Advertisement

Associations between Snacking and Weight Loss and Nutrient Intake among Postmenopausal Overweight to Obese Women in a Dietary Weight-Loss Intervention

      Abstract

      Snacking may play a role in weight control. The associations of timing and frequency of snacking with observed weight change and nutrient intake were assessed in an ancillary study to a 12-month randomized controlled trial in Seattle, WA. Overweight-to-obese postmenopausal women (n=123) enrolled in the two dietary weight-loss arms from 2007 to 2008 with complete data at 12 months were included in these analyses. Generalized linear models were used to test the associations between snacking and weight loss (percent) and nutrient intake at the 12-month time point. Participants were, on average, 58 years old and mainly non-Hispanic white (84%). Ninety-seven percent reported one or more snacks per day. Weight loss (percent) was significantly lower among mid-morning (10:30 am to 11:29 am) snackers (7.0%, 95% confidence interval: 4.3 to 9.7) compared to non–mid-morning snackers (11.4%, 95% confidence interval: 10.2 to 12.6; P=0.005). A higher proportion of mid-morning snackers reported more than one snack per day (95.7%), compared to afternoon (82.8%) and evening (80.6%) snackers, although differences were not statistically significant (P>0.05). Women who reported two or more snacks per day vs one or no snacks per day had higher fiber intake (P=0.027). Afternoon snackers had higher fruit and vegetable intake compared to non–afternoon-snackers (P=0.035). These results suggest that snack meals can be a source for additional fruits, vegetables, and fiber-rich foods; however, snacking patterns might also reflect unhealthy eating habits and impede weight-loss progress. Future dietary weight-loss interventions should evaluate the effects of timing, frequency, and quality of snacks on weight loss.
      Meal pattern characteristics such as frequency, timing (eg, breakfast, late-night eating), or type (eg, main meals, snacks) may play a role in weight control. Positive associations have been observed between snacking frequency and energy intake (
      • Piernas C.
      • Popkin B.M.
      Snacking Increased among US adults between 1977 and 2006.
      ,
      • Berteus Forslund H.
      • Torgerson J.S.
      • Sjostrom L.
      • Lindroos A.K.
      Snacking frequency in relation to energy intake and food choices in obese men and women compared to a reference population.
      ,
      • Kerver J.M.
      • Yang E.J.
      • Obayashi S.
      • Bianchi L.
      • Song W.O.
      Meal and snack patterns are associated with dietary intake of energy and nutrients in US adults.
      ,
      • Berteus Forslund H.
      • Lindroos A.K.
      • Sjostrom L.
      • Lissner L.
      Meal patterns and obesity in Swedish women—A simple instrument describing usual meal types, frequency and temporal distribution.
      ,
      • Popkin B.M.
      • Duffey K.J.
      Does hunger and satiety drive eating anymore? Increasing eating occasions and decreasing time between eating occasions in the United States.
      ), suggesting snacking could contribute to weight gain. However, the relationship of snacking with weight is inconclusive (
      • Berteus Forslund H.
      • Lindroos A.K.
      • Sjostrom L.
      • Lissner L.
      Meal patterns and obesity in Swedish women—A simple instrument describing usual meal types, frequency and temporal distribution.
      ,
      • Bes-Rastrollo M.
      • Sanchez-Villegas A.
      • Basterra-Gortari F.J.
      • Nunez-Cordoba J.M.
      • Toledo E.
      • Serrano-Martinez M.
      Prospective study of self-reported usual snacking and weight gain in a Mediterranean cohort: The SUN project.
      ,
      • Berteus Forslund H.
      • Klingstrom S.
      • Hagberg H.
      • Londahl M.
      • Torgerson J.S.
      • Lindroos A.K.
      Should snacks be recommended in obesity treatment? A 1-year randomized clinical trial.
      ,
      • Kant A.K.
      • Schatzkin A.
      • Graubard B.I.
      • Ballard-Barbash R.
      Frequency of eating occasions and weight change in the NHANES I Epidemiologic Follow-up Study.
      ,
      • Hampl J.S.
      • Heaton C.L.B.
      • Taylor C.A.
      Snacking patterns influence energy and nutrient intakes but not body mass index.
      ,
      • Ma Y.
      • Bertone E.R.
      • Stanek 3rd, E.J.
      • Reed G.W.
      • Hebert J.R.
      • Cohen N.L.
      • Merriam P.A.
      • Ockene I.S.
      Association between eating patterns and obesity in a free-living US adult population.
      ). In a recent review, 16 of 20 studies reported inverse associations between eating frequency and weight (
      • McCrory M.A.
      • Howarth N.C.
      • Roberts S.B.
      • Huang T.T.-K.
      Eating frequency and energy regulation in free-living adults consuming self-selected diets.
      ). However, if under-reporting of energy intake is not appropriately accounted for, snack intake may be underestimated (
      • McCrory M.A.
      • Howarth N.C.
      • Roberts S.B.
      • Huang T.T.-K.
      Eating frequency and energy regulation in free-living adults consuming self-selected diets.
      ,
      • Bellisle F.
      Impact of the daily meal pattern on energy balance.
      ,
      • Huang T.T.
      • Roberts S.B.
      • Howarth N.C.
      • McCrory M.A.
      Effect of screening out implausible energy intake reports on relationships between diet and BMI.
      ). A review of controlled feeding studies concluded that snacks offered little to no benefit to overall regulation of food intake (
      • Leidy H.J.
      • Campbell W.W.
      The effect of eating frequency on appetite control and food intake: Brief synopsis of controlled feeding Studies.
      ). Furthermore, a prospective analysis of National Health and Nutrition Examination Survey data detected no association of eating frequency with weight change (
      • Kant A.K.
      • Schatzkin A.
      • Graubard B.I.
      • Ballard-Barbash R.
      Frequency of eating occasions and weight change in the NHANES I Epidemiologic Follow-up Study.
      ). Conflicting findings may be a result of the lack of universal terms to define meals and snacks (
      • Berteus Forslund H.
      • Lindroos A.K.
      • Sjostrom L.
      • Lissner L.
      Meal patterns and obesity in Swedish women—A simple instrument describing usual meal types, frequency and temporal distribution.
      ,
      • Chapelot D.
      The role of snacking in energy balance: A biobehavioral approach.
      ) and may also vary by the setting and characteristics of the population studied (
      • Hartline-Grafton H.L.
      • Rose D.
      • Johnson C.C.
      • Rice J.C.
      • Webber L.S.
      The influence of weekday eating patterns on energy intake and BMI among female elementary school personnel.
      ).
      The present study examined the relationships of between-meal snacking and snacking frequency with weight loss (percent) and nutrient intake using a simple, reproducible meal pattern grid to assess snacking patterns (
      • Berteus Forslund H.
      • Lindroos A.K.
      • Sjostrom L.
      • Lissner L.
      Meal patterns and obesity in Swedish women—A simple instrument describing usual meal types, frequency and temporal distribution.
      ) among postmenopausal overweight-to-obese women in a year-long dietary weight-loss intervention. Little is known about snacking frequency and weight change in this population, a group at high risk for obesity and for chronic diseases from obesity, including diabetes, cardiovascular disease, and certain cancers such as breast, colon, and endometrium (
      • Dennis K.E.
      Postmenopausal women and the health consequences of obesity.
      ,
      Physical Activity Guidelines Advisory Committee
      Physical Activity Guidelines Advisory Committee Report, 2008.
      ). It was hypothesized that snacking would be positively related to weight loss, fiber, fruit and vegetable intake, and inversely related to fat intake.

