Objective: The objective was to assess the relevance of the recommendations of the Institute of Medicine (IOM), regarding gestational weight gain (GWG) for long-term BMI development.
Research Methods and Procedures: The Stockholm Pregnancy and Women's Nutrition is a follow-up study of 483 women who delivered children in 1984 to 1985. ANOVA was used to examine the change in body weight before pregnancy, at 6 months, and 1 year postpartum and 15 years after childbirth. Multiple linear regression was used to assess the predictors of BMI at 15-year follow-up.
Results: The weight increase from baseline to 15-year follow-up was 6.2 kg for IOM-insufficient, 6.7 kg for IOM-recommended, and 10.0 kg for IOM-excessive weight gain (p < 0.01). ANOVA showed a main effect of time, group and group by time interaction. The weight of the women who had excessive GWG was significantly greater at each time-point of follow-up than the weight of those who gained within or below recommendations. GWG was related to BMI at 15-year follow-up even after accounting for several confounders. Women who gained excessive weight during pregnancy had an increase of 0.72 kg/m2 in long-term BMI compared with women who gained within recommendations.
Discussion: The findings support the adequateness of IOM guidelines, not only for the pregnancy-related health matters, but also for preventing long-term weight retention after delivery. Healthcare providers should give women appropriate advice for controlling GWG and motivate them to lose pregnancy-related weight during postpartum to prevent future overweight.
Weight increases gradually during adult life (1), but for women, pregnancy can significantly alter the future weight gain trajectory (2). A large body of literature has supported the appropriateness of gestational weight gain (GWG)1 guidelines for positive pregnancy outcomes, suggested by the Institute of Medicine (IOM) (3). Studies have shown that GWG above the range recommended by IOM is associated with postpartum weight retention (4)(5)(6). In general, the studies have assessed postpartum weight retention at 6 months or 1 year after delivery. Only two studies have appraised the relationship between GWG and long-term weight development (7)(8).
In the first study, Rooney and Schauberger estimated the impact of excess pregnancy weight gain on weight retention and BMI at 8 to 10 years follow-up after controlling for weight loss by 6 months postpartum. However, the analysis was not adjusted for initial BM (7). In the second study, women were followed, on average, for 14.7 years (range, 10.1 to 16.3 years). The authors overcame the limitation of the initial study by controlling the analysis for pre-pregnancy BMI. They concluded that excessive gestational weight gain and failure to lose pregnancy-related weight by 6 months postpartum constitute important predictors of obesity in midlife (8). Despite the contribution of their findings, more detailed information on maternal weight changes after pregnancy is warranted. It is also important to consider that the postpartum period may comprise up to the first 12 months after delivery because many physiological changes due to pregnancy remain up to 1 year (9). After delivery, women may experience many social and behavioral changes (10)(11)(12), and 6 months might be a too short period to return to pre-pregnancy weight.
The most important limitations of analyzing the role of excessive GWG on long-term weight development are the time interval between the weight measurements and the wide variation of weight changes during the follow-up period. It is paramount to determine if maternal weight at long-term follow-up represents pregnancy weight retention or weight cycling, a regain of weight after an initial loss.
The objective of the present study is to explore the effect of GWG according to IOM recommendations on long-term BMI, accounting for several potentially confounding factors, including postpartum weight changes and pre-pregnancy BMI.
Research Methods and Procedures
The Stockholm Pregnancy and Women's Nutrition (SPAWN) is a longitudinal study of women who delivered children in 1984 to 1985 in 14 maternity units in Stockholm, Sweden. In the maternity unit, the staff invited the women to take part in the study at the first control visit after delivery. Up to that point, this study was retrospective in that information about weight development during pregnancy was collected from obstetrics records. Women were then prospectively monitored up to 1 year postpartum and 15 years later. Detailed information about the yearlong study is described elsewhere (13)(14)(15). The study comprised the follow-up period from delivery (1984 to 1985) to 15 years postpartum (1999 to 2000).
The final population was composed of 483 women. The initial population comprised 2342 women. After exclusions and dropouts, 1423 subjects completed the yearlong study. Fifteen years after childbirth, women were tracked by using the national identity number. A total of 269 subjects could not be found. Thus, invitation letters were sent to 1154 subjects, and 563 of these women agreed to participate. Of these, 80 women were excluded due to preterm delivery and lack of information about weight measurements during pregnancy and at follow-up occasions.
