High Respiratory Quotient Is Associated with Increases in Body Weight and Fat Mass in Young Adults

ORIGINAL ARTICLE High respiratory quotient is associated with increases in body weight and fat mass in young adults

RP Shook1, GA Hand2, AE Paluch3, X Wang3, R Moran4, JR Hébert4,5,6, JM Jakicic7 and SN Blair3,4

BACKGROUND/OBJECTIVES: Metabolic disturbances, such as reduced rates of fat oxidation (high respiratory quotient (RQ)) or low energy expenditure (low resting metabolic rate (RMR)), may contribute to obesity. The objective was to determine the association between a high RQ or a low RMR and changes in body weight and body composition over 1 year. SUBJECTS/METHODS: We measured RQ and RMR in 341 adults using indirect calorimetry, along with body weight/body composition using dual-energy X-ray absorptiometery, energy expenditure using an arm-based activity monitor and energy intake using dietary recalls. Participants were classiﬁed into low, moderate or high RQ and RMR (adjusted for age, sex, race and body composition) groups according to tertiles by sex. Follow-up measurements were completed every 3 months. RESULTS: Individuals with a high RQ had larger gains in body weight and fat mass compared with individuals with a low/moderate RQ at month 3, and increases in fat mass were more than double among individuals with a high RQ at 12 months (1.3 ± 3.0 vs 0.6 ± 3.7 kg, P = 0.03). Individuals with a low RMR did not gain more body weight nor fat mass compared with individuals with a moderate/high RMR. CONCLUSION: The primary ﬁnding is a high RQ is predictive of gains in body weight and fat mass over a 12-month period among young adults, with changes occurring as soon as 3 months. In addition, a low RMR was not associated with gains in body weight or fat mass over the same period. European Journal of Clinical Nutrition (2016) 70, 1197–1202; doi:10.1038/ejcn.2015.198; published online 25 November 2015

INTRODUCTION studies suggest that low RMR is predictive of subsequent weight 3,7–10 1,2,5 At the most basic level, obesity is the result of a chronic imbalance gain, whereas others do not. between energy intake and energy expenditure. However, the Thus, the purpose of the present study was to explore the exact etiology is considerably more complex and may involve a longitudinal associations of RQ and RMR on changes in body variety of physiological and behavioral factors. Metabolic dis- weight and body composition over 12 months in a group of healthy young adults. We also aimed to document temporal turbances, including reduced fat oxidation and reduced resting changes in body composition, RQ and RMR through rigorous and metabolic rate (RMR), have been identified as possible predictors repeated measures of each over 1 year. of changes in body weight and body composition. Respiratory quotient (RQ) reflects the ratio of carbohydrate to fat oxidation; when measured in a fasting state stored fat is the MATERIALS AND METHODS primary fuel source. If an individual has a low RQ, she/he oxidizes This manuscript reports findings from The Energy Balance Study, which has more stored fat at rest compared with an individual with a high been described in detail previously.11 Briefly, participants were healthy RQ and theoretically is protected against future fat accumulation. young adults aged ⩾ 21 and ⩽ 35 years and with a BMI ⩾ 20 and ⩽ 35 kg/ A relatively small body of research exists on the prospective m2. All study protocols were approved by the University of South Carolina relationship between RQ and weight gain, with early studies from Institutional Review Board, and informed consent was obtained from each two decades ago suggesting a positive association,1–3 but results participant before data collection. 4,5 A dual energy X-ray absorptiometer was used to measure bone mineral from more recent studies mixed. density, fat mass (FM) and fat-free mass (FFM). The scan was completed RMR, the amount of calories burned performing normal with a Lunar DPX system (version 3.6; Lunar Radiation Corp, Madison, WI). physiological functions (for example, respiration, brain activity), All anthropometric measurements were completed once every 3 months represents the largest contributor (60–80%) of total energy for the duration of the study. In addition, self-reported body weight expenditure in humans. Given the intricate balance of energy 12 months before baseline and self-reported weight gain 44.5 kg (10 intake and expenditure in the regulation of body weight, it is pounds) over the 3 months before baseline also were recorded. RQ and RMR were measured at baseline via indirect calorimetry using a hypothesized that small changes in RMR could result in a large 6 ventilated hood and an open-circuit system, TrueOne 2400 Metabolic reduction in the number of calories burned over time; however, Measurement Cart (ParvoMedics, Salt Lake City, UT, USA), following a this relationship is uncertain. For example, some prospective standardized protocol.12,13 Briefly, a 15-min resting period preceded