      Methods

       Participants

      Participants in this ancillary study were part of a larger, four-arm randomized controlled trial, testing the effects of nutrition and exercise-based interventions in breast cancer biomarkers and body composition (

      Foster-Schubert KE, Alfano CM, Duggan CR, Xiao L, Campbell KL, Kong A, Bain CE, Wang C-Y, Blackburn GL, McTiernan A. Effect of diet and exercise, alone or combined, on weight and body composition in overweight-to-obese postmenopausal women [published online ahead of print April 14, 2011]. Obesity. doi:10.1038/oby2011.76.

      ). The four arms were diet-induced weight loss (Diet); aerobic exercise; both interventions combined (Diet+Exercise); and control. The design of the parent trial was detailed previously (

      Foster-Schubert KE, Alfano CM, Duggan CR, Xiao L, Campbell KL, Kong A, Bain CE, Wang C-Y, Blackburn GL, McTiernan A. Effect of diet and exercise, alone or combined, on weight and body composition in overweight-to-obese postmenopausal women [published online ahead of print April 14, 2011]. Obesity. doi:10.1038/oby2011.76.

      ). Women enrolled in either the Diet or Diet+Exercise arms from June 2007 to August 2008 formed the Weight and Eating Behaviors cohort used for the current analyses. Women who completed the parent trial (n=92) prior to this ancillary study were not included in the Weight and Eating Behaviors cohort. The Fred Hutchinson Cancer Research Center Institutional Review Board approved all study procedures and study participants provided written informed consent.

       Lifestyle-Based Interventions

      A detailed description of the lifestyle-based interventions has been described elsewhere (

      Foster-Schubert KE, Alfano CM, Duggan CR, Xiao L, Campbell KL, Kong A, Bain CE, Wang C-Y, Blackburn GL, McTiernan A. Effect of diet and exercise, alone or combined, on weight and body composition in overweight-to-obese postmenopausal women [published online ahead of print April 14, 2011]. Obesity. doi:10.1038/oby2011.76.

      ). Briefly, the goal of the exercise intervention was ≥45 minutes of moderate-to-vigorous intensity aerobic exercise, 5 days per week for 12 months. Participants attended three or more sessions per week at the study facility, supervised by an exercise physiologist, and exercised the remaining sessions at home.
      The design and curriculum of the diet intervention were based on the Look AHEAD (Action for Health in Diabetes) and Diabetes Prevention Program studies (
      Look Ahead Research Group
      Look AHEAD (Action for Health in Diabetes): Design and methods for a clinical trial of weight loss for the prevention of cardiovascular disease in type 2 diabetes.
      ,
      Diabetes Prevention Program Research Group
      The Diabetes Prevention Program (DPP): Description of lifestyle intervention.
      ), with the following goals: total intake of 1,200 to 2,000 kcal/day based on baseline weight, <30% calories from fat, and 10% reduction in weight by 6 months with maintenance to 12 months. Although separate instruction groups were held for women in the Diet and Diet+Exercise groups to reduce contamination; registered dietitians (RDs) with training in behavior modification delivered the same curriculum to both groups. The curriculum covered topics such as reducing fat and improving fiber intake, self-monitoring of weight and food intake, goal setting, and problem solving. Women were asked to maintain a daily food journal for 6 months or until they reached their weight-loss goal (10%). No specific recommendations were made concerning snacking behavior. Women met individually with an RD on at least two occasions, followed by weekly group meetings, for up to 6 months. Thereafter, women met with an RD at least semi-monthly (eg, one in-person, plus e-mail or phone contact), where self-monitoring of weight and food intake were still encouraged and sessions focused on staying motivated and dealing with barriers and lapses. However, those struggling with initial or maintenance of weight loss received additional RD assistance.

       Demographic and Anthropometric Variables

      Self-reported information on age, race/ethnicity, marital status, and education level were collected as baseline measures. Anthropometric measurements were collected at baseline and at 12 months and performed with the participant in a hospital gown. Trained technicians obtained height and weight using a balance beam scale (DETECTO, Web City, MO) and stadiometer (Perspective Enterprises, Portage, MI), rounding up to the nearest 0.1 cm and 0.5 kg, respectively.

