Puberty and observed energy intake: boy, can they eat!^,

Abstract

Background: Anecdotal reports suggest that adolescent males consume large quantities of food to meet the growth demands of pubertal development. However, limited experimental data exist to support this impression.

Objective: The objective was to measure energy intakes of youth at different pubertal stages.

Design: Participants were 204 volunteers (50.5% male) aged 8–17 y. Pubertal development was categorized by physical examination into prepuberty (males: testes < 4 mL; females: Tanner breast stage 1), early–midpuberty (males: testes = 4–12 mL; females: Tanner breast stages 2–3), or late puberty (males: testes >12 mL; females: Tanner breast stages 4–5). Energy intake was measured as consumption from a 9835-kcal food array during 2 lunchtime meals.

Results: Males consumed more energy than did females across all pubertal stages (P < 0.001). Intake increased with pubertal development (P < 0.001), but the timing and magnitude of change varied by sex (P = 0.02). Males' unadjusted energy intake was greater in late puberty (mean ± SE: 1955 ± 70 kcal) than in prepuberty (1287 ± 90 kcal) or early–midpuberty (1413 ± 92 kcal) (P < 0.001). Females' unadjusted energy intake tended to be lower among prepubertal girls (905 ± 140 kcal) than among females in early–midpuberty (1278 ± 82 kcal, P = 0.07) or late puberty (1388 ± 68 kcal, P = 0.01). After adjustment for fat-free mass, fat mass, height, overweight status, race, and meal instruction, the main effect of sex (P < 0.001) remained significant, but the effect of puberty was not significant (P = 0.66).

Conclusions: The observed intake patterns are congruent with known sexual dimorphisms for body composition, peak growth velocity, and pubertal development. Consistent with their higher energy requirements, males can consume significantly larger amounts of food than females, especially during later puberty. This trial was registered at clinicaltrials.gov as NCT00320177.

INTRODUCTION

The onset of pubertal development is characterized by rapid physical changes in body size, shape, and composition in both females and males (1). However, there is marked sexual dimorphism in the relations between age, growth, and sexual maturation. Girls typically initiate rapid linear growth contemporaneously with the first appearance of breast development (Tanner breast stage 2) and complete most of their linear growth by age 13 y before completing breast development (Tanner breast stage 5). Males usually achieve peak linear growth velocity later in pubertal development and often grow most rapidly at around age 14 y. These physical changes are accompanied in both males and females by increased energy requirements (2, 3) that would be expected to necessitate greater food intake for growth. Yet, surprisingly few experimental data document the quantity and nature of the food intake changes that accompany puberty. It is commonly asserted anecdotally that adolescent males consume very large amounts of food compared either to adolescent females or to their counterparts who have not begun puberty (4). Empirical data supporting these claims are largely derived from self-report measures such as dietary recalls (5, 6) and food frequency questionnaires (7). Males of all ages self-report greater unadjusted energy intake than age-matched females (5–7). Similarly, on the basis of 24-h dietary recall methods from the National Health and Nutrition Examination Surveys, energy intake is reported to increase with age, with males reporting more intake than females, especially during adolescence (6). Furthermore, among adolescent males, reported unadjusted energy intake increases throughout adolescence, peaking at age 16–19 y; among adolescent females, reported energy intake increases during adolescence but appears to peak earlier, at age 12–15 y, and has been suggested to decline slightly at age 16–19 y (6).

Existing self-report data are limited by the very poor correspondence between self-reports of food intake and actual intake (8–11), making it difficult to establish to what extent sex differences reflect actual differences in intake compared with reporting biases. We therefore measured by direct observation the actual energy intakes of prepubertal, early–midpubertal, and late-pubertal males and females. We hypothesized that males would consume more energy overall than females and that energy intake would increase between prepuberty and late puberty. Consistent with popular lore (4) and prior self-report data (6), we further predicted that puberty and sex would interact such that males compared with females would show larger increases in intake across pubertal development. In addition, we investigated whether any observed puberty and sex differences could be accounted for by body composition. Specifically, we expected that accounting for individual variations in the fat-free mass and fat mass of youth would largely explain the greater energy intake of adolescent males (compared with adolescent females) and of pubertal (compared with nonpubertal) youth.