1Department of Kinesiology, Iowa State University, Ames, IA, USA; 2School of Public Health, University of West Virginia, Morgantown, WV, USA; 3Department of Exercise Science, Arnold School of Public Health, University of South Carolina, Columbia, SC, USA; 4Department of Epidemiology and Biostatistics, Arnold School of Public Health, University of South Carolina, Columbia, SC, USA; 5Department of Family and Preventive Medicine, University of South Carolina, Columbia, SC, USA; 6South Carolina Statewide Cancer Prevention and Control Program, Arnold School of Public Health, University of South Carolina, Columbia, SC, USA and 7Department of Health and Physical Activity, Physical Activity and Weight Management Research Center, University of Pittsburgh, Pittsburgh, PA, USA. Correspondence: Dr RP Shook, Department of Kinesiology, Iowa State University, 241 Forker Building, Ames, IA 50011, USA. E-mail: [email protected] Received 3 January 2015; revised 1 October 2015; accepted 8 October 2015; published online 25 November 2015 Fat oxidation, metabolic rate, weight gain RP Shook et al 1198 a 30-min data-collection period; the metabolic cart was calibrated before (46 years using NDSR) registered dietitians employing a multi-pass each test using known gas concentrations and volumes as recommended approach, which utilizes prompting to reduce food omissions and by the manufacturer; all measurements occurred in the morning standardizes the interview methodology across interviewers.23 Interviews (o0900 hours) following a 12-h dietary fasting state and at least 24 h occurred on randomly selected non-consecutive days over 14 days to after the last bout of any structured exercise; participants remained quiet minimize preparation that could bias recall.24 As described above, and still in the supine position throughout the entire procedure and were measurement of energy intake was completed once every 3 months. kept awake with continuous monitoring; and the room was maintained in Participant characteristics were based on demographic and physiologi- low light, noise was kept at a minimum and the temperature was maintained cal measurements using means and s.d.’s for continuous variables and at a constant 26–30 °C. RMR was determined as the average value of 10 percentages for categorical variables. Statistical significance for compar- consecutive minutes with the lowest coefficient of variation. RMR was ison between groups was tested using t-tests for continuous variables and calculated from O2 consumption and CO2 production as measured chi-square tests for categorical variables. Pearson’s correlation coefficients continuously during the testing period with a constant airflow rate into were calculated between the continuous variables and respiratory 12,14,15 the hood; RQ was calculated as VCO2/VO2. Given no widely accepted quotient. A linear mixed models regression random-intercept growth criteria exist to categorize RQ levels, participants were classified as ‘high’ if model was used to analyze the longitudinal data25 for changes in first total they were in the upper tertile for RQ among the entire cohort by sex (Males, body weight, followed by FM. Statistical adjustments were made for RQ ⩾ 0.807; Females, RQ ⩾ 0.805) or ‘Low/Moderate’ if they were in the potential confounders (sex, age, race and change in MVPA, energy intake bottom two tertiles at baseline. This process was repeated for RMR (adjusted and energy expenditure). All computations were performed using SAS 9.3 for age, sex, race, FFM and FM), with participants classified as ‘low’ if they (Cary, NC, USA). Because of differences in sample size between the groups, were in the bottom tertile for RMR among the entire cohort by sex (Males, a post hoc power analysis was performed using G*Power 3 (Germany), RMR ⩽ 1618.4; Females, RMR ⩽ 1275.7) or ‘moderate/high’ if they were in the which yielded a power of 0.88 to detect F-test differences between groups upper two tertiles at baseline. In addition, we estimated RMR based on two based on an effect size of 0.50. widely cited prediction equations: Harris–Benedict16 and Mifflin–St. Jeor.17 We repeated the primary analysis using other classification options (groups based on tertiles, quintiles and so on), and the results were similar to the two- RESULTS fi group analysis. Cardiorespiratory tness (CRF) testing was conducted at Selected demographic and anthropometric variables are reported baseline on a treadmill (Trackmaster 425, Carefusion, Newton, KS, USA) with respiratory gases sampled using a TrueOne 2400 Metabolic Measurement overall and for each RQ group in Table 1 (by the RMR group in Cart (ParvoMedics, Salt Lake City, UT, USA) using a Modified Bruce protocol, Supplementary Table 1). Overall, our participants (n = 419) were and all participants exercised to volitional fatigue. young adults (27.6 ± 3.8 years) with nearly equal numbers of males Energy expenditure was estimated using an arm-based monitor (48.7%) and females (51.3%), with no difference in body weight or (SenseWear Mini Armband, BodyMedia Inc. Pittsburgh, PA, USA). The body composition between groups. There was no difference in monitor is a portable, multi-sensor device worn on the upper arm, energy balance (body weight change in the previous 12 months incorporating tri-axial accelerometry and measures of heat flux, galvanic and percent of participants with weight gain 44.5 kg in the skin response, skin temperature and near-body ambient temperature. previous 3 months) between groups. However, participants in the These measures are entered in combination with demographic informa- low/moderate group lost a small amount of weight (−0.4 ± 0.9 kg) tion into an algorithm to estimate total energy expenditure and the time spent in moderate-to-vigorous physical activity (MVPA, defined as at least during the baseline period (10.2 ± 8.2 days), whereas the high RQ 8 of 10 consecutive minutes of activity at 43 metabolic equivalents of participants maintained their weight. task).18 The armband has been shown to be a valid device for measuring The RQ was significantly higher in the high-RQ group energy expenditure and activity with interclass correlations between (0.841 ± 0.032 vs 0.766 ± 0.025, Po0.0001) and is displayed in doubly labeled water (DLW) of 0.81–0.85,19–21 and comparison with DLW in Table 2. Unadjusted absolute resting energy expenditure (kJ/day) the current sample yielded an interclass correlation of 0.88 (unpublished was not different between the groups. However, RMR relative to data). The participants were asked to wear the armband for 10 days and body weight was slightly higher in the low/moderate RQ were deemed compliant if they completed 5 days of wear (including two compared with the high-RQ group (2.99 ± 0.35 vs 2.90 ± 0.35 ml/ ⩾ weekend days) with 18.5 h of wear time on each of the days. kg per min, P = 0.01) but not different relative to FFM (120.1 ± 14.0 Measurement of energy expenditure and the time spent in physical vs 121.7 ± 15.2, P = 0.28). There was no difference in total energy activity were completed once every 3 months for the duration of the study corresponding to measurement of energy intake and began immediately intake or in diet composition between the groups. In addition, following assessment of body composition and metabolic measurements. slightly higher levels of total energy expenditure (P = 0.11), time Energy intake was measured using interviewer-administered 24-h spent in MVPA (P = 0.09) and CRF (P = 0.10) were observed in the dietary recalls, the Nutrient Data System for Research software (NDSR, low/moderate-RQ group; however, differences did not reach Version 2012).22 Dietary recalls were collected by a team of experienced statistical significance. Individuals in the low-RMR group had a