       Meal Pattern Grid

      Meal pattern intake was measured at 12 months, using a self-administered, meal assessment grid used to describe typical frequency, types, and temporal distribution of meals. The instrument, developed by Berteus-Forslund and colleagues to assess meal patterns in obese and normal weight individuals in a Swedish population, demonstrated reasonable reliability (r=0.70) (
      • Berteus Forslund H.
      • Lindroos A.K.
      • Sjostrom L.
      • Lissner L.
      Meal patterns and obesity in Swedish women—A simple instrument describing usual meal types, frequency and temporal distribution.
      ) and ability to distinguish meal patterns between lean and obese adults (
      • Berteus Forslund H.
      • Torgerson J.S.
      • Sjostrom L.
      • Lindroos A.K.
      Snacking frequency in relation to energy intake and food choices in obese men and women compared to a reference population.
      ,
      • Berteus Forslund H.
      • Lindroos A.K.
      • Sjostrom L.
      • Lissner L.
      Meal patterns and obesity in Swedish women—A simple instrument describing usual meal types, frequency and temporal distribution.
      ). This grid is currently being evaluated against 24-hour recalls in two collaborating work-site studies in the United States (S.A.A. Beresford, personal communication, April 15, 2011).
      To complete the grid, the women indicated when they ate or drank during a typical 24-hour weekday, noting the time and type of meal (ie, main, light/breakfast, snack, beverage only). Participants were instructed to mark one box per line. If a respondent marked a meal and beverage box, the meal box was coded as the default. The grid provided examples of foods common to each meal type, but did not provide standard definitions. The only modification made to the original grid was to include foods familiar to an American audience (Figure 1).
      Figure thumbnail gr1
      Figure 1Meal pattern grid administered at 12 months to postmenopausal women participating in a dietary weight loss intervention.

       Dietary Intake

      Percent calories from fat, fiber (g/day), and fruit and vegetable intake (servings/day) were estimated at 12 months using the Women's Health Initiative food frequency questionnaire (
      • Patterson R.E.
      • Kristal A.R.
      • Tinker L.F.
      • Carter R.A.
      • Bolton M.P.
      • Agurs-Collins T.
      Measurement characteristics of the Women's Health Initiative food frequency questionnaire.
      ). The Women's Health Initiative food frequency questionnaire reasonably compares to food records and 24-hour recalls (mean intake based on 8 days worth of entries), with correlation coefficients for percent energy from fat of 0.62 and 0.70 for fiber (
      • Patterson R.E.
      • Kristal A.R.
      • Tinker L.F.
      • Carter R.A.
      • Bolton M.P.
      • Agurs-Collins T.
      Measurement characteristics of the Women's Health Initiative food frequency questionnaire.
      ). Women who reported values outside of the plausible range of energy intake (<600 kcal and >3,500 kcal) (
      • Patterson R.E.
      • Kristal A.R.
      • Tinker L.F.
      • Carter R.A.
      • Bolton M.P.
      • Agurs-Collins T.
      Measurement characteristics of the Women's Health Initiative food frequency questionnaire.
      ) were excluded from the main analysis (n=5).

       Statistical Analyses

      Distribution of main, light, and snack meals during the course of a typical 24-hour period was examined to establish the time periods for main meals (ie, breakfast, lunch, and dinner) and snacking (ie, mid-morning, afternoon, evening). The time periods were breakfast (5:00 am to 10:29 am), mid-morning snack (10:30 am to 11:29 am), lunch (11:30 am to 1:59 pm), afternoon snack (2:00 pm to 5:29 pm), dinner (5:30 pm to 8:59 pm), and evening snack (9:00 pm to 12:00 am). Frequencies for main, light, snack, and beverage-only meals were tabulated. In this study, a snack was identified as a “snack meal” on the meal pattern grid; and occurred between main meals. A cross tabulation between timing and frequency of snacking was conducted to further examine snacking patterns.
      Generalized linear models were used to examine associations at 12 months for snacking (ie, timing and frequency) with weight loss (%) and nutrient intake. A minimum difference of 4.2% in weight change (standard deviation [SD]=7%, two group samples of n=30 and 93) could be detected at 80% power (type 1 error 0.05). The models were also adjusted for baseline age, body mass index, study arm (Diet, Diet+Exercise), race/ethnicity, and food frequency questionnaire–estimated alcohol, sweetened beverage intake, and nutrient values, as well as for grid-estimated total number of main and light meals and snack occasions at other time periods. Fiber (g/day), percent calories from fat, and fruit and vegetable intake (servings/day) were assessed because they were emphasized during the dietary weight-loss intervention. Due to the skewed distribution of fiber, total fruit and vegetable intake, and total calories, these variables were log-transformed and geometric means were reported. All statistical tests were two-sided with an α<.05 and all analyses were performed using STATA (version 11.1, 2010, STATA Corp, College Station, TX).

