Obesity

Original Article CLINICAL TRIALS: BEHAVIOR, PHARMACOTHERAPY, DEVICES, SURGERY

Changes in Food Choice During a Successful Weight Loss Trial in Overweight and Obese Postpartum Women Ena Huseinovic, Anna Winkvist, Fredrik Bertz, and Hilde Kristin Brekke

Objective: To examine changes in intake across food groups during a weight loss trial that produced significant and sustainable weight loss in lactating women receiving dietary treatment. Methods: At 10-14 wk postpartum, 61 overweight and obese lactating Swedish women were randomized to a 12-wk dietary (D), exercise (E), combined (DE), or control (C) treatment. Food intake was assessed by 4-d weighed diet records which were used to examine changes in intake across seven food groups from baseline to 12 wk and 1 y after randomization. Differences in changes in food choice between women receiving dietary treatment (D1DE) and no dietary treatment (E1C) were examined using multivariate linear regression. Results: At baseline, sweets and salty snacks contributed to 21610 percent of total energy intake (E%). During the intervention period, women receiving dietary treatment reduced their E% from sweets and salty snacks and caloric drinks and increased their E% from vegetables more than did women not receiving dietary treatment (all P < 0.010). At 1 y, the increased E% from vegetables was maintained significantly higher among women receiving dietary treatment (P 5 0.002). Conclusions: Lactating women receiving dietary treatment achieved sustainable weight loss through changes in food choice in line with current dietary guidelines. Obesity (2014) 22, 2517–2523. doi:10.1002/oby.20895

Introduction Pregnancy has been identified as a significant contributor to the development of excess weight among women (1,2). During pregnancy many women experience excessive weight gain which is retained postpartum, and this weight development is particularly prominent among women entering pregnancy overweight or obese (3,4). Among overweight and obese women 63% and 46%, respectively, gain more pregnancy weight than recommended (5) and approximately 25% of all women experience a weight retention of 4.55 kg or more 1 y postpartum (6). Thus, for many women postpartum weight retention contributes to increased weight with each reproductive cycle, thereby increasing the risk of maternal and neonatal complications during subsequent pregnancies (7,8) and providing important implications for long-term maternal health (9,10). The postpartum period may be a time window when women are motivated to undertake lifestyle changes to lose the extra weight gained during pregnancy (11). Previous studies have reported poor dietary intake among postpartum women (12-15) which indicate the value and potential of dietary changes during this period. Epidemiological as well as experimental studies suggest that weight loss efforts may be most effec-

tive when particular foods and beverages are targeted for increased or decreased consumption (16-19). Unfortunately, evidence for food choices that facilitate weight loss following pregnancy is limited and few studies report changes in food choice made by those successful in achieving weight loss through diet interventions (20-22). In the recent LEVA (Lifestyle for effective weight loss during lactation) trial, overweight and obese lactating women received a 12-wk diet intervention at 10-14 wk postpartum and achieved significant weight loss of 7.663.6 kg (9%) that was sustained at 8.566.3 kg (10%) at the 1-y follow-up. Weight loss among women not receiving the dietary treatment was 1.563.2 kg (2%) and 1.766.3 kg (2%) after 12 wk and 1 y, respectively (P < 0.001 between groups at both times) (23). In this report, we aim to examine changes in food choice achieved in the LEVA-trial. First, we describe the food choices among all women at baseline. Thereafter we evaluate changes in food choice during the intervention period among women receiving dietary treatment and compare these to changes in food choice among women not receiving the 12-wk dietary treatment. Finally, we report on changes in food choice achieved among the two groups at the 1-y follow-up.

Department of Internal Medicine and Clinical Nutrition, The Sahlgrenska Academy, University of Gothenburg, Gothenburg, Sweden. Correspondence: Ena Huseinovic ([email protected]) Funding agencies: The study was supported by grants from the Swedish Research Council (K2009-70X-21091-01-03), the Swedish Council for Working Life and Social Research (2006-0339), and the Swedish Nutrition Foundation. Disclosure: The authors declared no conflict of interest. Author contributions: AW, HKB, and FB conceived and conducted the LEVA-trial; EH analyzed the data; all authors were involved in writing the paper. Additional Supporting Information may be found in the online version of this article. Received: 24 June 2014; Accepted: 18 August 2014; Published online 19 September 2014. doi:10.1002/oby.20895

www.obesityjournal.org

Obesity | VOLUME 22 | NUMBER 12 | DECEMBER 2014

2517

Obesity

Food Choices and Weight Loss Postpartum Huseinovic et al.