Table 1. Baseline participant anthropometric characteristics overall and by the RQ group

All (N = 419) Low/mod RQ (n = 278) High RQ (n = 141) P-value between-group (mean ± s.d.) (mean ± s.d.) (mean ± s.d.) differences

Percent female (%) 51.3 51.1 51.4 0.94 Percent white (%) 66.6 68.0 63.8 0.02 Age (years) 27.6 ± 3.8 27.4 ± 3.9 28.0 ± 3.6 0.13 Body Mass Index (kg/m2) 25.6 ± 3.8 25.6 ± 3.8 25.7 ± 3.8 0.81 Height (cm) 171.1 ± 9.4 171.3 ± 9.7 170.6 ± 9.0 0.53 Weight (kg) 75.2 ± 13.8 75.3 ± 14.1 74.9 ± 13.1 0.76 FFM (kg) 53.9 ± 12.1 54.4 ± 12.4 52.8 ± 11.3 0.22 FM (kg) 21.3 ± 9.9 20.9 ± 10.0 22.0 ± 9.7 0.29 Body fat (%) 28.1 ± 10.9 27.5 ± 10.9 29.2 ± 10.9 0.16 Weight change, past year (kg) 0.1 ± 5.6 0.3 ± 5.5 − 0.3 ± 5.6 0.34 Percent with weight gain 44.5 kg, past 3 months (%) 10.8 12.3 7.8 0.16 Weight gain during baseline period (kg)a 0.3 ± 0.9 − 0.4 ± 0.9 0.0 ± 0.8 0.0002 Abbreviations: FFM, fat-free mass; FM, fat mass; RQ, respiratory quotient. aMean ± s.d. days to complete baseline measurements = 10.2 ± 8.2, no difference between groups.

European Journal of Clinical Nutrition (2016) 1197 – 1202 © 2016 Macmillan Publishers Limited, part of Springer Nature. Fat oxidation, metabolic rate, weight gain RP Shook et al 1199 lower RMR compared with individuals in the moderate/high-RMR with the low/moderate group. At the end of the 12-month follow- group, both measured (5707 ± 892 vs 6719 ± 1028 kJ/day, up period, the average unadjusted weight gain for all participants Po0.001) and predicted from the Harris–Benedict equation was 1.0 ± 3.7 kg, with no significant differences between the RQ (6059 ± 410 vs 6782 ± 533 kJ/day, Po0.001) and the Mifflin– groups (Table 4). However, individuals in the high-RQ group St. Jeor equation (5794 ± 552 vs 6657 ± 657 kJ/day, Po0.001; gained more than twice the amount of fat mass compared with Supplementary Table 2). In addition, measured RMR was lower the low/moderate-RQ group (1.3 ± 3.0 vs 0.6 ± 3.7 kg, P = 0.03). than predicted RMR in the low-RMR group using both equations There were no significant differences between baseline and (Harris–Benedict, Po0.001; Mifflin–St. Jeor, P = 0.06), whereas 12 months in the change of variables potentially predictive of there was no difference between measured and estimated RMR weight change between the RQ groups (energy intake, diet in the moderate/high group (Harris–Benedict, P = 0.13; Mifflin–St. composition, energy expenditure and the time spent in physical Jeor, P = 0.16). Given the agreement between the classification of activity). There were no differences in either body weight or fat low RMR using the tertile method and deviation from estimated mass gain in the low-RMR group vs the mod/high-RMR group RMR using equations, we conducted all further analyses using the (Supplementary Table 3). tertile method. After adjustment for differences in sex, age, race and change in Pearson's correlations were calculated for RQ and possible MVPA, energy intake and energy expenditure from baseline determinants based on variables previously identified in the between the low/moderate and the high-RQ groups, the high-RQ literature, both overall and by group (Table 3). Weak, but group had gained more body weight at each time point (Figure 1, statistically significant, associations were observed between 12-month weight gain difference 1.41 ± 0.25 vs 0.89 ± 0.19 kg, percent of kilocalories as carbohydrates and RQ, and baseline P = 0.05). The high-RQ group also gained more fat mass at each weight change and RQ, both overall (r = 0.16, P = 0.0015; r = 0.21, time point compared with the low/moderate-RQ group (Figure 1, Po0.0001) and for the low/moderate group (r = 0.18, P = 0.002; 12-month fat mass gain difference 1.47 ± 0.25 vs 0.62 ± 0.19 kg, r = 0.12, Po0.05) but not the high group (r = 0.10, P = 0.26; r = 0.10, P = 0.001). This difference was significant beginning at the first P = 0.24). follow-up period (month 3) and was maintained through month The following number of participants provided complete data 12. The results were similar when examined as relative change at each time point: 3 months = 402; 6 months = 374; 9 months = (percent) for both body weight and fat mass (results not 352; 12 months = 344. Participants in the high RQ at baseline had presented). In terms of RMR, there were no differences in changes significantly higher RQ values at month 6 (0.809 vs 0.789, in body weight (1.02 ± 0.24 vs 1.07 ± 0.21 kg, P = 0.85) nor in the fat Po0.0001) and month 12 (0.809 vs 0.797, P = 0.02) compared mass (1.05 ± 0.21 vs 0.73 ± 0.23, P = 0.24) between the low and