      Results and Discussion

      Baseline characteristics are presented in Table 1. Women were on average 58 years old and had a mean body mass index of 31.3. Eighty-four percent were non-Hispanic white, 58% were married or living with a partner, 67% had at least a college degree, and 59% were employed full-time. There were no significant differences in demographic characteristics between the Diet (n=59) vs Diet+Exercise (n=64) arms. No significant differences were observed between those enrolled (n=143) vs not enrolled (n=92) in the Weight and Eating Behaviors cohort; however, noncompleters (ie, enrolled in Weight and Eating Behaviors cohort, but did not complete 12-month measures) (n=20) had significantly higher mean body mass indexes than completers (n=123) (data not shown).
      Table 1Baseline demographic and anthropometric characteristics of postmenopausal women participating in a year-long dietary weight-loss intervention and the percent weight loss experienced at 6 and 12 months, stratified by intervention arm
      P value >0.05; no differences were found by group for any of the characteristics or percent weight change.
      Total (n=123)Diet (n=59)Diet+Exercise (n=64)
      Demographic characteristicsmean±SD
      SD=standard deviation; comparisons were made between Diet and Diet+Exercise groups.
      Age (y)58.0±5.157.7±6.058.2±4.2
      n (%)
      Race/ethnicity
      White, non-Hispanic103 (83.7)49 (83.1)54 (84.4)
      Non-white20 (16.3)10 (16.9)10 (15.6)
      Education
      College graduate or more82 (66.7)36 (61.0)46 (71.9)
      Some college or less41 (33.3)23 (39.0)18 (28.1)
      Employment (n=109)
      Not employed15 (13.8)9 (14.1)6 (9.4)
      Part-time30 (27.5)14 (23.7)16 (25.0)
      Full-time64 (58.7)29 (49.2)35 (54.7)
      Marital status (n=122)
      Presently married/living with partner71.0 (58.2)25 (42.4)26 (41.3)
      Single/divorced/widowed51.0 (41.8)34 (57.6)37 (58.7)
      Body mass index4.3 (31.3)3.8 (30.8)4.7 (31.6)
      mean±SD
      % Weight loss at 6 mo
      (Baseline weight−6 months weight)/baseline weight.
      8.8±5.58.3±6.29.2±4.8
      % Weight loss at 12 mo
      (Baseline weight−12 month weight)/baseline weight.
      10.7±7.19.6±7.711.6±6.5
      a P value >0.05; no differences were found by group for any of the characteristics or percent weight change.
      b SD=standard deviation; comparisons were made between Diet and Diet+Exercise groups.
      c (Baseline weight−6 months weight)/baseline weight.
      d (Baseline weight−12 month weight)/baseline weight.
      Participants had a mean percent weight change of 8.8% (SD=5.5) at 6 months and 10.7% (SD=7.1) at 12 months (Table 1). No statistically significant differences were observed by study arm (Diet vs Diet+Exercise) at either time point. In the present study's subsample of 123 women, approximately half met the 10% weight-loss goal by 6 months, with no difference between Diet and Diet+Exercise arms, and 75% were actively trying to lose weight between months 6 and 12 (either because they had not met the goal by 6 months or because they opted to lose >10% of baseline weight) (data not shown).
      Participants reported a mean of 6 (SD=1.29) meals/day, including beverage-only occasions: 1.4 (SD=0.5) main meals/day, 1.5 (SD=0.6) light meals/day, 2.1 (SD=1.0) snacks/day, and 1.00 (SD=0.98) beverage only/day (data not shown). Snacking was highly prevalent: 97% (n=119) reported one or more “snack meal” per day (data not shown). The most common snacking period was in the afternoon: 76% (n=93) of women reported a snack meal in the afternoon period (2:00 pm to 5:29 pm) (Figure 2). Only 19% (n=23) reported a mid-morning (10:30 am to 11:29 am) snack and almost 30% (n=36) reported one or more snack meals after 9 pm (Figure 2). Compared to afternoon (82.8%) and late-evening (80.6%) snackers, more mid-morning snackers (95.7%) reported more than one snack per day (P>0.05) (Figure 2).
      Figure thumbnail gr2
      Figure 2Proportion of weight-loss participants reporting either one snack per day, two snacks per day, or three or more snacks per day within the following time intervals: midmorning (10:30 am to 11:29 am), afternoon (2:00 pm to 5:29 pm), and evening (9:00 pm to 12:00 am). The dark gray bar represents the proportion of participants who reported one snack per day, the light gray bar represents two snacks per day, and the white bar represents three or more snacks per day. χ2 P>0.05; proportions not statistically significantly different.
      Women who reported mid-morning snacking lost significantly less weight (7.0%, 95% confidence interval [CI]: 4.1 to 9.8) compared to non–mid-morning-snackers (11.5%, 95% CI: 10.2 to 12.7; P=0.005) (Table 2). Afternoon snackers reported higher fiber intake (P=0.047) and fruit and vegetable intake (P=0.035) compared to non–afternoon-snackers.
      Table 2Mean comparison of percent weight loss, percent calories from fat, fiber (g/day), and servings of fruit and vegetables between snackers and nonsnackers during between-meal snack periods and by snacking frequency categories among postmenopausal women participating in a year-long dietary weight-loss intervention
      Weight loss reflects percent change from baseline to 12 months; means reported for dietary variables reflect data collected at 12 months. The following dietary variables were log-transformed: fiber (baseline and 12 months), total calories (12 months), total fruits and vegetables (servings/day—at baseline and 12 months).
      Between-meal snacking periodsn
      Sample sizes for nutrient variables (percent fat intake, fiber, fruit/vegetables) and midmorning snacks: no snacks, n=94; ≤1 snack, n=22; afternoon snack: no snack, n=25, ≤1 snack, n=92; evening snack: no snack, n=83; ≤1 snack, n= 34; snacking frequency: ≤1 snack, n=29; 2 snacks, n=48; ≥3 snacks, n=40.
      % Weight Loss
      Adjusted for study arm, baseline body mass index (BMI), number of main and light meals, other snacking periods (except for total snacking frequency model), alcohol, sweetened-beverage intake, and demographic covariates (ie, age, education, marital status, race/ethnicity).
      % Fat Intake
      Adjusted for study arm, baseline BMI, number of main and light meals, other snacking periods (except for total snacking frequency model), and demographic covariates, plus baseline percent fat intake.
      Fiber (g/day)
      Geometric mean reported and adjusted for study arm, baseline BMI, demographic covariates, log of total calories (12 months), log of baseline fiber (g/day), number of main and light meals, and other snacking periods (except for total snacking frequency model).
      Fruits/Vegetables (Servings/Day)
      Geometric mean reported and adjusted for study arm, demographic covariates, log of total calories (12 months), log of baseline fruit and vegetable (servings/day), number of main and light meals, and other snacking periods (except for total snacking frequency model).
