Abstract
Background: Estimates of energy intake (EI) in humans have limited validity.
Objective: The objective was to test the accuracy and precision of the intake-balance method to estimate EI during weight gain induced by overfeeding.
Design: In 2 studies of controlled overfeeding (1 inpatient study and 1 outpatient study), baseline energy requirements were determined by a doubly labeled water study and caloric titration to weight maintenance. Overfeeding was prescribed as 140% of baseline energy requirements for 56 d. Changes in weight, fat mass (FM), and fat-free mass (FFM) were used to estimate change in energy stores (ΔES). Overfeeding EI was estimated as the sum of baseline energy requirements, thermic effect of food, and ΔES. The estimated overfeeding EI was then compared with the actual EI consumed in the metabolic chamber during the last week of overfeeding.
Results: In inpatient individuals, calculated EI during overfeeding determined from ΔES in FM and FFM was (mean ± SD) 3461 ± 848 kcal/d, which was not significantly (−29 ± 273 kcal/d or 0.8%; limits of agreement: −564, 505 kcal/d; P = 0.78) different from the actual EI provided (3490 ± 729 kcal/d). Estimated EI determined from ΔES in weight closely estimated actual intake (−7 ± 193 kcal/d or 0.2%; limits of agreement: −386, 370 kcal/d; P = 0.9). In free-living individuals, estimated EI during overfeeding determined from ΔES in FM and FFM was 4123 ± 500 kcal/d and underestimated actual EI (4286 ± 488 kcal/d; −162 ± 301 kcal or 3.8%; limits of agreement: −751, 427 kcal/d; P = 0.003). Estimated EI determined from ΔES in weight also underestimated actual intake (−159 ± 270 kcal/d or 3.7%; limits of agreement: −688, 370 kcal/d; P = 0.001).
Conclusion: The intake-balance method can be used to estimate EI during a period of weight gain as a result of 40% overfeeding in individuals who are inpatients or free-living with only a slight underestimate of actual EI by 0.2–3.8%. This trial was registered at clinicaltrials.gov as NCT00565149 and NCT01672632.
See corresponding editorial on page 744.
INTRODUCTION
In 2011–2012, more than 78 million adult Americans were obese (1). Obesity prevalence continues to rise and is projected at 42% of adult Americans by the year 2050 (2). Obesity ensues from a sustained imbalance of energy whereby energy intake (EI)4 exceeds energy expenditure (EE), which results in an energy surplus stored as fat. Although the jury is still out on the cause(s) of this energy imbalance, there is increasing consensus that increased EI outweighs lower levels of physical activity, and thereby reduced EE is the most significant determinant of weight gain in most people (3, 4). Understanding the magnitude of the hyperphagia or overeating during periods of weight gain will help researchers to develop carefully controlled intervention studies and evidenced-based policy for obesity prevention through a reduction in dietary intake.
Assessing EI is a major shortcoming in obesity research in humans. For decades, scientific studies have relied on instruments that obtain estimates of EI from a self-report, in the form of food diaries or dietary recall. Although self-reported estimates of EI are somewhat acceptable in individuals with specific training such as dietitians (5), for the most part, the self-reported data show no differences in EI between BMI classifications (6), which, if viewed alone, would suggest that EI is not a contributing factor for weight gain.
In the early 1980s, the use of doubly labeled water (DLW) for the assessment of EE in free-living individuals was developed and validated (7). DLW is now considered the gold-standard method for estimating EI in weight-stable individuals. Comparisons of EI measured by DLW and self-report clearly indicated that individuals grossly underreport EI by an estimated 18% (8, 9), and this underreporting is more pronounced in overweight and obese individuals (10, 11). Despite this well-known finding, many investigators still rely on subjective instruments to assess EI. The study of food intake behavior during weight gain is hindered by the lack of accurate methods to quantify EI in free-living conditions (12), leaving the need for validated methods to estimate EI during weight gain. Several groups have demonstrated the validity and utility of the intake-balance method [EI = total daily EE + change in body energy stores (ES)] to estimate EI in free-living individuals during weight stability or weight loss induced by dietary restriction (13–16), but a novel application of the intake-balance method is to estimate EI during weight gain induced by supervised overfeeding.