A dropout analysis, based on the total sample (n = 2342), showed no significant differences between completers and non-completers in relation to the circumstance of pregnancy, such as fecundity history, high blood pressure, edema, delivery methods, dietary advice during pregnancy, and social factors. However, the completers were older (29.4 vs. 30.0 years) and non-response was more common among non-Nordic citizens. More information about the background of SPAWN and the dropout analyses are published in related articles (16)(17)(18)(19).
Definition of Variables
The main outcome variable was BMI at 15-year follow-up and was used as a continuous variable. The major independent variable was total GWG categorized as insufficient, recommended, and excessive weight gain according to IOM guidelines, which are based on pre-pregnancy BMI (20). The IOM suggests that women of normal BMI before pregnancy should aim for a weight gain between 11.5 and 16.0 kg. A slightly higher and lower target range is recommended for underweight and overweight women, respectively. In our study, overweight and obese women were collapsed into one category, because few women were obese and there is no recommended upper limit of GWG for them. Self-reported pre-pregnancy weight (PPW) at first prenatal visit was used to calculate pre-pregnancy BMI. Women were classified according to IOM definitions as: underweight (<19.8), normal (19.8−26.0), and overweight/obese (≥26.1) (20).
Seven different variables were created to explore the postpartum weight changes. The variables are shown in Table 1. Duration and intensity of lactation were expressed in scores (0 to 48 points). Every month with full lactation was given 4 points and every month with mixed feeding was given 2 points (15). This score was used as a rough estimate of the total energy expenditure for milk production, but has been suggested to be a reasonable estimate (13)(16). The lactation score was explored as a continuous and categorical variable (0 to 9, 10 to 19, 20 to 29, ≥30). Physical activity during leisure time was defined as inactive, 4 to 6 hours of light activity/wk and regular exercise (e.g., jogging, gymnastics). Women were asked about frequency of breakfast, lunch, and snacks; frequency of cooked meals; breakfast quality; and self-perceived change of meal sizes. More details about the food habits questionnaire can be found elsewhere (14). Questions concerning physical activity, eating, and smoking habits were answered for five study periods: before and during pregnancy and 6 months and 1 and 15 years after.
Table 1. Definition of postpartum weight variables
|Δ weight at 6 mo postpartum||Weight at 6 mo postpartum − PPW|
|Δ weight at 1 yr postpartum||Weight at 1 yr postpartum − PPW|
|Δ weight from 6 mo to 1 yr postpartum||Weight at 1 yr − weight at 6 mo|
|Δ weight at 15 yr||Weight at 15 yr postpartum − PPW|
|Weight change at 6 mo postpartum*||Derived from Δ weight values at 6 mo postpartum|
| Below PPW||<−0.5kg|
| Returned to PPW||Between − 0.5 and 0.5 kg|
| Above PPW||>0.5 kg|
|Weight change at 1 yr postpartum*||Derived from Δ weight values at 1 yr postpartum|
| Below PPW||<−0.5 kg|
| Returned to PPW||Between−0.5 and 0.5 kg|
| Above PPW||>0.5 kg|
|Weight changes up to 1 yr postpartum|| |
| Below PPW weight at 6 mo and 1 yr||Below PPW at both time points|
| Below at 6 and returned to PPW at 1 yr||Below PPW weight at 6 mo and able to return to their PPW by 1 yr postpartum|
| Kept PPW at 6 mo and 1 yr||Able to return to PPW at 6 mo and maintained their weight up to 1 yr|
| Above at 6 mo and returned to PPW at 1 yr||Above PPW at 6 mo and able to return to their PPW at 1 yr|
| Kept over PPW at 6 mo and 1 yr||Above PPW at both time points|
The analyses were performed using SPSS 11.0 for Windows (SPSS Inc., Chicago, IL) and Stata 9.0 (StataCorp LP., College Station, TX). A mixed ANOVA was applied to examine the change in weight at the four time-points of assessment (before pregnancy, 6 months, 1 and 15 years after pregnancy; within-subjects factor). Between-subjects factor was based on GWG according to IOM guidelines. Furthermore, the association between IOM recommended GWG and BMI at 15-year follow-up was assessed by multiple linear regression. The first procedure in the analysis was to screen all potential predictors individually for relationship to the dependent variable. Variables that showed p value ≤0.25 in the univariate analysis were selected. The selected variables were gradually introduced in the model in hierarchical causal position: first, the distal variables; second, the intermediate variables; and last, the proximal variables. To appraise if pre-pregnancy BMI moderates the relationship between GWG and long-term BMI, three different models were fitted. Model I included all significant predictors without considering pre-pregnancy BMI. Model II was similar to Model I except for the inclusion of initial BMI as a continuous variable. Model III accounted for pre-pregnancy BMI as a component of the outcome variable, which was no longer long-term BMI, but the difference between BMI at 15-year follow-up and pre-pregnancy BMI − Δ long-term BMI.