Table 2. Descriptive statistics overall and by the RQ group at baseline

All (N = 419) Low/mod RQ High RQ P-value (mean ± s.d.) (n = 278) (n = 141) between-group (mean ± s.d.) (mean ± s.d.) differences

Respiratory quotient 0.792 ± 0.045 0.766 ± 0.025 0.841 ± .032 o0.0001 Resting energy expenditure (kJ/day) 6385.2 ± 1093.2 6420.4 ± 1138.9 6315.7 ± 998.2 0.35 Resting metabolic rate/body weight (ml/kg per minute) 2.96 ± 0.36 2.99 ± 0.35 2.90 ± 0.35 0.014 Resting energy expenditure/FFM (kJ/kg per day) 120.6 ± 14.4 120.1 ± 14.0 121.7 ± 15.2 0.28 Cardiorespiratory ﬁtness (ml/kg per minute) 38.4 ± 9.8 39.0 ± 9.7 37.3 ± 9.9 0.10 Energy intake (kJ/day) 8718.2 ± 2837.5 8828.3 ± 3008.4 8500.9 ± 2462.4 0.23 Carbohydrates (% of total kJ) 47.2 ± 9.9 46.7 ± 10.0 48.4 ± 9.6 0.10 Fat (% of total kJ) 32.9 ± 7.8 32.9 ± 7.7 32.7 ± 7.3 0.79 Protein (% of total kJ) 17.2 ± 4.9 17.5 ± 5.3 16.6 ± 4.0 0.07 Alcohol (% of total kJ) 2.8 ± 4.5 3.0 ± 4.8 2.4 ± 3.9 0.17 Energy expenditure (kJ/day) 11474 ± 2142 11593 ± 2193 11237 ± 2023 0.11 MVPA (minutes/day) 77.8 ± 64.1 81.5 ± 64.3 70.3 ± 63.2 0.09 Abbreviations: FFM, fat-free mass; MVPA, moderate-to-vigorous physical activity. Mean armband wear time = 23.3 ± 0.6 h/day; mean dietary recalls completed = 2.81 ± 0.4.

Table 3. Correlation coefﬁcients between respiratory quotient and selected variables

Respiratory quotient

All (P-value) Low/mod RQ (P-value) High RQ (P-value)

FM (kg) 0.0495 (0.31) 0.0734 (0.23) − 0.0844 (0.32) FFM (kg) − 0.0481 (0.33) − 0.0521 (0.39) 0.0883 (0.30) Carbohydrates (% of total kJ) 0.1548 (0.002) 0.1841 (0.002) 0.0955 (0.26) MVPA (minutes/day) − 0.0734 (0.14) − 0.0489 (0.42) 0.0455 (0.60) Cardiorespiratory ﬁtness (ml/kg/min) − 0.0756 (0.12) − 0.1112 (0.062) 0.1239 (0.14) Energy expenditure (kJ/day) − 0.0442 (0.37) − 0.0329 (0.59) 0.1360 (0.18) Weight change, past year (kg) − 0.0765 (0.12) − 0.0596 (0.33) − 0.0702 (0.41) Baseline weight change (kg) 0.2095 (o0.0001) 0.1198 (0.05) 0.1003 (0.24) Abbreviations: FFM, fat-free mass; FM, fat mass; MVPA, moderate-to-vigorous physical activity.