The aim of this work was therefore to assess the accuracy and precision of the intake-balance method to assess EI during a period of controlled overfeeding. We hypothesize that estimation of EE and deposition during controlled overfeeding would yield valid estimates of EI by using the intake-balance method. Furthermore, the EI estimated by the intake-balance method would more closely approximate food provision during overfeeding in individuals confined to a metabolic ward compared with those who were free-living.
SUBJECTS AND METHODS
Validation study I: weight gain induced by overfeeding in participants confined to a metabolic ward
This study was approved by the Pennington Biomedical Research Center Institutional Review Board (www.clinicaltrials.gov; NCT00565149). All volunteers provided written informed consent before participation.
Design
The parent study was a randomized controlled trial to test the effect of dietary protein on body composition and EE during overfeeding (17). Participants resided on a metabolic ward for a total of ∼10 wk. Participants achieved weight maintenance through caloric titration the first 13–25 d and consumed 140% of baseline energy requirements the remaining 8 wk (56 d). All food was prepared by the metabolic kitchen, and participants were required to be sedentary. Alcohol and caffeine were prohibited throughout the study.
Participants
Data from the 8 healthy individuals (5 men and 3 women) randomly assigned to the normal protein diet (∼15% dietary protein) were included in this study. Participants had to be aged 18–35 y and have a BMI (in kg/m2) between 19 and 30 at enrollment. Anthropometric and clinical characteristics of the participants are summarized in Table 1.
TABLE 1.
Characteristics of the study participants at baseline1
| Validation study I (inpatient)(n = 8) | Validation study II (outpatient)(n = 35) | |
| Sex [n (%)] | ||
| Male | 5 (63) | 29 (83) |
| Female | 3 (38) | 6 (17) |
| Race2 [n (%)] | ||
| Black | 7 (88) | 14 (40) |
| White | 0 (0) | 21 (60) |
| Other | 1 (12) | 0 (0) |
| Age (y) | 22 ± 3 (18–28)3 | 27 ± 5 (20–40) |
| Weight (kg) | 77.3 ± 13.8 (63.2–107.2) | 79.6 ± 10.2 (58.2–100.0) |
| BMI4 | 26.0 ± 2.9 (22.2–29.6) | 25.6 ± 2.3 (22.5–31.4) |
| Normal [n (%)] | 4 (50) | 20 (57) |
| Overweight [n (%)] | 4 (50) | 13 (37) |
| Obese [n (%)] | 0 | 2 (6) |
| Percentage fat | 24.9 ± 9.1 (11.7–39.2) | 21.8 ± 7.1 (10.4–36.6) |
| FM (kg) | 19.3 ± 7.0 (9.0–28.4) | 17.2 ± 5.2 (7.9–25.8) |
| FFM (kg) | 58.1 ± 12.9 (44.0–84.9) | 62.4 ± 10.6 (38.9–77.8) |
FFM, fat-free mass; FM, fat mass.
Self-reported.
Mean ± SD; range in parentheses (all such values).
Normal: 18.5–24.9 kg/m2; overweight: 25–30 kg/m2; obese: ≥30 kg/m2.
Determination of baseline energy requirements
The baseline energy requirement was defined by the EI needed to achieve weight stability (±1 kg) over a 10-d period on the ward. The 24-h EE measured in the metabolic chamber, multiplied by an activity factor of 1.15 to account for physical activity on a metabolic ward (18), was used to calculate the kilocalories for the initial weight-stabilizing diet. For participants with weight fluctuations ±1 kg over the first 3 d, EI was titrated up or down until weight stability was achieved. During this baseline period, participants completed a 9-d DLW study.
Energy intake during overfeeding
The overfeeding diet prescription was calculated as 140% of the baseline energy requirement. All meals were prepared by a metabolic kitchen and provided in a 5-d rotation. Individual diets were prepared in duplicate, composited, and analyzed for macronutrient content by Covance Laboratories. The analyzed diets provided 15% of energy from protein, 41% of energy from carbohydrate, and 44% of energy from fat. Participants were observed during mealtimes (+60 min) by the dietary staff to ensure that all foods were properly consumed.