The study was approved by the Ethics Committee of the Karolinska University Hospital, and all participants gave written informed consent.
Characteristics of Women
The socio-demographic and other characteristics of the sample are shown in Table 2. The mean (± standard deviation) age at baseline was 30.0 ± 4.6 years, and 52% of the women had no child before the index pregnancy. The mean gestational length, total GWG, and birth weight were 39.8 ± 1.3 weeks, 14.2 ± 4.1 kg, and 3506 ± 470 g, respectively. The pre-pregnancy BMI was 21.5 ± 2.4 kg/m2.
Table 2. General characteristics of the sample
|During pregnancy|| || |
| Elementary or high school||256||53.6|
|Civil status||477|| |
| Married/live with someone||449||94.1|
|Prepregnancy BMI||483|| |
| Normal weight||355||73.5|
|IOM GWG groups||483|| |
|At 15 yr follow-up|| || |
| <40 yr||48||9.9|
| 40–50 yr||349||72.3|
| ≥50 yr||86||17.8|
|Lactation score||483|| |
|Total parity||472|| |
| 1 child||50||10.6|
| 2–3 children||378||80.1|
| ≥4 children||44||9.3|
Postpartum Weight Retention
The mean postpartum weight retention was 1.3 ± 3.5 kg and 0.5 ± 1.3 kg at 6-month and 1-year follow-up, respectively. Thus, women lost a further 0.9 ± 2.2 kg from 6 months to 1 year postpartum. Although the average weight loss was modest, the weight change between 6 months and 1 year postpartum ranged markedly from −8.9 to 7.9 kg. After 6 months 15.1% and after 1 year 33.3% of the women had returned to PPW. All women who returned to PPW at 6 months were able to maintain this weight up to 1-year postpartum (Table 3).
Table 3. Postpartum weight changes at 6 and 12 months of follow-up
|At 6 mo postpartum||483|| |
| Below PPW||119||24.6|
| Returned to PPW||73||15.1|
| Above PPW||291||60.2|
|At 1 yr postpartum||483|| |
| Below PPW||87||18.0|
| Returned to PPW||161||33.3|
| Above PPW||235||48.7|
|Up to 1 yr postpartum||483|| |
| Below PPW weight at 6 mo and 1 yr||87||18.0|
| Below at 6 mo and returned to PPW at 1 yr||32||6.6|
| Kept PPW at 6 mo and 1 yr||73||15.1|
| Above at 6 mo and returned to PPW at 1 yr||56||11.6|
| Kept over PPW at 6 mo and 1 yr||235||48.2|
The 15-year weight increase was, on average, 7.6 ± 7.5 kg. Table 4 presents the unadjusted mean of long-term weight retention and BMI by subject characteristics. Age, civil status during pregnancy, income at 15-year follow-up, parity, and physical activity at 6 months and 1 year postpartum were not significantly associated with weight development (data not shown).