Table 4. Changes in key variables between baseline and 12 months both overall and by the RQ group

All (N = 344) Low/mod RQ (n = 231) High RQ (n = 113) P-value between-group (mean ± s.d.) (mean ± s.d.) (mean ± s.d.) differences

Weight (kg) 1.0 ± 3.7 0.8 ± 3.8 1.4 ± 3.5 0.16 FM (kg) 0.8 ± 3.5 0.6 ± 3.7 1.3 ± 3.0 0.03 Body fat (%) 0.7 ± 3.4 0.4 ± 3.6 1.2 ± 2.9 0.04 Energy intake (kJ/day) − 36.8 ± 2525.8 − 58.6 ± 2513.0 7.5 ± 2559.5 0.82 Carbohydrates (% of total kJ) − 1.4 ± 10.0 − 1.2 ± 10.3 − 1.8 ± 9.2 0.63 Fat (% of total kJ) 0.4 ± 8.1 0.4 ± 6.2 0.8 ± 4.5 0.51 Protein (% of total kJ) 0.6 ± 5.6 0.4 ± 8.2 0.4 ± 7.8 0.96 Alcohol (% of total kJ) 0.4 ± 4.4 0.4 ± 4.2 0.6 ± 4.8 0.66 Energy expenditure (kJ/day) − 90.9 ± 1027.1 − 96.3 ± 1025.8 − 80.0 ± 1034.2 0.89 MVPA (minutes/day) − 6.6 ± 44.9 − 6.0 ± 47.8 − 7.9 ±38.3 0.68 Abbreviations: FM, fat mass; MVPA, moderate-to-vigorous physical activity, RQ, respiratory quotient.

Table 5. Linear mixed model results predicting change in fat mass from baseline to 12 months

Independent variable F-value P-value

Sex 0.34 0.56 Age 3.73 0.05 Race 0.89 0.49 Changes in MVPA 151.48 o0.0001 Changes in energy intake 8.25 0.004 Changes in energy expenditure 140.17 o0.0001 Baseline adjREE 1.05 0.31 Baseline RQ 4.82 0.03 Abbreviations: adjREE, Resting energy expenditure adjusted for fat-free mass and fat mass; MVPA, moderate-to-vigorous physical activity; RQ, respiratory quotient.

participants with a high RQ at baseline continued to have higher RQ values at both month 6 and 12, and changes in body weight and composition that were evident at the third month of follow- up were maintained for the remainder of the follow-up period. Overall, individuals with a high RQ in the present study experienced on average a nearly 0.9 kg greater increase in fat Figure 1. Changes in body weight and fat mass (adjusted mean ± s.e.) mass compared with those with a low or moderate RQ. In at each time point for each group after adjustment for gender, age, addition, a low RMR was not associated with gains in body weight race and changes in energy intake, energy expenditure and the over the same period. These findings suggest that lower levels of time spent in MVPA. Between-group differences for each time point: fat oxidation, independent of changes in energy intake, energy *Po0.05, **Po0.001. expenditure, macronutrient composition of the diet and physical activity, contribute to changes in body weight and fat mass, whereas lower energy expenditure from RMR does not. moderate/high groups after adjustment for the previously Previous research has been ambiguous regarding the role of mentioned covariates. Finally, linear mixed model was also fasting substrate oxidation on subsequent weight gain. In the performed for the cohort as a whole sample (that is, not stratified current study, we found that individuals in the top tertile for RQ by group) and the results indicate changes in MVPA, changes in (mean RQ = 0.841 ± 0.032) had larger gains in body weight and fat energy expenditure, changes in energy intake, and RQ at baseline mass after 12 months compared with individuals in the bottom were all associated with increases in fat mass (Table 5). Percent of two tertiles (Figure 1), confirming several previous examinations. kJ from carbohydrates, baseline weight change and CRF were Studies involving Pima Indians found RQ to be an independent added to a subsequent model but were not observed to be predictor of gains in both body weight (Po0.001) and fat mass associated with change in fat mass (F-value = 1.58, P = 0.21, (P = 0.004) at 25 months, suggesting that weight gain is a result of F-value = 1.31, P = 0.25 and F-value = 3.42, P = 0.07, respectively). reduced rates of fat oxidation.3 The Baltimore Longitudinal Study of Aging is the largest study to examine the role of RMR and RQ on weight gain in men (N = 775) over 10.3 years of follow-up.2 DISCUSSION After multivariate adjustments for age, BMI and fat-free mass, RQ was The primary finding of the current study is that a high baseline RQ significantly associated with weight change (β coefficient = 11.54, was predictive of gains in body weight and fat mass over Po0.001) but RMR was not (β coefficient = 0.05, P = 0.06). In a 12-month period among young adults. We have answered addition, those with an RQ of 40.85 (individuals with low rates of a critical unanswered question regarding temporality, as fat oxidation) were 2.42 times more likely to gain at least 5 kg