SUBJECTS AND METHODS

Participants

Participants were healthy volunteers recruited for an observational protocol through flyers posted on public bulletin boards at the National Institutes of Health (NIH), at local libraries and supermarkets, and on school parent e-mail listservs in the Washington, DC, greater metropolitan area. The study was also advertised in a local newspaper and posted on free websites. The flyer and advertisement explained that the study was investigating eating behavior in youth and that no treatment would be provided. As described in detail previously (12), healthy males and females of any race or ethnicity who had a body mass index (BMI) >5th percentile for age and sex according to the Centers for Disease Control and Prevention standards (13) were eligible for participation. Individuals were excluded if they were under a physician's care for a chronic illness; were shown to have abnormal pulmonary, hepatic, renal, or endocrine function; were taking medication known to affect body weight; had recently lost weight; or had a psychiatric disorder that might impede protocol adherence. Youth provided written assent, and parents gave written consent for participation in the study. Children and adolescents were financially compensated for their participation. This study was approved by the Eunice Kennedy Shriver National Institute of Child Health and Human Development Institutional Review Board.

Procedures

As described previously (12), all assessments were conducted in an outpatient clinic at the Hatfield Clinical Research Center at the NIH. Participants were instructed to adhere to an overnight fast initiated at 2200 the night before visiting the NIH to be screened for eligibility. Height was measured 3 times to the nearest millimeter by a stadiometer (Holtain, Crymmych, Wales, United Kingdom) calibrated before each participant was measured. Weight was measured to the nearest 0.1 kg with a calibrated digital scale (Scale-Tronix, Wheaton, IL). Height and weight were used to compute BMI, calculated as weight (kg) divided by the square of height (m). BMI SD (BMIz) scores were calculated according to the Centers for Disease Control and Prevention 2000 standards (13).

Fat-free mass (kg) and body-fat mass (kg) were assessed with air-displacement plethysmography (Life Measurement Inc, Concord, CA). Measurements were obtained as recommended; participants were studied after an overnight fast while wearing only underclothes (14).

All participants underwent a medical history and a physical examination performed by an endocrinologist or trained nurse practitioner. Testicular volume (mL) was measured by using a set of orchidometer beads as standards according to Prader (15), and breast development was assigned according to the 5 stages of Tanner (16, 17). Testicular volume for males and Tanner breast staging for females were used to categorize youth as those in prepuberty (males: testes <4 mL; females: Tanner stage 1), early–midpuberty (males: testes = 4–12 mL; females: Tanner stages 2–3), or late puberty (males: testes >12 mL; females: Tanner stages 4–5). In cases in which the stage was discordant between the right and left testes/breasts, the higher stage was assigned. Ratings of pubic hair and age at menarche also were obtained, but genital/breast measures were used because there is some suggestion that these are the most accurate single indicators of sexual maturity (18).

Participants returned 2 more times for visits scheduled ≥48 h apart, during which they were asked to consume their lunch ad libitum from a multiple-item buffet test meal. Participants were instructed to adhere to an overnight fast initiated at 2200 the night before each meal appointment. On each test meal day, youth were served a standard breakfast consisting of 280 kcal (240 g apple juice, 1 English muffin, and 6 g butter) at 0840. After the breakfast, subjects remained at the NIH to participate in sedentary activities (eg, playing computer games, reading, or doing arts and crafts). Youth were observed to ensure that there was no food intake until the afternoon lunch meal. There was a 6-h period of food deprivation to ensure that participants would be hungry for the experimental test meal. At 1430, participants were presented with a 9835-kcal buffet test meal with individual items that varied in macronutrient composition (12% protein, 51% carbohydrate, and 37% fat across all foods) and contained a wide assortment of foods (Figure 1) (19, 20). When the subject entered the room containing the buffet, she or he received one of 2 meal instructions (in random order): “Eat as much as you would at a normal meal“ or ”Let yourself go and eat as much as you want.” Participants were then left alone in the room containing the buffet to eat ad libitum with the instruction to open the room door when they were finished eating. The amount of each food and beverage consumed was measured as the difference in weight (g) for each food item before and after the meal. Energy content (kcal) and macronutrient composition (% protein, % carbohydrate, and % fat) for each food were determined with data from the USDA National Nutrient Database for Standard Reference (release 16; USDA Agricultural Research Service, Beltsville, MD) and manufacturer information when available. In some analyses, we also examined participants' observed percentage intake of estimated energy requirements (EER; kcal/d). EER was calculated according to the Institute of Medicine guidelines (21): for males, ERR = 88.5 − (61.9 × age in y) + 1.13 × [(26.7 × weight in kg) + (903 × height in m)] + 25; for females, EER = 135.3 − (30.8 × age in y) + 1.16 × [(10.0 × weight in kg) + (934 × height in m)] + 25.