Table 4. Maternal long-term weight retention and BMI by subject characteristics
|Education*|| ||3.00|| ||2.67|
| Elementary/high school||8.6 ± 8.2||0.001||24.7 ± 4.0||0.008|
| University||6.5 ± 6.7|| ||23.8 ± 3.2|| |
|Prepregnancy BMI|| ||0.34|| ||87.93|
| Underweight||7.3 ± 6.0||0.715||21.5 ± 2.4||0.000|
| Normal weight||7.8 ± 7.6|| ||24.8 ± 3.2|| |
| Overweight/obese||6.6 ± 12.5|| ||30.5 ± 5.1|| |
|IOM GWG groups|| ||13.35|| ||22.59|
| Insufficient||6.2 ± 6.8||0.000||23.5 ± 3.7||0.000|
| Recommended||6.7 ± 6.8|| ||23.6 ± 3.0|| |
| Excessive||10.3 ± 8.5|| ||25.9 ± 3.9|| |
|Lactation score|| ||5.10|| ||4.93|
| 0–9||10.2 ± 8.6||0.002||25.6 ± 4.7||0.002|
| 10–19||9.2 ± 8.0|| ||25.0 ± 4.0|| |
| 20–29||6.9 ± 7.1|| ||23.8 ± 3.4|| |
| ≥30||6.6 ± 7.1|| ||24.0 ± 3.2|| |
|Physical activity at 15 yr postpartum|| ||1.41|| ||5.29|
| Inactive||8.2 ± 8.8||0.246||25.4 ± 4.1||0.005|
| 4–6 hours of light activity/wk||7.9 ± 7.4|| ||24.3 ± 3.7|| |
| Regular activity||6.7 ± 6.8|| ||23.8 ± 3.2|| |
|Weight change at 6 mo postpartum|| ||15.93|| ||9.56|
| Below PPW||4.8 ± 6.4||0.000||23.6 ± 3.5||0.000|
| Returned to PPW||6.3 ± 6.6|| ||23.2 ± 3.3|| |
| Above PPW||9.1 ± 7.8|| ||24.9 ± 3.7|| |
|Weight change at 1 yr postpartum|| ||14.26|| ||9.93|
| Below PPW||5.0 ± 6.7||0.000||23.8 ± 3.5||0.000|
| Returned to PPW||6.6 ± 6.7|| ||23.5 ± 3.5|| |
| Above PPW||9.4 ± 7.9|| ||25.0 ± 3.7|| |
|Weight changes up to 1 yr postpartum|| ||8.32|| ||5.82|
| Below PPW weight at 6 mo and 1 yr||4.9 ± 6.7||0.000||23.8 ± 3.5||0.000|
| Below at 6 mo and returned to PPW at 1 yr||4.7 ± 5.6|| ||22.9 ± 3.6|| |
| Kept PPW at 6 mo and 1 yr||6.3 ± 6.6|| ||23.2 ± 3.3|| |
| Above at 6 mo and returned to PPW at 1 yr||8.0 ± 7.3|| ||24.2 ± 3.6|| |
| Kept over PPW at 6 mo and 1 yr||9.4 ± 7.9|| ||25.0 ± 3.7|| |
Educational level was significantly associated with weight retention. Less educated women retained 2 kg more than women who had a university degree. Women with a higher lactation score retained less weight at the three follow-up periods. The p values of ANOVA for Δ weight at 6 months, 1 and 15 years and the lactation score were 0.049, 0.039, and 0.002, respectively (data not shown). Women who gained excessive weight during pregnancy had the greater weight increase (10.3 kg) at long-term follow-up and retained 4.1 kg and 3.6 kg more at 15-year follow-up than women who gained less or recommended weight, respectively (p = 0.00).
Women who were above their PPW at 6 months or 1 year postpartum retained more weight at 15-year follow-up than women who were able to lose pregnancy-related weight. Women who returned to PPW at 1 year retained 0.3 kg more than those who returned to PPW at 6 months, but the difference was not significant (p = 0.38). Likewise, although women who were above PPW at 6 months but were able to return to their PPW at 1 year postpartum retained 1.7 kg more than women who returned to PPW at 6 months, the difference was not statically significant (p = 0.08).
The results regarding long-term BMI were similar, with the exception of pre-pregnancy BMI and physical activity at 15-year follow-up, which showed significant effect: women who were inactive and overweight/obese before pregnancy had higher long-term BMI than active and normal or underweight women.
Weight Development After Index Pregnancy
A mixed ANOVA with one repeated measures factor, time (weight before pregnancy, 6 months, 1 and 15 years after), and one between-subjects factor (insufficient, recommended, and excessive GWG) showed a main effect of time [F(9.024) = 113.728, p = 0.000] and a significant time group interaction [F(6.129) = 77.237, p = 0.000]. The weight of the women who gained excessive weight during pregnancy was significantly greater at each time-point [main effect of group: F(10.550) = 870.023 p = 0.000]. The analysis was adjusted for pre-pregnancy BMI.
The mean BMI increased significantly from 21.5 ± 2.4 before gestation to 24.3 ± 3.7 kg/m2 at 15-year follow-up. Fifty-two percent of women remained within the normal range at both time-points, 21.7% of under or normal weight women became overweight/obese by the end of follow-up, 1.2% of initially overweight/obese subjects became normal weight, and 3% of women were overweight/obese at both time-points.