European Journal of Clinical Nutrition (2016) 1197 – 1202 © 2016 Macmillan Publishers Limited, part of Springer Nature. Fat oxidation, metabolic rate, weight gain RP Shook et al 1201 compared with those with an RQ of o0.76 (individuals with high groups did consume a higher percentage of their diet as rates of fat oxidation). A study of Italian women (N = 58) found carbohydrates; however, this difference was not statistically similar results, with those who gained 43 kg over a 3-year follow- significant (P = 0.1010). However, we did observe a weak but a up period having an RQ of 0.91 vs 0.84 of those who did not,1 and positive correlation with the percent of kilocalories from high RQ was significantly associated with gains in fat mass at both carbohydrates overall (Table 3, r = 0.1548, P = 0.002) but not 1 and 2 years among premenopausal women.4 The current study among the high-RQ group. None of the other previously cited supports these previous findings; that is, oxidation of energy determinants of RQ were found to be correlated in the present stores, not RMR, is predictive of weight gain. These results are in study (sex, fat mass, CRF). Protein oxidation was not measured in direct contrast to the Québec Family Study, which found no the current study, which limits our understanding of its role as a association of RMR or RQ on changes in body weight or fat during stimulant of energy expenditure, although previous research a 5.5-year follow-up.5 In that study, the correlations between suggests that it varies little between periods of overfeeding, measures of body weight/fatness (for example, weight, BMI or underfeeding or energy balance, except in extreme conditions.36 sum of skinfolds) and RMR were low (r = − 0.03 to 0.16, not Likewise, we did not measure thyroid hormone levels, another significant) or RQ (r = − 0.05 to 0.12, not significant). Neither RMR potential predictor of weight gain,37 in the current study, although nor RQ was a significant predictor of the increases in body weight we did exclude any individuals who were taking thyroid or fatness from Cox regressions. medications. Sympathetic nerve activity was also not measured, The variations in findings across studies are likely due to many which may be an independent determinant of both RQ38 and factors. The studies that have shown significant relationships RMR39 and a potential marker of future weight gain. between RMR and weight gain have consisted of young adult In summary, the current study has found that a high RQ is populations; it is likely that this effect in this age group is predictive of gains in body weight and fat mass over a 12-month explained by the fact that they are more likely to gain weight period among young adults when compared with individuals with compared with older adults.26 For example, the mean age of the a low/moderate or low-RQ value. A low RMR was not associated Pima Indians in a separate study was 26 years with 11.9% of with gains in body weight or fat mass over the same period. These participants gaining at least 10 kg over 4 years of follow-up.8 By findings support previous research, which suggests that lower contrast, the mean age of participants in the Baltimore Long- levels of fat oxidation, independent of changes in energy intake, itudinal Study of Aging was 49 years and weight gain was 0.07 kg energy expenditure, macronutrient composition of the diet and over a mean follow-up of 10.3 years, and the mean age in the physical activity, contribute to changes in body weight and fat Québec Family Study was 39 years and mean weight gain was mass. This study expands previous findings to document the 2.8 kg for men and 3.5 kg for women over 5.5 years. In the current temporal changes in both body composition and RQ, identifying study of young adults (mean age = 27.7 ± 3.8 years), the mean increases in fat mass as early as 3 months and persistent weight gain for all participants was 1.0 ± 3.4 kg after 12 months. elevations of RQ up to 12 months following initial assessment. There also are methodological differences across studies. For example, RMR was assessed in the Pima Indians using indirect calorimetry, whereas RMR in the Baltimore Longitudinal Study of CONFLICT OF INTEREST Aging was measured using multiple techniques over a 19-year This study was supported by an unrestricted