FIGURE 1.

Buffet test meal consisting of a wide assortment of foods that varied in macronutrient composition (12% protein, 51% carbohydrate, 37% fat; 9835 kcal). Foods included bread, deli meats and cheese, chicken nuggets, vegetables, fruit, chips, cookies, candy, condiments, water, juice, lemonade, and milk.

Statistical methods

Given the expectation of a 450-kcal SD, a power calculation suggested that a sample of 192 youths with similar numbers of prepubertal, early–midpubertal, and late-pubertal males and females would have 95% power to detect 150-kcal differences among groups by analysis of variance (22). Analyses were performed with SPSS 16.0 (SPSS, Chicago, IL) (23). Data screening indicated no influential outliers and acceptable levels of skew and kurtosis (24). Linear mixed modeling with repeated measures was used to examine the hypothesized puberty- and sex-related differences in total energy intake at both lunchtime meals. The dependent variable was total energy intake (kcal); the fixed-factor main effects were sex (male, female) and puberty (prepuberty, early–midpuberty, late puberty); and the fixed-factor interaction was sex by puberty. Other interactions were not modeled. The same model was run examining 2 sets of covariates. The first adjustment included overweight status (BMI <85th percentile, BMI ≥85th percentile), race (white, non-Hispanic, other), and meal instruction. The second adjustment included overweight status, race, and meal instruction, as well as fat-free mass (kg), fat mass (kg), and height (cm). This approach allowed us to determine whether any observed effects in sex and puberty could be explained specifically by body composition. Also, as a follow-up to any observed effects, we tested whether the estimates were significant after inclusion of weight and height rather than body composition (fat mass, fat-free mass, and height). Pairwise comparison tests with a Bonferroni correction were used to follow up on planned comparisons among prepuberty, early–midpuberty, and late puberty for males and females separately (25, 26). Although the primary variable of interest was total observed energy intake, we also conducted parallel linear mixed modeling and follow-up analyses to examine puberty and sex differences in percentage intake of EER (total kcal/d) and percentage macronutrient intakes from carbohydrate, fat, and protein. Arcsine transformations were performed for analyses of the percentages of macronutrient intake. Back-transformed values are shown in the figures and tables.

RESULTS

A total of 225 children and adolescents attended a screening appointment, and 19 children and adolescents did not complete all of the study procedures for the following reasons: 9 were unable to consume 50% of the high-calorie shake and were excluded, 4 did not schedule their test-meal visits within 6 mo of the screening appointment, 2 scheduled test-meal appointments but did not attend, and 4 decided not to continue the study but did not provide a reason. In addition, 2 participants were excluded because one did not eat at the test meals and the other did not follow study procedures. The final sample consisted of 103 males and 101 females between age 8 and 17 y. There were approximately equal numbers of nonoverweight youth (BMI <85th percentile for age and sex; 52.5%) and overweight or obese youth (BMI ≥85th percentile for age and sex; 47.5%). Sample demographic and anthropometric characteristics by pubertal status and sex are described in Table 1. As expected, females had greater fat mass (kg) than males (P = 0.02), and conversely, males had greater fat-free mass as a percentage of total mass than females (P < 0.001).

TABLE 1.