      Mean95% CIP value
      No snacks or 1 snack (where appropriate) was used as a reference group, P value based on Wald test and reported whether coefficients were equal to 0.
      Mean95% CIP value
      No snacks or 1 snack (where appropriate) was used as a reference group, P value based on Wald test and reported whether coefficients were equal to 0.
      Mean95% CIP value
      No snacks or 1 snack (where appropriate) was used as a reference group, P value based on Wald test and reported whether coefficients were equal to 0.
      Mean95% CIP value
      No snacks or 1 snack (where appropriate) was used as a reference group, P value based on Wald test and reported whether coefficients were equal to 0.
      Mid-morning: 10:30 am to 11:29 am
      No snacks10011.510.2-12.70.00527.125.7-28.50.7920.519.0-22.20.434.94.4-5.40.07
      ≥1 Snacks237.04.1-9.826.623.6-29.622.218.7-26.26.14.9-7.5
      Afternoon: 2:00 pm to 5:29 pm
      No snacks3012.39.7-15.00.1828.225.3-31.10.3418.115.8-22.10.0474.23.4-5.20.035
      ≥1 Snacks9310.39.0-11.626.625.2-28.021.819.8-23.55.44.9-6.0
      Evening: 9:00 pm to 12:00 am
      No snacks8710.38.8-11.70.3026.625.0-28.10.3220.318.6-22.10.314.94.4-5.50.39
      ≥1Snacks3611.79.5-14.028.125.6-30.522.219.3-25.55.44.7-6.5
      Total snacking frequency
      Total snacking frequency includes individuals who reported snack intake between and during regular meals.
      n
      n=number of women who reported snack meals reported at any time period (includes snack meals reported during regular meal periods).
      ≥1 Snacks/day3311.28.7-13.70.6627.524.9-30.20.4817.115.1-20.3
      2 snacks/day group was significantly different from reference (0 to 1 snacks/day) (P=0.008) and 3 snacks/day group was significantly different from reference group (P=0.002), but not significant for trend.
      0.0274.33.5-5.10.06
      2 Snacks/day5010.08.2-11.927.625.6-29.621.719.8-24.9
      2 snacks/day group was significantly different from reference (0 to 1 snacks/day) (P=0.008) and 3 snacks/day group was significantly different from reference group (P=0.002), but not significant for trend.
      5.24.5-6.0
      ≥3 Snacks4011.19.0-13.226.823.6-28.023.219.7-25.2
      2 snacks/day group was significantly different from reference (0 to 1 snacks/day) (P=0.008) and 3 snacks/day group was significantly different from reference group (P=0.002), but not significant for trend.
      5.74.9-6.7
      a Weight loss reflects percent change from baseline to 12 months; means reported for dietary variables reflect data collected at 12 months. The following dietary variables were log-transformed: fiber (baseline and 12 months), total calories (12 months), total fruits and vegetables (servings/day—at baseline and 12 months).
      b Adjusted for study arm, baseline body mass index (BMI), number of main and light meals, other snacking periods (except for total snacking frequency model), alcohol, sweetened-beverage intake, and demographic covariates (ie, age, education, marital status, race/ethnicity).
      c Adjusted for study arm, baseline BMI, number of main and light meals, other snacking periods (except for total snacking frequency model), and demographic covariates, plus baseline percent fat intake.
      d Geometric mean reported and adjusted for study arm, baseline BMI, demographic covariates, log of total calories (12 months), log of baseline fiber (g/day), number of main and light meals, and other snacking periods (except for total snacking frequency model).
      e Geometric mean reported and adjusted for study arm, demographic covariates, log of total calories (12 months), log of baseline fruit and vegetable (servings/day), number of main and light meals, and other snacking periods (except for total snacking frequency model).
      f Sample sizes for nutrient variables (percent fat intake, fiber, fruit/vegetables) and midmorning snacks: no snacks, n=94; ≤1 snack, n=22; afternoon snack: no snack, n=25, ≤1 snack, n=92; evening snack: no snack, n=83; ≤1 snack, n= 34; snacking frequency: ≤1 snack, n=29; 2 snacks, n=48; ≥3 snacks, n=40.
      g No snacks or 1 snack (where appropriate) was used as a reference group, P value based on Wald test and reported whether coefficients were equal to 0.
      h Total snacking frequency includes individuals who reported snack intake between and during regular meals.
      i n=number of women who reported snack meals reported at any time period (includes snack meals reported during regular meal periods).
      low asterisk 2 snacks/day group was significantly different from reference (0 to 1 snacks/day) (P=0.008) and 3 snacks/day group was significantly different from reference group (P=0.002), but not significant for trend.
      There were no statistically significant differences in percent weight loss, percent fat intake, or fruit and vegetable intake by snacking frequency. Fiber intake was higher among women who reported two snacks per day (21.7 g/day, 95% CI: 19.8 to 24.9) and three snacks per day (23.2 g/day, 95% CI: 19.7 to 25.2) compared to women who reported zero to one snack per day (17.1 g/day, 95% CI: 15.1 to 20.3; P=0.027) (Table 2).
      The prevalence of snacking in this study was comparable to 2003-2006 National Health and Nutrition Examination Survey data, in which 97.3% of adults (65 years of age or older) reported snacking (
      • Zizza C.A.
      • Arsiwalla D.D.
      • Ellison K.J.
      Contribution of snacking to older adults' vitamin, carotenoid, and mineral intakes.
      ). In this study, less weight loss was observed among mid-morning snackers vs non–mid-morning snackers. Snacking frequency was associated with higher intake of fiber and in particular, afternoon snackers had greater intakes of fiber, fruits, and vegetables.
      Previous research on morning-time eating patterns and weight change have only focused on breakfast intake (
      • Raynor H.A.
      • Jeffery R.W.
      • Ruggiero A.M.
      • Clark J.M.
      • Delahanty L.M.
      Weight loss strategies associated with BMI in overweight adults with type 2 diabetes at entry into the Look AHEAD (Action for Health in Diabetes) trial.
      ,
      • Purslow L.R.
      • Sandhu M.S.
      • Forouhi N.
      • Young E.H.
      • Luben R.N.
      • Welch A.A.
      • Khaw K.T.
      • Bingham S.A.
      • Wareham N.J.
      Energy intake at breakfast and weight change: Prospective study of 6,764 middle-aged men and women.
      ), but not on morning-time snack intake. For example, a prospective observational study in 6,874 European adults aged 40 to 75 years reported an inverse association between percent energy intake consumed at breakfast and weight gain (
      • Purslow L.R.
      • Sandhu M.S.
      • Forouhi N.
      • Young E.H.
      • Luben R.N.