Methods Study design and subjects The LEVA-trial was a randomized controlled factorial trial among overweight and obese lactating Swedish women conducted 20072011. It aimed to investigate the separate and interactive effects of a 12-wk diet and exercise behavior modification treatment on body weight and body composition in both the short (12 wk) and long (1 y) term. The design, methods, and primary outcome have been described in detail previously (23,24). In short, 68 women with a self-reported prepregnancy body mass index (BMI) of 25-35 kg/m2 and the intention to breastfeed for at least 6 months were recruited during pregnancy or up to 8 wk postpartum in Gothenburg, Sweden. At 10-14 wk postpartum, the women were randomly assigned to four study groups; dietary behavior modification group (D), physical exercise behavior modification group (E), dietary and physical exercise behavior modification group (DE), or control group (C). Study measurements were conducted prior to randomization (i.e., 8-12 wk postpartum, indicated as baseline), at intervention termination (indicated as 12 wk) and 9 months post-treatment (indicated as 1 y). The study was approved by the regional ethical committee in Gothenburg and informed consent was obtained from all participants.

Intervention groups Women randomized to the diet group (D) received 2.5 h of individual dietary behavior modification counseling by a dietitian; 1.5 h at the start of the intervention and 1 h at a follow-up home visit after 6 wk of intervention. The women were instructed to implement a 12-wk dietary modification plan in order to achieve an energy reduction of 500 kcal/d and a nutrient composition according to the Nordic Nutrition Recommendations 2004 (25). The diet plan was communicated in terms of foods and consisted of four key dietary principles to be implemented one at a time, according to a step-wise body weight-determined plan in order to achieve the weekly weight loss goal of 0.5 kg and the final weight loss goal of 6 kg at 12 wk. Women were instructed to self-weigh 3 times/wk and to use body weight as a proxy for energy balance in order to adjust the energy intake during the intervention by a step-wise introduction of the key dietary principles. The four key dietary principles to be introduced were: (1) limit sweets, salty snacks, and caloric drinks to 100 g/wk, to be consumed during one day of the wk, (2) substitute regular foods with low-fat and/or low-sugar alternatives marked with the “Green keyhole”, a labeling symbol provided by the Swedish National Food Administration indicating foods that contain less sugar, fat, and salt and more whole grain and fiber, (3) cover onehalf of the plate with vegetables at lunch and dinner by applying the “Plate model”, an illustration of the proportions between the meal components, and (4) reduce portion sizes. The women were provided with suggestions on concrete changes in food choice to the reported baseline diet in accordance with the four key dietary principles and calculations on the weekly and total weight loss that would be achieved if the principles were implemented. During the second dietary counseling section at 6 wk, the diet plan was followed up and feedback was provided on the current diet. Women in the exercise group (E) received an individual physical exercise modification plan by a physical therapist. The goal of the exercise intervention was to implement a 45-min brisk walk 4 d/wk at 60-70% of maximum heart rate. Women in the combined diet and exercise group (DE) received both diet and exercise intervention for

2518

Obesity | VOLUME 22 | NUMBER 12 | DECEMBER 2014

a total of 5 h of individual counseling. The control group (C) did not receive any intervention (i.e., usual care) and were asked not to engage in any other lifestyle modification program during the intervention period.

Measurements Body weight was determined after an overnight fast to the nearest 0.1 kg by using an electronic scale (MC 180 MA; Tanita, Tokyo, Japan), with women wearing light underclothing. Height was measured by using a wall-mounted stadiometer. Body composition was measured at all three time points using dual-energy X-ray absorptiometry (DXA) (Lunar Prodigy; GE Lunar). Total energy expenditure (TEE) was measured at baseline and 12 wk using the doubly labeled water method (DLW) (26).