Demographic and anthropometric characteristics of the study sample¹

	Prepuberty2		Early–midpuberty3		Late puberty4
	Males (n = 25)	Females (n = 13)	Males (n = 28)	Females (n = 35)	Males (n = 49)	Females (n = 51)
Age (y)	9.5 ± 1.05	9.6 ± 1.1	11.8 ± 2.0	11.5 ± 1.7	15.4 ± 1.5	14.9 ± 1.7
Race (% white)	56.0	69.2	71.4	45.7	63.3	62.7
SES (median)6	2	2	2	2	2	2
Weight (kg)	38.2 ± 15.8	30.1 ± 5.8	54.6 ± 34.2	56.6 ± 19.1	77.2 ± 22.0	74.9 ± 30.8
Height (cm)	138.5 ± 9.1	135.6 ± 8.2	150.9 ± 11.8	150.4 ± 8.7	172.2 ± 7.7	163.3 ± 7.3
BMI (kg/m2)	19.4 ± 5.1	16.2 ± 1.7	23.1 ± 11.4	24.7 ± 7.1	25.9 ± 7.0	27.9 ± 10.7
BMI z score	0.7 ± 1.1	−0.3 ± 0.7	0.7 ± 1.3	1.3 ± 1.1	1.1 ± 1.1	1.1 ± 1.1
BMI percentile	0.7 ± 0.3	0.4 ± 0.2	0.6 ± 0.3	0.8 ± 0.2	0.8 ± 0.2	0.8 ± 0.2
Overweight (%)7	40.0	7.7	42.9	57.1	49.0	54.9
Fat mass (kg)	10.6 ± 10.1	6.5 ± 4.1	17.8 ± 23.2	20.5 ± 13.3	18.7 ± 16.8	27.5 ± 20.4
Fat-free mass (kg)	26.9 ± 6.0	30.9 ± 13.2	36.2 ± 12.1	35.5 ± 7.2	57.7 ± 9.8	48.3 ± 13.5
Fat-free mass (%)	75.6 ± 11.6	82.4 ± 8.5	73.5 ± 12.8	66.5 ± 12.5	78.4 ± 13.6	67.0 ± 10.4

¹SES, socioeconomic status.

²Prepuberty defined for males as testicular volume <4 mL and for females as Tanner breast stage 1.

³Early–midpuberty defined for males as testicular volume of 4–12 mL and for females as Tanner breast stage 2–3.

⁴Late puberty for males defined as testicular volume >12 mL and for females as Tanner breast stage 4–5.

⁵Mean ± SD (all such values).

⁶SES as assessed with the Hollingshead Index (33).

⁷Overweight defined as BMI ≥85th percentile.

In unadjusted analyses for total energy intake, there were main effects for sex and puberty (Ps < 0.001) and an interaction of sex by puberty (P = 0.02; Figure 2A). Males ate more than females at all stages of puberty and, overall, energy intake increased across puberty. However, the magnitude of intake across puberty varied by sex. Follow-up subgroup analyses by sex corrected for multiple comparisons indicated a main effect of puberty among males (F = 19.6, P < 0.001) such that males' energy intake was significantly greater in late puberty compared with prepuberty (difference = 704.7 ± 124.4 kcal, P < 0.001) or early–midpuberty (difference = 534.1 ± 170.6, P < 0.001); males' intake in prepuberty compared with early–midpuberty was not significantly different (difference = 170.6 ± 139.1, P = 0.67). There also was a main effect of puberty among females (F = 5.1, P = 0.01). However, in contrast, compared with prepubertal females, females in early–midpuberty tended to eat more (difference = 347.6 ± 150.6, P = 0.07) and females in late puberty ate significantly more (difference = 457.4 ± 144.1, P = 0.01); there was no difference in females' intake between early–mid- and late puberty (difference = 109.8 ± 101.8, P = 0.85).

FIGURE 2.

Mean (±SE) total energy intake derived from linear mixed modeling analyses among males (n = 25) and females (n = 13) in prepuberty (males: testicular volume <4 mL; females: Tanner breast stage 1), males (n = 28) and females (n = 35) in early–midpuberty (males: testicular volume = 4–12 mL; females: Tanner breast stage 2–3), and males (n = 49) and females (n = 51) in late puberty (males: testicular volume >12 mL; females: Tanner breast stage 4–5). Values that do not share a common letter were significantly different at P < 0.05. A: Unadjusted total energy intake for prepubertal males (1242.2 ± 95.9 kcal) and females (930.1 ± 132.5 kcal), early–midpubertal males (1412.7 ± 90.3 kcal) and females (1277.7 ± 80.7 kcal), and late-pubertal males (1946.8 ± 68.2 kcal) and females (1387.5 ± 66.9 kcal). B: Adjusted total intake (covariates: race, overweight status, and meal instruction) for prepubertal males (1276.7 ± 87.6 kcal) and females (1073.4 ± 123.9 kcal), early–midpubertal males (1493.9 ± 83.5 kcal) and females (1244.6 ± 73.9 kcal), and late-pubertal males (1986.4 ± 62.9 kcal) and females (1413.9 ± 61.9 kcal). C: Adjusted total intake (covariates: fat-free mass, fat mass, height, race, overweight status, and meal instruction) for prepubertal males (1550.5 ± 98.1 kcal) and females (1313.1 ± 129.8 kcal), early–midpubertal males (1550.2 ± 77.6 kcal) and females (1330.9 ± 72.1 kcal), and late-pubertal males (1773.3 ± 85.3 kcal) and females (1270.4 ± 61.2 kcal).