      • Welch A.A.
      • Khaw K.T.
      • Bingham S.A.
      • Wareham N.J.
      Energy intake at breakfast and weight change: Prospective study of 6,764 middle-aged men and women.
      ). Observing a lower percent of weight loss in mid-morning snackers, while novel, may not relate necessarily to the time of day, but rather to the short interval between breakfast and lunch. In this study, most mid-morning snackers reported eating both breakfast and lunch, as well as eating more than one snack per day, which suggests that snacking was an additional eating occasion. Recent research has reported an increase in eating occasions among United States adults in the past few decades (
      • Piernas C.
      • Popkin B.M.
      Snacking Increased among US adults between 1977 and 2006.
      ,
      • Popkin B.M.
      • Duffey K.J.
      Does hunger and satiety drive eating anymore? Increasing eating occasions and decreasing time between eating occasions in the United States.
      ). An analysis of National Health and Nutrition Examination Survey data found that the time interval between meals was 1 hour less in the 2003-2006 survey compared with the 1977 survey (
      • Popkin B.M.
      • Duffey K.J.
      Does hunger and satiety drive eating anymore? Increasing eating occasions and decreasing time between eating occasions in the United States.
      ). Research suggests that this may be a reflection of “recreational” or “mindless” eating habits, where individuals are influenced to eat by a variety of situational and external cues (eg, food availability, time of day, other people) other than hunger and satiety (
      • Popkin B.M.
      • Duffey K.J.
      Does hunger and satiety drive eating anymore? Increasing eating occasions and decreasing time between eating occasions in the United States.
      ,
      • Chapelot D.
      The role of snacking in energy balance: A biobehavioral approach.
      ,
      • Wansink B.
      • Payne C.R.
      • Shimizu M.
      “Is this a meal or snack?” Situational cues that drive perceptions.
      ,
      • Vartanian L.R.
      • Herman C.P.
      • Wansink B.
      Are we aware of the external factors that influence our food intake?.
      ). In a cross-over experiment in males between the ages of 20 to 25 years, snacks administered in a non-hungry state did not influence satiety in the following meal; findings were consistent regardless of the nutrient composition of the snack (eg, high carbohydrate or high protein) (
      • Marmonier C.
      • Chapelot D.
      • Fantino M.
      • Louis-Sylvestre J.
      Snacks consumed in a nonhungry state have poor satiating efficiency: Influence of snack composition on substrate utilization and hunger.
      ). This suggests that additional snacking occasions could contribute to excess calories, particularly if they are not consumed in response to internal hunger cues. However, further studies are needed to assess the effect of time between eating occasions on unhealthful eating behaviors and on weight and evaluating the reasons for snacking.
      Marginally higher reported intake of fiber and servings of fruits and vegetables were found among afternoon snackers and higher fiber intake was observed among women who reported greater snacking frequency (two or more snacks per day vs zero or one snack per day). Previous use of the meal pattern grid to examine snacking with fiber intake revealed an inverse (
      • Berteus Forslund H.
      • Torgerson J.S.
      • Sjostrom L.
      • Lindroos A.K.
      Snacking frequency in relation to energy intake and food choices in obese men and women compared to a reference population.
      ) rather than positive association. In the previous study, the relationship was assessed prior to the intervention, which might help explain differences from this study. In addition, in a weight loss intervention context, participants are typically encouraged to choose fruits and vegetables over higher energy-dense options. If snacks are used to incorporate low-energy density, high-fiber foods, eating more snacks may be less likely to contribute to weight gain.
      Using the simple meal pattern grid (
      • Berteus Forslund H.
      • Lindroos A.K.
      • Sjostrom L.
      • Lissner L.
      Meal patterns and obesity in Swedish women—A simple instrument describing usual meal types, frequency and temporal distribution.
      ) to assess meal and snack patterns had strengths and limitations. The grid can distinguish between various meal types, which is helpful, given the inconsistency of the definition of snacking across studies (
      • Piernas C.
      • Popkin B.M.
      Snacking Increased among US adults between 1977 and 2006.
      ,
      • Kerver J.M.
      • Yang E.J.
      • Obayashi S.
      • Bianchi L.
      • Song W.O.
      Meal and snack patterns are associated with dietary intake of energy and nutrients in US adults.
      ,
      • Cross A.T.
      • Babicz D.
      • Cushman L.F.
      Snacking patterns among 1,800 adults and children.
      ). The simplicity of the grid also made it easy for participants to complete without much assistance. The developers of this tool suggest that the risk of under-reporting may be lessened because individuals do not need to recall specific food items to complete the grid (
      • Berteus Forslund H.
      • Lindroos A.K.
      • Sjostrom L.
      • Lissner L.
      Meal patterns and obesity in Swedish women—A simple instrument describing usual meal types, frequency and temporal distribution.
      ). Under-reporting of energy intake is of particular concern among individuals involved in weight-loss interventions for the following reasons: women in behavioral weight-loss studies may under-report dietary intake more at the end of the study (
      • Johnson R.K.
      • Friedman A.B.
      • Harvey-Berino J.
      • Gold B.C.
      • McKenzie D.
      Participation in a behavioral weight-loss program worsens the prevalence and severity of underreporting among obese and overweight women.
      ), less successful dieters are more likely to under-report (
      • Muhlheim L.S.
      • Allison D.B.
      • Heshka S.
      • Heymsfield S.B.
      Do unsuccessful dieters intentionally underreport food intake?.
      ), plus foods of low social desirability (eg, certain snack foods, high-fat foods, sweets, etc.) are often under-reported (
      • Lissner L.
      Measuring food intake in studies of obesity.
      ). A limitation of the grid was that it did not include specific types and amounts of foods and, therefore, the types of foods associated with the main, light, or snacking occasions could not be determined. Although not previously validated against other methods; it is currently being evaluated against 24-hour recalls in two other US populations (S.A.A. Beresford, personal communication, April 15, 2011).