In the multivariate regression model, GWG was a significant predictor of BMI at long-term follow-up, even when pre-pregnancy BMI was accounted for as a covariate in Model II or as part of the outcome variable (Δ long-term BMI) in Model III. When adjusting for pre-pregnancy BMI, the effect of GWG on long-term BMI was greatly attenuated but still significant. The Models I and II showed that women who gained excessive weight had an increase of 1.81 and 0.72 kg/m2, respectively, in long-term BMI compared with women who gained within recommendations. The adjusted R2 increased significantly from Model I to Model II. The main final model (Model II) was statistically significant and explained 51.3% of the variance in long-term BMI (Table 5).
Table 5. Multivariate regression of predictor factors of long-term BMI and Δ long-term BMI
|IOM GWG groups|| ||0.000|| ||0.033|| ||0.042|
| Insufficient||−0.05 (−0.81, 0.72)|| ||0.01 (−0.56, 0.59)|| ||0.02 (−0.56, 0.59)|| |
| Recommended||Reference|| ||Reference|| ||Reference|| |
| Excessive||1.81 (1.07, 2.57)|| ||0.72 (0.15, 1.30)|| ||0.68 (0.11, 1.24)|| |
|Education|| ||0.028|| ||0.014|| ||0.015|
| Elementary or high school||0.70 (0.76, 1.33)|| ||0.59 (0.12, 1.06)|| ||0.59 (0.11, 1.06)|| |
| University||Reference|| ||Reference|| ||Reference|| |
|Lactation score||−0.03 (−0.06, −0.01)||0.046||−0.03 (−0.06, −0.01)||0.010||−0.03 (−0.05, −0.01)||0.010|
|Δ weight at 6 mo postpartum||3.29 (1.35, 5.23)||0.001||2.12 (0.64, 3.57)||0.005||2.06 (0.60, 3.52)||0.006|
|Δ weight from 6 mo to 1 yr postpartum||4.87 (1.85, 7.89)||0.002||2.96 (0.68, 5.24)||0.011||2.89 (0.61, 5.16)||0.013|
|Prepregnancy BMI|| || ||0.96 (0.86, 1.06)||0.000|| || |
|Adjusted R2||14.0|| ||51.3|| ||14.3|| |
Women who gained weight above IOM recommendations and had a low educational level had a significantly higher long-term BMI than women who gained within or below recommendations and had a university degree. A higher lactation score was associated with lower long-term BMI. Greater initial BMI, postpartum retention at 1 year, and weight gain between 6 months and 1 year postpartum were associated with greater long-term BMI.
Other potential confounding variables were tested, but they were not retained in the final model. Age, civil status, nationality, smoking habits, and physical activity at 6 months postpartum and frequency of breakfast and lunch at 15-year follow-up were not significantly associated with long-term BMI in the univariate analyses (p >0.25). Parity, hypertension during pregnancy, frequency of snacks at 1 year postpartum, and physical activity at 1 and 15 years postpartum were significant in the univariate analyses (p ≤0.25); however, they did not show significant effect in the multivariate model. No interaction effect was found between variables.
This study shows that excess weight gain during pregnancy is related to higher BMI at 15-year follow-up even after controlling for education, lactation score, pre-pregnancy BMI, weight retention at 6 months postpartum, and weight gain between 6 months and 1 year postpartum. The analysis found that the inclusion of pre-pregnancy BMI in the Model II attenuates the association between GWG and long-term BMI and reduced the coefficient regression from Model I to Model II by more than one-half. Although the GWG was not the major predictor in multivariate model, it still was significantly associated with long-term BMI.
It is argued that total number of children and subsequent gestations after the index pregnancy may have an impact on long-term BMI. However, this effect was not verified in this study. Corroborating our findings, Rooney and Schauberger did not find a significant effect of parity in the multivariate model predicting long-term BMI (7)(8).
In addition, the results revealed that high-intensity and frequency of breastfeeding is associated with less postpartum weight retention on all follow-up occasions. However, lactation seems to have a more significant effect at long-term follow-up than short-term. Rooney and Schauberger have found similar results (7). Their study revealed that, although breastfeeding may not have an impact on short-term weight loss, women who breast-fed their infants for at least 3 months had a significantly lower weight gain at 8-year follow-up (7). They also found that short breastfeeding duration is related to high BMI at 15-year follow-up (8). It is possible that women with intensive and lengthy lactation have a high health consciousness and dietary habits that facilitates the long-term weight control, rather than the early energy losses caused by milk production.