In analyses including adjustment for race, overweight status, and meal instruction, there were significant effects for all covariates on total energy intake (Ps < 0.01), with nonwhite youth eating more than white youth, overweight youth consuming more than nonoverweight, and greater intake with the instruction to “let yourself go” (compared with “eat as you would at a normal meal”). The main effects of sex and puberty (Ps < 0.001) and the interaction of sex by puberty (P = 0.03) remained significant (Figure 2B). Follow-up comparisons by sex indicated parallel findings to those in unadjusted analyses. In addition, accounting for body composition (fat-free mass, fat mass, and height) in the model revealed an effect for fat mass (P < 0.001) and height (P = 0.04), but not for fat-free mass (P = 0.38). The same main effect for sex (P < 0.001) was observed, but the effect of puberty (P = 0.66) and the interaction effect of sex by puberty (P = 0.08) were no longer significant (Figure 2C). Follow-up comparisons separated by sex revealed no significant effects of puberty among males or females after adjustment for body composition (P > 0.58). When weight and height instead of body composition (fat mass, fat-free mass, and height), as well as race, overweight status, and meal instruction, were included in the model, the results were the same: there was a main effect of sex on total intake (P < 0.001), but no effect of puberty (P = 0.82) or sex by puberty (P = 0.11).

Mean ± SE observed percentage intake of estimated energy requirements (EERs) estimated on the basis of sex, age, weight, and height for males and females in prepuberty, early–midpuberty, and late puberty is shown in Table 2. Males and females at all stages of puberty ate a similar percentage of their predicted EERs at lunchtime meals (range: 53–62%) in unadjusted as well as in both sets of analyses adjusting for covariates.

TABLE 2.

Linear mixed modeling analyses examining sex, puberty, and their interaction for observed percentage energy and macronutrient intake of males and females in prepuberty, early–midpuberty, and late puberty¹

	Prepuberty2		Early–midpuberty3		Late puberty4		P values
Energy intakes	Males (n = 25)	Females (n = 13)	Males (n = 28)	Females (n = 35)	Males (n = 49)	Females (n = 51)	Sex	Puberty	Sex × puberty
EER (%)5	60.5 ± 3.46	54.6 ± 4.7	56.7 ± 3.2	60.6 ± 2.9	60.2 ± 2.5	59.2 ± 2.4	0.71	0.81	0.38
Adjusted7	60.8 ± 3.2	54.2 ± 4.6	58.4 ± 3.1	60.8 ± 2.8	61.5 ± 2.4	60.9 ± 2.3	0.52	0.52	0.44
Adjusted8	59.6 ± 4.1	53.3 ± 5.4	58.2 ± 3.2	60.4 ± 3.0	62.2 ± 3.5	61.6 ± 2.5	0.55	0.59	0.49
Carbohydrate (%)	50.8 ± 1.4	49.6 ± 2.0	48.4 ± 1.3	48.4 ± 1.2	48.6 ± 1.0	48.8 ± 1.0	0.76	0.47	0.89
Adjusted7	50.7 ± 1.4	48.9 ± 2.0	48.2 ± 1.4	48.6 ± 1.2	48.6 ± 1.0	48.8 ± 1.0	0.73	0.65	0.75
Adjusted8	50.6 ± 1.8	49.9 ± 2.3	47.9 ± 1.4	48.2 ± 1.3	49.0 ± 1.5	48.1 ± 1.1	0.73	0.43	0.88
Protein (%)	13.8 ± 0.6x	14.8 ± 0.9	15.3 ± 0.6x,y	14.9 ± 0.5	15.7 ± 0.4y	13.8 ± 0.4	0.37	0.49	0.059
Adjusted7	13.8 ± 0.6x	14.7 ± 0.9	15.2 ± 0.6x,y	14.9 ± 0.5	15.6 ± 0.4y	13.7 ± 0.4	0.41	0.48	0.049
Adjusted8	13.8 ± 0.8	14.2 ± 1.0	15.3 ± 0.6	15.2 ± 0.6	15.3 ± 0.7	14.0 ± 0.5	0.55	0.20	0.39
Fat (%)	36.6 ± 1.1	36.8 ± 1.5	37.5 ± 1.0	38.0 ± 0.9	36.9 ± 0.8	38.7 ± 0.8	0.32	0.53	0.68
Adjusted7	36.7 ± 1.1	37.7 ± 1.5	37.9 ± 1.0	37.8 ± 0.9	37.0 ± 0.8	38.8 ± 0.8	0.32	0.76	0.59
Adjusted8	37.0 ± 1.3	37.4 ± 1.8	38.1 ± 1.1	37.9 ± 1.0	36.9 ± 1.2	39.0 ± 0.8	0.36	0.80	0.48