Food assessment and analysis Food intake was assessed using a weighed diet record during four consecutive days, preferably Wednesday through Saturday. Women were provided with an electronic kitchen scale (HR2395, Philips, Drachten, Netherlands) and instructed to register all foods and beverages in as much detail as possible and to weigh the amounts to the nearest one gram. Women were told not to divert from their usual food choices or habits at baseline as their diet record would be used to construct the diet plan if they were randomized to the dietary treatment. In this report, the diet records were used to examine changes in food choice from baseline to 12 wk and 1 y. Food items were manually categorized into seven major food groups mainly on the basis of underlying hypotheses related to the key dietary principles of the diet intervention; [1] sweets and salty snacks, [2] caloric drinks, [3] fruit, [4] vegetables, [5] potatoes/pasta/bread, [6] meat and meat products, and [7] dairy (Supporting Information Table S1). In addition, the energy density of dairy products (calculated as kcal/g) was assessed to further evaluate the dietary recommendation to substitute regular foods with low-fat and low-sugar alternatives. Energy intake and quantity consumed from each food group are expressed as mean intake during the four registered days and were calculated using the software Dietist XP (version 3.2, Kost och N€aringsdata, Bromma, Sweden), based on the Swedish Food Database 2010 and data from food manufacturers. The percentage of total energy intake (E%) contributed by each food group was calculated as the food group energy intake divided by the total energy intake. In addition, the proportion of women reaching an intake of 500 g fruit and vegetables/d, a recommendation set by the Swedish National Food Administration, was assessed. This recommendation includes fruits, berries, juice (maximum 100 g/d), dried fruits, root vegetables, and legumes. However, consistent with the key dietary principles, juice was categorized as a “caloric drink” in this report.

Statistical analysis Food choices at baseline, 12 wk, and 1 y as well as changes in food choice from baseline to 12 wk and 1 y are presented for women receiving dietary treatment (D1DE, D-groups) and no dietary treatment (E1C, ND-groups). Changes in food choice were calculated as the value obtained at 12 wk and 1 y minus the baseline value. Student’s t-test and Mann Whitney U-test were used to compare continuous variables between groups and paired samples t-test and

www.obesityjournal.org

Original Article

Obesity

CLINICAL TRIALS: BEHAVIOR, PHARMACOTHERAPY, DEVICES, SURGERY

Figure 1 Food choices presented as percentage of total energy intake (E%) at baseline, 12 wk and 1 y for women receiving dietary treatment (D-groups) and women receiving no dietary treatment (ND-groups) in the LEVA-trial. Median values are presented for all food groups.

Wilcoxon signed rank test were used to compare continuous variables within groups. Pearson chi-square and Fischer’s exact test were used for categorical variables. Mean6SD values are presented for normally distributed variables, median (1st; 3rd quartile) values are presented for non-normally distributed variables and proportions are presented for categorical variables. For simplicity, median values are presented in Figure 1. Multiple linear regression was used to examine differences in changes in food choice between D-groups and ND-groups, with changes in food choice as dependent variable (ranks were created for non-normally distributed variables) and group assignment (Dand ND-groups) as independent variable. All regression models were adjusted for baseline intake of the dependent variable and an estimate of the energy intake underreporting at 12 wk and 1 y based on mean energy requirements (kcal/kg/day) corrected for energy cost of lactation and changes in body tissue energy stores as measured by DXA. The calculation of this estimate has been described in detail previously and validated against underreporting calculated using DLW-measured TEE at 12 wk (r50.9; P < 0.001) (24). Data were analyzed using SPSS version 21.0 (IBM, Somers, NY, USA). Statistical significance was considered at P < 0.05.