      Conclusions

      In this study, both positive and negative outcomes related to snacking were observed. Snacks can be used to incorporate healthy foods such as fruits and vegetables; however, snacking patterns might also reflect unhealthy eating habits (eg, mindless eating, overeating) and impede weight-loss progress. With the high prevalence of snacking in this study population and among US adults, future dietary weight loss interventions should evaluate of the effects of timing, frequency, and quality of snacks on weight loss. Individuals undergoing dietary weight-loss programs should be educated on ways to healthfully incorporate snacks into the diet.
      STATEMENT OF POTENTIAL CONFLICT OF INTEREST: No potential conflict of interest was reported by the authors.
      FUNDING/SUPPORT: This study was funded by National Cancer Institute (NCI) National Institutes of Health (NIH) grants R01 CA105204-01A1 and U54-CA116847 (Transdisciplinary Research on Energetics and Cancer). K.E.F.-S. received support from NIH 5KL2RR025015-03 and A.K. was supported by NCI R25CA094880 at the time of this study and is currently supported by NCI 2R25CA057699-16 .
      ACKNOWLEDGEMENTS: The authors wish to thank the participants and study staff for their time and dedication to the study.

      References

        • Piernas C.
        • Popkin B.M.
        Snacking Increased among US adults between 1977 and 2006.
        J Nutr. 2010; 140: 325-332
        • Berteus Forslund H.
        • Torgerson J.S.
        • Sjostrom L.
        • Lindroos A.K.
        Snacking frequency in relation to energy intake and food choices in obese men and women compared to a reference population.
        Int J Obes (Lond). 2005; 29: 711-719
        • Kerver J.M.
        • Yang E.J.
        • Obayashi S.
        • Bianchi L.
        • Song W.O.
        Meal and snack patterns are associated with dietary intake of energy and nutrients in US adults.
        J Am Diet Assoc. 2006; 106: 46-53
        • Berteus Forslund H.
        • Lindroos A.K.
        • Sjostrom L.
        • Lissner L.
        Meal patterns and obesity in Swedish women—A simple instrument describing usual meal types, frequency and temporal distribution.
        Eur J Clin Nutr. 2002; 56: 740-747
        • Popkin B.M.
        • Duffey K.J.
        Does hunger and satiety drive eating anymore?.
        Am J Clin Nutr. 2010; 91: 1342-1347
        • Bes-Rastrollo M.
        • Sanchez-Villegas A.
        • Basterra-Gortari F.J.
        • Nunez-Cordoba J.M.
        • Toledo E.
        • Serrano-Martinez M.
        Prospective study of self-reported usual snacking and weight gain in a Mediterranean cohort: The SUN project.
        Clin Nutr. 2010; 29: 323-330
        • Berteus Forslund H.
        • Klingstrom S.
        • Hagberg H.
        • Londahl M.
        • Torgerson J.S.
        • Lindroos A.K.
        Should snacks be recommended in obesity treatment?.
        Eur J Clin Nutr. 2007; 62: 1308-1317
        • Kant A.K.
        • Schatzkin A.
        • Graubard B.I.
        • Ballard-Barbash R.
        Frequency of eating occasions and weight change in the NHANES I Epidemiologic Follow-up Study.
        Int J Obes Relat Metab Disord. 1995; 19: 468-474
        • Hampl J.S.
        • Heaton C.L.B.
        • Taylor C.A.
        Snacking patterns influence energy and nutrient intakes but not body mass index.
        J Hum Nutr Diet. 2003; 16: 3-11
        • Ma Y.
        • Bertone E.R.
        • Stanek 3rd, E.J.
        • Reed G.W.
        • Hebert J.R.
        • Cohen N.L.
        • Merriam P.A.
        • Ockene I.S.
        Association between eating patterns and obesity in a free-living US adult population.
        Am J Epidemiol. 2003; 158: 85-92
        • McCrory M.A.
        • Howarth N.C.
        • Roberts S.B.
        • Huang T.T.-K.
        Eating frequency and energy regulation in free-living adults consuming self-selected diets.
        J Nutr. 2011; 141: 148-153
        • Bellisle F.
        Impact of the daily meal pattern on energy balance.
        Scand J Nutr. 2004; 48: 114-118
        • Huang T.T.
        • Roberts S.B.
        • Howarth N.C.
        • McCrory M.A.
        Effect of screening out implausible energy intake reports on relationships between diet and BMI.
        Obes Res. 2005; 13: 1205-1217
        • Leidy H.J.
        • Campbell W.W.
        The effect of eating frequency on appetite control and food intake: Brief synopsis of controlled feeding Studies.
        J Nutr. 2011; 141: 154-157
        • Chapelot D.
        The role of snacking in energy balance: A biobehavioral approach.
        J Nutr. 2011; 141: 158-162
        • Hartline-Grafton H.L.
        • Rose D.
        • Johnson C.C.
        • Rice J.C.
        • Webber L.S.
        The influence of weekday eating patterns on energy intake and BMI among female elementary school personnel.
        Obesity. 2010; 18: 736-742
        • Dennis K.E.
        Postmenopausal women and the health consequences of obesity.
        J Obstet Gynecol Neonatal Nurs. 2007; 36: 511-519
        • Physical Activity Guidelines Advisory Committee
        Physical Activity Guidelines Advisory Committee Report, 2008.
        US Department of Health and Human Services, Washington, DC2008
      1. Foster-Schubert KE, Alfano CM, Duggan CR, Xiao L, Campbell KL, Kong A, Bain CE, Wang C-Y, Blackburn GL, McTiernan A. Effect of diet and exercise, alone or combined, on weight and body composition in overweight-to-obese postmenopausal women [published online ahead of print April 14, 2011]. Obesity. doi:10.1038/oby2011.76.