¹Arcsine-transformed carbohydrate, protein, and fat were used in the analyses, but untransformed values are presented for ease of interpretation. EER, estimated energy requirement. For males, values not sharing a common superscript letter were significantly different at P < 0.05.

²Prepuberty defined for males as testicular volume <4 mL and for females as Tanner breast stage 1.

³Early–midpuberty for males defined as testicular volume of 4–12 mL and for females as Tanner breast stage 2–3.

⁴Late puberty for males defined as testicular volume >12 mL and for females as Tanner breast stage 4–5.

⁵Observed percentage intake of EER (kcal/d) calculated on the basis of sex, age, weight, and height (21).

⁶Observed mean ± SE (all such values).

⁷Adjusted for race (white, other), overweight status (BMI: <85th percentile, ≥85th percentile), and meal instruction.

⁸Adjusted for fat-free mass (kg), fat mass (kg), height (cm), race, overweight status, and meal instruction.

⁹Follow-up analyses by sex indicated no significant differences between females.

Percentage macronutrient consumption from carbohydrate (range: 48–51%) and fat (range: 37–39%) was similar among males and females (main effect sex, P > 0.31) and among youth in all stages of puberty (main effect puberty, P > 0.42). However, in the unadjusted model predicting percentage protein intake, there was an interaction of sex by puberty (P = 0.05). Follow-up analyses indicated a significant main effect of puberty among males only (F = 3.2, P = 0.04). Late-puberty males consumed greater energy intake from protein than prepuberty males (difference = 1.9 ± 0.8%, P = 0.04); early–midpuberty protein intake of males did not differ significantly from that of males in prepuberty (difference = 1.5 ± 0.9, P = 0.25) or of males in late puberty (difference = 0.4 ± 0.7, P = 1.0). The percentage protein intake of females was similar at all pubertal stages (F = 1.3, P = 0.27). The overall sex by puberty interaction remained significant (P = 0.04) in analyses predicting protein intake after adjustment for race, overweight status, and meal, but not after adding fat-free mass, fat mass, and height to the model (P = 0.39). In contrast, if the model included weight and height rather than body composition, as well as the other covariates, the effect of sex by puberty was significant (P = 0.05).

We performed a number of follow-up analyses. First, we conducted a series of follow-up analyses to test whether the observed findings in total and macronutrient energy intake for puberty and sex were consistent across weight status (defined as BMI <85th percentile compared with ≥85th percentile). There was no significant weight status by puberty interaction in any model (P = 0.46–0.92). The only significant weight status by puberty by sex interaction was in the model examining the percentage intake of EERs (P = 0.05). Further investigation indicated that, among nonoverweight youth, males tended to eat a greater percentage of their EERs at meals than females (65.5 ± 2.8% compared with 58.1 ± 2.9%, P = 0.07) whereas the reverse was true among overweight youth: overweight females tended to eat a greater percentage of their EERs at meals than overweight males (62.6 ± 5.1% compared with 53.2 ± 2.2%, P = 0.09). Neither puberty nor sex by puberty was significant among nonoverweight or overweight youth (P > 0.60). No other 3-way interaction among overweight status, sex, and puberty was significant (P = 0.14–0.84). Although these analyses may be somewhat under-powered, the findings at least suggest that the observed patterns in male and female total and percentage macronutrient intakes at pre-, early–mid-, and late puberty may be similar for nonoverweight and overweight or obese youth. We conducted a second set of follow-up analyses to determine whether the observed effects of sex and puberty were independent of age. Sex, puberty, and their interaction remained significant (P < 0.03) after accounting for age (P = 0.42) in the prediction of total energy intake. The sex by puberty interaction for percentage intake from protein also remained significant (P = 0.05) after including age (P = 0.37) in the model.