research grant from the Coca-Cola period. Company. SNB received book royalties (o$5000 per year) from Human Kinetics, The determinants of elevated RQ are not well understood. The honoraria for service on the Scientific/Medical Advisory Boards for Technogym, studies involving Pima Indians found ~ 28% of the variability in RQ Santech, Cancer Fit Steps for Life and Clarity, and honoraria for lectures and fi values was due to family membership, and 18% was due to prior consultations from scienti c, educational and lay groups. During the past 5-year 3 period he has received research grants from the National Institutes of Health, change in weight, energy balance, body fat and sex, whereas 24- Department of Defense, Body Media and the Coca-Cola Company. In the last 3 years, h energy balance measured in a metabolic chamber was positively GAH has received funding from the National Institutes of Health, Health Resources associated with RQ in obese men and women (r = 0.55, Po0.01) in and Services Administration, American Heart Association, the Coca Cola Company 27 a separate study. Although we did not measure 24-h energy and TechnoGym. JRH was supported by an Established Investigator Award in Cancer balance, individuals in the low/moderate-RQ group did lose a Prevention and Control from the Cancer Training Branch of the National Cancer small amount of weight during the baseline data collection period Institute (K05 CA136975). The remaining authors declare no conflict of interest. (~10 days before RQ measurement), suggesting a negative energy balance, whereas the high-RQ group was weight stable. Linear ACKNOWLEDGEMENTS mixed modeling found no association between baseline weight change and weight change during the following 12 months. We thank the participants of the present study for volunteering their time and effort. More recent studies support the genetic influence on RQ In addition, we are grateful to the project staff for their tireless efforts with data fi levels28 and perhaps fitness and fat mass.29–33 Individuals with collection and study management, speci cally Patrick Crowley, Madison Demello, high levels of CRF and low levels of fat mass have been shown to Beth Lach, Reena Patel and Sarah Schumacher. be ‘metabolically flexible’, indicating that their skeletal muscle possesses the ability to switch between glucose and fat oxidation REFERENCES in response to homeostatic signals, such as after a meal or during fi 29 1 Marra M, Scal L, Covino A, Esposito-Del Puente A, Contaldo F. Fasting respiratory a dietary fast. When glucose is not present (that is, following an quotient as a predictor of weight changes in non-obese women. Int J Obes Relat overnight fast), a transition to reliance on fat oxidation occurs in Metab Disord 1998; 22: 601–603. 34 metabolically flexible individuals as seen in low RQ values. 2 Seidell JC, Muller DC, Sorkin JD, Andres R. Fasting respiratory exchange ratio and Among obese and unfit individuals with low metabolic flexibility, resting metabolic rate as predictors of weight gain: the Baltimore Longitudinal this transition in fuel source is less robust, and there is a blunted Study on Aging. Int J Obes Relat Metab Disord 1992; 16:667–674. preference on fat oxidation as seen in high RQ values.29 In the 3 Zurlo F, Lillioja S, Esposito-Del Puente A, Nyomba BL, Raz I, Saad MF et al. current study, it is possible that individuals in the low/moderate- Low ratio of fat to carbohydrate oxidation as predictor of weight gain: study of 259 – RQ group were more metabolically flexible because of higher 24-h RQ. Am J Physiol 1990; (5 Pt 1): E650 E657. 4 Ellis AC, Hyatt TC, Hunter GR, Gower BA. Respiratory quotient predicts fat mass levels of CRF and physical activity at baseline, although these 18 – fi gain in premenopausal women. Obesity (Silver Spring) 2010; : 2255 2259. differences did not reach statistical signi cance (P = 0.10 and 5 Katzmarzyk PT, Perusse L, Tremblay A, Bouchard C. No association between P = 0.09, respectively, Table 2). resting metabolic rate or respiratory exchange ratio and subsequent changes in Diet composition, specifically high levels of carbohydrate, have body mass and fatness: 5-1/2 year follow-up of the Quebec family study. Eur J Clin 35 also been shown to result in elevated RQ levels. The high-RQ Nutr 2000; 54:610–614.

