-term obesity, the guidelines should be reviewed again. Health care providers need to be aggressive about their recommendations regarding weight gain during pregnancy and should refer overweight or obese women to weight counseling both during and after pregnancy. Some studies have found that weight gain during pregnancy can be safely and effectively controlled by health care providers and patients together, and this may be needed to slow the rate of obesity and lessen the risk of future chronic illness.30
1. Colditz GA, Willett WC, Rotnitzky A, Manson JE. Weight gain as a risk factor for clinical diabetes mellitus in women. Ann Intern Med 1995;122:481–6.
2. Huang Z, Willett WC, Manson JE, Rosner B, Stampfer M, Speizer F, et al. Body weight, weight change and hypertension in women. Ann Intern Med 1998;128:81–8.
3. Lapidus L, Bengtsson C, Larsson B, Pennert K, Rybo E, Sjostrom L. Distribution of adipose tissue and risk of cardiovascular disease and death: a 12-year follow up of participations in the population study of women in Gothenburg, Sweden. Br Med J (Clin Res Ed) 1984;289:1257–61.
4. Must A, Spadano J, Coakley EH, Field AE, Colditz G, Dietz WH. The disease burden associated with overweight and obesity. JAMA 1999;282:1523–9.
5. Sweeney C, Blair CK, Anderson KE, Lazovich D, Folsom AR. Risk factors for breast cancer in elderly women. Am J Epidemiol 2004;160:868–75.
6. Manson JE, Willett WC, Stampfer MJ, Colditz GA, Hunter DJ, Hankinson SE, et al. Body weight and mortality among women. N Engl J Med 1995;333:677–85.
7. Calle EE, Rodriguez C, Walker-Thurmond K, Thun MJ. Overweight, obesity, and mortality from cancer in a prospectively studied cohort of U.S. adults. N Engl J Med 2003;348:1625–38.
8. Strawbridge WJ, Wallhagen MI, Shema SJ. New NHLBI clinical guidelines for obesity and overweight: will they promote health? Am J Public Health 2000;90:340–3.
9. Flegal KM, Graubard BI, Williamson DF, Gail MH. Excess deaths associated with underweight, overweight, and obesity. JAMA 2005;293:1861–7.
10. Flegal KM, Carroll MD, Ogden CL, Johnson CL. Prevalence and trends in obesity among US adults, 1999–2000. JAMA 2002;288:1723–7.
11. Freedman DS, Khan LK, Serdula MK, Galuska DA, Dietz WH. Trends and correlates of class 3 obesity in the United States from 1990 through 2000. JAMA 2002;288:1758–61.
12. Williamson DF, Kahn HS, Remington PL, Anda RF. The 10-year incidence of overweight and major weight gain in US adults. Arch Intern Med 1990;150:665–72.
13. Mokdad AH, Serdula MK, Dietz WH, Bowman BA, Marks JS, Koplan JP. The spread of the obesity epidemic in the United States, 1991–1998. JAMA 1999;282:1519–22.
14. Bradley PJ. Conditions recalled to have been associated with weight gain in adulthood. Appetite 1985;6:235–41.
15. Linne Y, Barkeling B, Rossner S. Long-term weight development after pregnancy. Obes Rev 2002;3:75–83.
16. Schauberger CW, Rooney BL, Brimer LM. Factors that influence weight loss in the puerperium. Obstet Gynecol 1992;79:424–9.
17. Rooney BL, Schauberger CW. Excess pregnancy weight gain and long-term obesity: one decade later. Obstet Gynecol 2002;100:245–52.
18. Grundy SM, Cleeman JI, Merz CN, Brewer HB, Clark LT, Hunninghake DB, et al. Implications of recent clinical trials for the National Cholesterol Education Program Adult Treatment Panel III guidelines. Circulation 2004;110:227–39.
19. Institute of Medicine. Nutrition During Pregnancy: Part I, Weight Gain. Washington, DC: National Academy Press; 1990. p. 10.
20. Willett WC, Manson JE, Stampfer MJ, Colditz GA, Rosner B, Speizer FE, et al. Weight, weight change, and coronary heart disease in women: risk within the “normal” weight range. JAMA 1995;273:461–5.
21. Weng HH, Bastian LA, Taylor DH, Moser BK, Ostbye T. Number of children associated with obesity in middle-aged women and men: results from the health and retirement study. J Womens Health 2004;13:85–91.
22. Lawlor DA, Emberson JR, Ebrahim S, Whincup PH, Wannamethee SG, Walker M, et al. Is the association between parity and coronary heart disease due to biological effects of pregnancy or adverse lifestyle risk factors associated with child-rearing? Findings from the British Women’s Heart and Health Study and the British Regional Heart Study. Circulation 2003;107:1260–4.
23. Sakurai Y. Duration of obesity and risk of non-insulin-dependent diabetes mellitus. Biomed Pharmacother 2000;54:80–4.
24. Gore SA, Brown DM, West DS. The role of postpartum weight retention in obesity among women: a review of the evidence. Ann Behav Med 2003;26:149–59.
25. Walker LO. Weight gain after childbirth: a women’s health concern? Ann Behav Med 1995;17:132–41.
26. Alexander JK. Obesity and coronary heart disease. Am J Med Sci 2001;321:215–24.
27. Folsom AR, Kaye SA, Sellers TA, Hong CP, Cerhan JR, Potter JD, et al. Body fat distribution and 5-year risk of death in older women [published erratum in JAMA 1993;269:1254]. JAMA 1993;269:483–7.
28. Kinnunen TI, Luoto R, Gissler M, Hemminki E. Pregnancy weight gain from 1960s to 2000 in Finland. Int J Obes Relat Metab Disord 2003;27:1572–7.
29. Schieve LA, Cogswell ME, Scanlon KS. Trends in pregnancy weight gain within and outside ranges recommended by the Institute of Medicine in a WIC population. Matern Child Health J 1998;2:111–6.
30. Cogswell ME, Scanlon KS, Fein SB, Schieve LA. Medically advised, mother’s personal target, and actual weight gain during pregnancy. Obstet Gynecol 1999;94:616–22.