Results Study subjects Among the 68 randomized women in the LEVA-trial, 61 women completed the diet record at baseline and 12 wk while 54 of these also completed the diet record at 1 y. During the intervention one woman was excluded because of pregnancy and one woman because of prescription of a metabolism-affecting drug while four dropped out for other reasons. In this report, one additional woman was excluded because of a missing diet record at 12 wk. Between 12 wk and 1 y, five women were excluded because of new pregnancies and

www.obesityjournal.org

none dropped out; however, two additional women were excluded in this report because of missing diet records at 1 y. Baseline characteristics for the 61 included women are found in Table 1. The women had a mean age of 33.264.1 y, a prepregnancy BMI of 28.4 (27.1; 30.8) kg/m2, a baseline BMI of 30.162.7 kg/m2 and the majority was highly educated (72%). There were no statistically significant differences in baseline characteristics between Dgroups and ND-groups.

Food choices at baseline The women reported an energy intake of 26556497 kcal/d at baseline with sweets and salty snacks contributing to 21.4610.0 E%. The mean intake of fruit and vegetables was 3386176 g/d and 21% reached the recommended intake of 500 g fruit and vegetables/d. Food choices did not differ between D-groups and ND-groups except for a higher E% from meat and meat products among D-groups (P 5 0.022).

Food choices at 12 wk and changes in food choice during the intervention After the intervention, i.e., at 12 wk, sweets and salty snacks contributed to 11.169.3 and 20.9611.8 E% in D-groups and ND-groups, respectively (P 5 0.001, Figure 1). Further, D-groups had a mean intake of 4736277 g fruit and vegetables/d while ND-groups had an intake of 3206161 g/d (P 5 0.011). The percentage of women meeting the recommended intake of fruit and vegetables was significantly higher in D-groups than in ND-groups (45.2% vs 16.7%, P 5 0.016). During the intervention, D-groups reduced their E% from sweets and salty snacks (P < 0.001), caloric drinks (P 5 0.001) and increased their E% from fruit (P 5 0.039), vegetables (P < 0.001), and potatoes/pasta/bread (P 5 0.010) as compared to baseline. In ND-groups, the E% from meat and meat products was increased during the intervention (P 5 0.006). Further, D-groups reduced their

Obesity | VOLUME 22 | NUMBER 12 | DECEMBER 2014

2519

Obesity

Food Choices and Weight Loss Postpartum Huseinovic et al.

TABLE 1 Baseline characteristics of women receiving dietary treatment (D-groups) and women receiving no dietary treatment (ND-groups) in the LEVA-trial

Variable

D-groups (n 5 31)

ND-groups (n 5 30)

P-value

33.8 6 4.3 1.0 (1.0; 2.0) 28.7 (26.9; 30.1) 30.0 6 2.3 84.6 6 8.6

32.5 6 3.9 1.0 (1.0; 2.0) 28.2 (27.2; 32.6) 30.2 6 3.1 86.7 6 11.0

0.222 0.908 0.634 0.738 0.416

3.2 (1) 22.6 (7) 74.2 (23)

13.3 (4) 16.7 (5) 70.0 (21)

0.331

100.0 (31) 0.0 (0) 0.0 (0)

90.0 (27) 3.3 (1) 6.7 (2)

0.196

96.8 (30) 3.2 (1)

93.3 (28) 6.7 (2)

0.612

Agea (y) Parityb (n) Prepregnancy BMIb (kg/m2) BMI at baselinea (kg/m2) Weight at baselinea (kg) Education % (n) Short education at high school  3 y beyond high school > 3 y beyond high school Marital status % (n) Married or cohabitating Couple with different residences Single Lactation status % (n) Exclusive Partial a

Presented as mean 6 SD. Presented as median (first; third quartiles).

b

E% from sweets and salty snacks and caloric drinks and increased their E% from vegetables more than did ND-groups (all P 0.010, Table 2). In addition, the energy density of dairy products decreased more in D-groups than in ND-groups (P 5 0.016).