        • Look Ahead Research Group
        Look AHEAD (Action for Health in Diabetes): Design and methods for a clinical trial of weight loss for the prevention of cardiovascular disease in type 2 diabetes.
        Control Clin Trials. 2003; 24: 610-628
        • Diabetes Prevention Program Research Group
        The Diabetes Prevention Program (DPP): Description of lifestyle intervention.
        Diabetes Care. 2002; 25: 2165-2171
        • Patterson R.E.
        • Kristal A.R.
        • Tinker L.F.
        • Carter R.A.
        • Bolton M.P.
        • Agurs-Collins T.
        Measurement characteristics of the Women's Health Initiative food frequency questionnaire.
        Ann Epidemiol. 1999; 9: 178-187
        • Zizza C.A.
        • Arsiwalla D.D.
        • Ellison K.J.
        Contribution of snacking to older adults' vitamin, carotenoid, and mineral intakes.
        J Am Diet Assoc. 2010; 110: 768-772
        • Raynor H.A.
        • Jeffery R.W.
        • Ruggiero A.M.
        • Clark J.M.
        • Delahanty L.M.
        Weight loss strategies associated with BMI in overweight adults with type 2 diabetes at entry into the Look AHEAD (Action for Health in Diabetes) trial.
        Diabetes Care. 2008; 31: 1299-1304
        • Purslow L.R.
        • Sandhu M.S.
        • Forouhi N.
        • Young E.H.
        • Luben R.N.
        • Welch A.A.
        • Khaw K.T.
        • Bingham S.A.
        • Wareham N.J.
        Energy intake at breakfast and weight change: Prospective study of 6,764 middle-aged men and women.
        Am J Epidemiol. 2008; 167: 188-192
        • Wansink B.
        • Payne C.R.
        • Shimizu M.
        “Is this a meal or snack?”.
        Appetite. 2010; 54: 214-216
        • Vartanian L.R.
        • Herman C.P.
        • Wansink B.
        Are we aware of the external factors that influence our food intake?.
        Health Psychol. 2008; 27: 533-538
        • Marmonier C.
        • Chapelot D.
        • Fantino M.
        • Louis-Sylvestre J.
        Snacks consumed in a nonhungry state have poor satiating efficiency: Influence of snack composition on substrate utilization and hunger.
        Am J Clin Nutr. 2002; 76: 518-528
        • Cross A.T.
        • Babicz D.
        • Cushman L.F.
        Snacking patterns among 1,800 adults and children.
        J Am Diet Assoc. 1994; 94: 1398-1403
        • Johnson R.K.
        • Friedman A.B.
        • Harvey-Berino J.
        • Gold B.C.
        • McKenzie D.
        Participation in a behavioral weight-loss program worsens the prevalence and severity of underreporting among obese and overweight women.
        J Am Diet Assoc. 2005; 105: 1948-1951
        • Muhlheim L.S.
        • Allison D.B.
        • Heshka S.
        • Heymsfield S.B.
        Do unsuccessful dieters intentionally underreport food intake?.
        Int J Eat Disord. 1998; 24: 259-266
        • Lissner L.
        Measuring food intake in studies of obesity.
        Public Health Nutr. 2002; 5: 889-892

      Biography

      A. Kong is a postdoctoral research associate, Cancer Education and Career Development Program, Institute for Health Research and Policy, University of Illinois at Chicago
      S. A. A. Beresford is a professor, Department of Epidemiology, School of Public Health and Community Medicine, University of Washington, Seattle, WA
      C. M. Alfano is program director, Office of Cancer Survivorship, Division of Cancer Control and Population Sciences, National Cancer Institute, National Institutes of Health, Bethesda, MD
      K. E. Foster-Schubert is assistant professor of medicine, University of Washington, VA Puget Sound Health Care, Seattle, WA
      M. L. Neuhouser is an associate member, Cancer Prevention Program, Fred Hutchinson Cancer Research Center, Seattle, WA
      C. Duggan is senior staff scientist, Epidemiology Program, Division of Public Health Sciences, Fred Hutchinson Cancer Research Center, Seattle, WA
      C.-Y. Wang is a member, Biostatistics and Biomathematics Program, Division of Public Health Sciences, Fred Hutchinson Cancer Research Center, Seattle, WA
      L. Xiao is a statistical research associate, Fred Hutchinson Cancer Research Center, Seattle, WA
      C. E. Bain is study coordinator, Fred Hutchinson Cancer Research Center, Seattle, WA
      A. McTiernan is director, Prevention Center and member, Epidemiology Program, Division of Public Health Sciences, Fred Hutchinson Cancer Research Center, Seattle, WA
      D. B. Johnson is an associate professor, Nutritional Sciences and associate director, Center for Public Health Nutrition, University of Washington, Seattle, WA

      Linked Article

      • Erratum
        Journal of the Academy of Nutrition and DieteticsVol. 114Issue 6
        • Preview
          The Research and Professional Brief article, “Associations between Snacking and Weight Loss and Nutrient Intake among Postmenopausal Overweight to Obese Women in a Dietary Weight-Loss Intervention,” by Kong and colleagues that appeared in the December 2011 issue of the Journal (pp 1898-1903) contains an error on page 1902. In Table 2, the first category under “Total snacking frequency” is incorrectly listed as “≥1 Snacks/day.” The category should be correctly listed as “≤1 Snacks/day.”
        • Full-Text
        • PDF