© 2016 Macmillan Publishers Limited, part of Springer Nature. European Journal of Clinical Nutrition (2016) 1197 – 1202 Fat oxidation, metabolic rate, weight gain RP Shook et al 1202 6 Ravussin E, Bogardus C. Relationship of genetics, age, and physical fitness 23 Dwyer J, Ellwood K, Leader NP, Moshfegh AJ, Johnson CL. Integration of the to daily energy expenditure and fuel utilization. Am J Clin Nutr 1989; 49(5 Suppl): Continuing Survey of Food Intakes by Individuals and the National Health And 968–975. Nutrition Examination Survey. J Am Diet Assoc 2001; 101: 1142–1143. 7 Astrup A, Gotzsche PC, van de Werken K, Ranneries C, Toubro S, Raben A et al. 24 Hebert JR, Ebbeling CB, Matthews CE, Hurley TG, Ma Y, Druker S et al. Systematic Meta-analysis of resting metabolic rate in formerly obese subjects. Am J Clin Nutr errors in middle-aged women's estimates of energy intake: comparing three 1999; 69: 1117–1122. self-report measures to total energy expenditure from doubly labeled water. 8 Ravussin E, Lillioja S, Knowler WC, Christin L, Freymond D, Abbott WG et al. Ann Epidemiol 2002; 12:577–586. Reduced rate of energy expenditure as a risk factor for body-weight gain. N Engl J 25 West BT, Welch KB, Galecki AT. In: Linear Mixed Models: a Practical Guide Using Med 1988; 318: 467–472. Statistical Software. Chapman & Hall/CRC: Boca Raton, 2007. 9 Leibel RL, Rosenbaum M, Hirsch J. Changes in energy expenditure resulting from 26 Sheehan TJ, DuBrava S, DeChello LM, Fang Z. Rates of weight change for black altered body weight. N Engl J Med 1995; 332: 621–628. and white Americans over a twenty year period. Int J Obes Relat Metab Disord 10 Ravussin E, Lillioja S, Anderson TE, Christin L, Bogardus C. Determinants of 24-hour 2003; 27: 498–504. energy expenditure in man: methods and results using a respiratory chamber. 27 Toubro S, Christensen NJ, Astrup A. Reproducibility of 24-h energy expenditure, J Clin Invest 1986; 78: 1568–1578. substrate utilization and spontaneous physical activity in obesity measured in a 11 Hand GA, Shook RP, Paluch AE, Baruth M, Crowley EP, Jaggers J et al. The Energy respiration chamber. Int J Obes Relat Metab Disord 1995; 19:544–549. Balance Study: the design and baseline results for a longitudinal study of energy 28 Loos RJ, Ruchat S, Rankinen T, Tremblay A, Perusse L, Bouchard C. Adiponectin balance. Res Q Exerc Sport 2013; 84:1–12. and adiponectin receptor gene variants in relation to resting metabolic rate, 12 Branson RD, Johannigman JA. The measurement of energy expenditure. Nutr Clin respiratory quotient, and adiposity-related phenotypes in the Quebec Pract 2004; 19:622–636. Family Study. Am J Clin Nutr 2007; 85:26–34. 13 Henry CJ. Basal metabolic rate studies in humans: Measurement and develop- 29 Kelley DE. Skeletal muscle fat oxidation: timing and flexibility are everything. J Clin ment of new equations. Public Health Nutr 2005; 8:1133–1152. Invest 2005; 115: 1699–1702. 14 Weir JB. New methods for calculating metabolic rate with special reference to 30 Ukropcova B, McNeil M, Sereda O, de Jonge L, Xie H, Bray GA et al. protein metabolism. J Physiol 1949; 109:1–9. Dynamic changes in fat oxidation in human primary myocytes mirror metabolic 15 Weir JB. New methods for calculating metabolic rate with special reference to characteristics of the donor. J Clin Invest 2005; 115:1934–1941. protein metabolism. 1949. Nutrition 1990; 6:213–221. 31 Colberg SR, Simoneau JA, Thaete FL, Kelley DE. Skeletal muscle utilization of free 16 Roza AM, Shizgal HM. The Harris Benedict equation reevaluated: resting fatty acids in women with visceral obesity. J Clin Invest 1995; 95: 1846–1853. energy requirements and the body cell mass. Am J Clin Nutr 1984; 40: 32 Kunz I, Schorr U, Rommling K, Klaus S, Sharma AM. Habitual fat intake and basal 168–182. fat oxidation in obese and non-obese Caucasians. Int J Obes Relat Metab Disord 17 Mifflin MD, St Jeor ST, Hill LA, Scott BJ, Daugherty SA, Koh YO. A new predictive 2002; 26: 150–156. equation for resting energy expenditure in healthy individuals. Am J Clin Nutr 33 Astrup A, Buemann B, Western P, Toubro S, Raben A, Christensen NJ. Obesity as 1990; 51: 241–247. an adaptation to a high-fat diet: evidence from a cross-sectional study. Am J Clin 18 Troiano RP, Berrigan D, Dodd KW, Masse LC, Tilert T, McDowell M. Physical activity Nutr 1994; 59: 350–355. in the United States measured by accelerometer. Med Sci Sports Exerc 2008; 40: 34 Henriksson J. Muscle fuel selection: effect of exercise and training. Proc Nutr Soc 181–188. 1995; 54: 125–138. 19 St-Onge M, Mignault D, Allison DB, Rabasa-Lhoret R. Evaluation of a portable 35 Schutz Y. Abnormalities of fuel utilization as predisposing to the development of device to measure daily energy expenditure in free-living adults. Am J Clin Nutr obesity in humans. Obes Res 1995; 3(Suppl 2): 173S–178S. 2007; 85: 742–749. 36 Calles-Escandon J, Goran MI, O'Connell M, Nair KS, Danforth E Jr. Exercise 20 Welk GJ, McClain JJ, Eisenmann JC, Wickel EE. Field validation of the MTI increases fat oxidation at rest unrelated to changes in energy balance or lipolysis. Actigraph and BodyMedia armband monitor using the IDEEA monitor. Obesity Am J Physiol 1996; 270(6 Pt 1): E1009–E1014. (Silver Spring) 2007; 15:918–928. 37 Lowell BB, Spiegelman BM. Towards a molecular understanding of adaptive 21 Johannsen DL, Calabro MA, Stewart J, Franke W, Rood JC, Welk GJ. Accuracy of thermogenesis. Nature 2000; 404: 652–660. armband monitors for measuring daily energy expenditure in healthy adults. Med 38 Snitker S, Tataranni PA, Ravussin E. Respiratory quotient is inversely associated Sci Sports Exerc 2010; 42: 2134–2140. with muscle sympathetic nerve activity. J Clin Endocrinol Metab 1998; 83:3977–3979. 22 Thompson FE, Subar AF. Chapter 1—Dietary assessment methodology. In: Nutri- 39 Ravussin E. Low resting metabolic rate as a risk factor for weight gain: role of the tion in the Prevention and Treatment of Disease, 3rd edn. Academic Press: Oxford, sympathetic nervous system. Int J Obes Relat Metab Disord 1995; 19(Suppl 7): UK, 2013, pp 5–46. S8–S9.

Supplementary Information accompanies this paper on European Journal of Clinical Nutrition website (http://www.nature.com/ejcn)