DISCUSSION

In the current study, we observed the actual energy intakes of normal weight and overweight males and females in prepuberty, early–midpuberty, and late puberty during 2 lunchtime meals. Males and females at all pubertal stages consumed a similar percentage (≈55–60%) of their total daily estimated energy needs at lunchtime laboratory test meals. However, congruent with our predictions based on prior self-report data (5–7), males consumed more absolute total energy than females at all stages of puberty. This main effect remained even after adjusting for individual differences in body composition or body mass. Males do not simply report eating more (5–7); indeed, males do eat more than girls at lunchtime laboratory test meals.

Consistent with the notion that puberty is a developmental period characterized by rapid physical growth in body size, shape, and composition (1), youths' total energy intake increased throughout the stages of pubertal development. Compared with boys and girls in prepuberty, their late-puberty counterparts consumed ≈52% more in total energy intake in unadjusted analyses. However, the timing of significant increases in energy intake was notably different for males compared with females. For males, there was a nonsignificant 10% increase in energy intake between prepuberty and early–midpuberty whereas, for females, there was a 41% increase between pre- and early–midpuberty. During late puberty, adolescent males consumed 38% more total energy intake than early–midpubertal boys whereas late-pubertal adolescent females showed only 10% greater intake relative to early–midpubertal adolescent females. These patterns correspond to the periods of most rapid growth for males and females, with males typically experiencing their most significant growth spurt during late puberty, 2–3 y after the growth spurt of girls (27, 28). Conversely, females' most rapid growth usually occurs in early–midpuberty, and females have completed most of their linear growth by the later stages of pubertal development (27, 28). The differences observed—greater energy intake for males during late puberty and for females during early–midpuberty—appear to parallel the increased energy needs to support such periods of most rapid physical change.

The observation that males ate more than females throughout pubertal development was still observed after adjusting for body composition or body mass variables. This finding corresponds to the expected increase in males' total daily EERs from prepuberty through late puberty (for our sample: 2096 ± 113–3221 ± 85 kcal) being greater than that of females in prepuberty through late puberty (1689 ± 157–2332 ± 84 kcal). Throughout pubertal development, males have significant, additional growth needs that are not accounted for by their current body size.

In contrast, taking into account the variability in fat-free mass and fat mass or weight completely explained the main effect of puberty on energy intake and attenuated the puberty-associated differences in total intake among males and among females. These results suggest that the increasingly greater energy intake during pubertal development is largely a reflection of gains in body mass, with greater accumulation of low-energy–expending fat mass during pubertal development in females and greater increases in fat-free mass in males (29). Indeed, adolescent males in late puberty consumed a greater percentage of their intake from protein than prepubertal boys after adjustment for body weight, but not after adjustment for body composition. Adolescent males' greater percentage intake in protein consumption thus appears to correspond to the gains in body mass, especially lean body mass, which characterizes late puberty in males.

The current findings should be considered in light of this study's limitations. Although the hypotheses were planned and based on sound knowledge about pubertal development, the cross-sectional and correlational design of the study limits any causal conclusions that can be drawn about the effects of pubertal development on energy intake. For example, other third variables might explain the associations among sex, puberty, and energy intake, such as sex differences in patterns of physical activity during puberty. Although participants' intake during test-meal studies have been shown to reflect their eating behavior outside of the laboratory (30, 31), there remains the possibility that such studies may not yield the same eating patterns that occur in the natural environment. For example, because participants were aware that food intake was being measured, it cannot be ruled out that social desirability factors did not exert an effect. Furthermore, lunchtime meal intake may not necessarily be reflective of total energy intake measured over many days or months, and the prolonged 6-h duration between the standard breakfast and lunch test meal might have promoted greater intake. Finally, although it is plausible that adolescent males have energy needs for growth greater than that which can be explained by their current size, we could not relate intake to actual growth velocity because we did not have longitudinal measurements of linear growth from participants. The benefits of laboratory studies include the ability to control the exact conditions, to directly observe eating behaviors, and to precisely measure energy intake (32). Such methodology has advantages over questionnaire measures of intake, in which individuals frequently underestimate food consumption (9–11).

In conclusion, males, especially those in later adolescence, consume considerably more than females of a similar age and pubertal stage at lunchtime meals. When health care professionals counsel families regarding what might be appropriate or expected for the food intake of adolescents, both sex and pubertal stage should be taken into consideration.