Although in the previous publications of the yearlong study Ohlin and Rossner found a weak association between trends in eating and exercising patterns and weight retention (13)(14), these behavioral aspects seem have no relationship with long-term BMI. Boardley et al. demonstrated that caloric intake and postpartum physical activity were not significantly associated with weight change over 12 months postpartum in a multivariate model (21). In contrast, Rooney and Schauberger found that participation in aerobic exercise has a nonsignificant effect on short-term weight loss but has a significant impact on long-term BMI (7). Hence, it remains controversial whether lifestyle factors during pregnancy and postpartum period play a role in long-term weight development.
Several methodological problems are involved in the present study, which limited the appreciation of real effect of diet and physical activity. Data were based on reported, rather than recorded, behaviors. Women are prone to under-report some important behaviors related to weight gain (e.g., alcohol and sweets consumption) (22)(23). In addition, the crude questionnaires could not measure such small changes in everyday lifestyle, which still have significant impact on long-term weight development (18).
The other limitation of this study is the use of self-reported PPW. Although women are prone to underestimate their PPW, the validation study showed that the information was reasonable and probably had a minor impact on pre-pregnancy BMI classification (13). Thirty-three percent of the sample lived too far away from the clinic to attend in person and weight and height at 15 years were self-reported, rather than objectively measured. However, there were no differences between those who sent in forms and those who visited our unit with respect to pre-pregnancy BMI, weight gain during pregnancy, infant weight, gestational length, weight at 1-year follow-up, age, and BMI at follow-up (18). The self-reported data were furthermore compared with national self-report data and found to be representative (1).
It has been argued that total GWG, defined by subtracting pre-pregnancy from the weight in late pregnancy, is highly influenced by the interval between the last measurement and the date of delivery. In Sweden, all women have their weight measured and recorded at the maternity unit admission. This procedure ensures that the women did not experience additional weight gain that was not accounted for.
The availability of weight at 6 months and 1 year postpartum enabled us to assess the weight cycling during this period. It is worth noting that some women may continue losing pregnancy weight after 1-year follow-up. However, the data on weight change after 1 year postpartum were unavailable. Weight measurements collected in between 1 and 15-year follow-up would help to better understand the maternal weight cycling.
In this study, it would be possible to assess the long-term weight retention using Δ weight at 15 years in a multivariate model. However, GWG and Δ weight are overlapping variables since they were derived from the same baseline measurements (PPW). The use of overlapping variables results in part-whole bias (24)(25).
An additional concern in interpreting our study is the large number of dropouts. However, the extensive dropout analyses indicated that, for most variables under study, the remaining women were representative for the initial sample (16)(17)(18)(19). Nonetheless, the prevalence of overweight/obesity at baseline in this study was lower than the national prevalence (1). It seems that the higher risk group (overweight/obese women) was not recruited in the follow-up study, which probably underestimated our findings.
The prevalence of obesity in Sweden is lower than prevalences seen in many other Western countries. Meanwhile, it is worrying the increasing trend of obesity before and during pregnancy in Sweden (1). Thus, pre-pregnancy BMI is probably much higher nowadays than in 1984 to 1985. Furthermore, the weight gain data were recorded before the release of IOM guidelines. By that time, there were no formal recommendations for gestational weight gain in Sweden. In light of these particularities, it is unknown if these results could be generalized to other populations.
To our knowledge, the SPAWN is the longest follow-up study regarding weight development after childbirth, in which the whole sample was monitored for 15 years. This is a unique study that evaluated the impact of excessive GWG on long-term weight development accounting for two important factors, such as pre-pregnancy BMI and postpartum weight changes.
In summary, IOM recommendations, designed to ensure a healthy pregnancy outcome, also seem appropriate for preventing the later development of maternal overweight. The findings indicated that women who gained excess weight during pregnancy by IOM standards have a higher BMI at 15-year follow-up than women who gained within or below IOM recommendations, even after accounting for several confounding factors. More attention should be given to a health policy concerning effective strategies for controlling GWG and preventing postpartum weight retention. It seems advisable that women should lose pregnancy-related weight gradually during the postpartum period and reach their PPW within 1 year. The effects of lifestyle factors during and after pregnancy on long-term weight development warrant further studies.
The authors thank Agneta Öhlin, who was responsible for the data collection until 1990. This study was supported by Brazilian Foundation for training of researchers in Doctoral Exchange Programe (CAPES) (to A.R.A.) and by Arbeitsmarknadens Forsakrings-och Aktiebolag (AFA) (to M.N.).