Food choices at 1 y and changes in food choice during the 1-y follow-up At 1 y, sweets and salty snacks contributed to 13.8611.3 and 18.3611.1 E% in D-groups and ND-groups respectively (P 5 0.146, Figure 1). Further, D-groups had a mean intake of 3556165 g fruit and vegetables/d while ND-groups had an intake of 2936140 g/d (P 5 0.139). The difference in the percentage of women meeting the recommended intake of 500 g/d was not significant (18.5% vs 7.4%, P 5 0.420). Compared to baseline, women in D-groups reduced their E% from sweets and salty snacks (P 5 0.011) and increased their E% from vegetables (P < 0.001) and potatoes/pasta/bread (P 5 0.016) at 1 y. Among ND-groups, women reported a reduced E% from sweets and salty snacks (P 5 0.011) and an increased E% from meat and meat products (P 5 0.014). Further, D-groups reported a greater increase in the E% from vegetables as compared to ND-groups (P 5 0.002, Table 2). There was no change in E% from caloric drinks between baseline and 1 y; however, when alcoholic drinks were excluded from the food group, a significant reduction emerged within both groups (P < 0.001 vs P 5 0.009). Still, the nonsignificant difference between groups remained (P 5 0.615).

Discussion We examined changes in food choice among overweight and obese lactating women receiving dietary treatment that led to clinically

2520

Obesity | VOLUME 22 | NUMBER 12 | DECEMBER 2014

relevant and sustainable weight loss in the LEVA-trial. We found that the energy contribution from sweets and salty snacks and caloric drinks decreased and that the energy contribution from vegetables increased more among those receiving dietary treatment than among those not receiving it during the intervention period. At the 1-y follow-up, the increased energy contribution from vegetables was maintained significantly higher among the women that had received dietary treatment. The recently updated Nordic Nutrition Recommendations 2012 emphasize an increased intake of vegetables, legumes, fruits and low-fat dairy and a decreased intake of foods and beverages with added sugars to prevent diet-related chronic disease (27). However, previous studies have reported low compliance to dietary guidelines among postpartum women (12-15,28). Durham et al. reported substantial amounts of refined grains, sweetened beverages, and desserts and a fruit and vegetable intake below recommendations among 450 overweight and obese US women at 2 months postpartum (29). We found that sweets and salty snacks contributed to 21 E% and that only one in five women met the recommended intake of 500 g fruit and vegetables (juice excluded) at baseline. This high energy contribution from energy-dense and nutrient-poor foods corresponds to previous reports from the LEVA-trial where the intake of sucrose and saturated fat were above and intake of fiber, folate, vitamin D, and iron were below recommended levels at baseline (24). Further, since the majority of women in the LEVA-trial were exclusively breastfeeding, well-educated and relatively physically active at baseline, which are all factors associated with healthy eating habits (14,30), the overall diet quality might be lower in a more diverse population. Thus, these results suggest that overweight and obese women should be encouraged to increase fruits, vegetables, low-fat dairy, and whole grain and to decrease sweets and salty snacks to improve diet quality during lactation.

www.obesityjournal.org

Original Article

Obesity

CLINICAL TRIALS: BEHAVIOR, PHARMACOTHERAPY, DEVICES, SURGERY

TABLE 2 Food choices at baseline and changes in food choice from baseline to 12 wk (D 12 wk) and 1 y (D 1 y) for women

receiving dietary treatment (D-groups) and women receiving no dietary treatment (ND-groups) in the LEVA-trial D-groups (n 5 31)

Mean 6 SD Sweets and salty snacks (E%) Baseline D 12 wk D 1 yc Caloric drinks (E%) Baseline* D 12 wk* D 1 yc* Fruit (E%) Baseline D 12 wk D 1 yc Vegetables (E%) Baseline* D 12 wk* D 1 yc* Potatoes/pasta/bread (E%) Baseline D 12 wk D 1 yc Meat and meat products (E%) Baseline D 12 wk D 1 yc Dairy (E%) Baseline D 12 wk D 1 yc

Median (1st; 3rd quartile)

ND-groups (n 5 30) Pa Within group difference

20.3 6 10.0 17.8 (12.9; 25.7) 29.2 6 11.2 210.3 (216.5; 20.6) 27.3 6 13.8 28.3 (217.1; 1.9)

Changes in food choice during a successful weight loss trial in overweight and obese postpartum women.

To examine changes in intake across food groups during a weight loss trial that produced significant and sustainable weight loss in lactating women re...
202KB Sizes 0 Downloads 7 Views