Clinical assessments
Vital signs and weight were measured each morning following an overnight fast. Body composition was measured by dual-energy X-ray absorptiometry (DXA) at baseline and then every other week during overfeeding.
Validation study II: weight gain induced by overfeeding in free-living participants
This study was approved by the Pennington Biomedical Institutional Review Board (www.clinicaltrials.gov; NCT01672632). Volunteers provided written informed consent before participation.
Design
The second parent study used in this analysis was a prospective study to determine the effect of overfeeding on fat cell size and insulin sensitivity. Participants were provided with all meals during a baseline period (1 wk) and during the 56 d of overfeeding. They were outpatients, but they reported to the research kitchen at mealtimes where food was consumed under the supervision of a dietitian.
Participants
Thirty-five healthy, weight-stable individuals (29 men and 6 women) completed the study. Participants had to be aged 18–40 y and have a BMI between 22 and 32 at enrollment. Characteristics of the participants at baseline are shown in Table 1.
Determination of baseline energy requirements
Baseline energy requirements were calculated as the average between the estimates of EI from a prediction equation (19) and the EI consumed for weight maintenance during days 8–14 of a DLW study. During the first week of baseline and the DLW study (days 1–7), participants consumed a habitual diet in which the caloric intake was determined from a previously published equation for free-living, nonobese individuals (19). Beginning the second week (days 8–14), based on the total daily energy expenditure measured in days 1–7, EI was titrated up or down to weight stability with a diet consisting of 60% carbohydrate, 25% fat, and 15% protein.
Energy intake during overfeeding
The calorie intake of overfeeding was calculated as 140% of the baseline energy requirement. All meals were prepared by the metabolic kitchen and provided in a 5-d rotation with overfeeding kilocalories in proportion to baseline energy requirements. The macronutrient composition of the study diet was identical to that in validation study I: 15% of energy from protein, 41% from carbohydrate, and 44% from fat (as analyzed by Covance Laboratories). Participants were instructed to consume all food provided. Additional food (including alcohol) was permitted and reported by participants during overfeeding.
Clinic assessments
Weight was measured at screening, daily for 7 d during energy balance, before overfeeding, and after overfeeding. Body composition was measured by DXA before overfeeding (baseline), at week 5 of overfeeding, and during the last week of overfeeding.
Description of outcome measures common to both trials
Twenty-four-hour energy expenditure and respiratory quotient by indirect room calorimetry
Twenty-four-hour metabolic chamber stays were completed at baseline and during the last week of overfeeding. Oxygen and carbon dioxide levels in the chambers were measured by using a Magnos 4G magneto-pneumatic oxygen analyzer and a Uras 3G infrared CO2 analyzer (Hartmann and Braun), both of which sample O2 and CO2 concentrations 60 times per second. Every 10 s, a computer program averaged these values, calculated the volumes of O2 consumption and CO2 production, and plotted the average values at 10-min intervals. Energy expenditure and substrate oxidations (including respiratory quotient) were calculated from oxygen consumption, CO2 production, and 24-h urinary nitrogen excretion by using the equations established by Acheson et al (20).
Total daily energy expenditure by DLW
Total daily energy expenditure was measured according to a standard procedure at Pennington Biomedical Research Center as previously described (19). In brief, after 2 baseline urine samples were collected, participants consumed a cocktail containing 0.2 g/kg of total water of 1H218O (Cambridge Isotopes) and 0.115 g/kg of total body water of 2H2O (Isotec Inc). Isotope elimination rates of 2H and 18O (kH and ko, respectively) were measured in urine collected at 4.5 and 6 h and then daily for the next 9 d (study 1) and on days 7 and 14 (study 2) by using linear regression based on the isotopic enrichment. Modifications of Schoeller (21) equations by Racette et al (22) were used to calculate carbon dioxide production. Total daily energy expenditure was calculated by multiplying the rate of carbon dioxide by the energy equivalent of carbon dioxide based on the measured respiratory quotient in the metabolic chamber.
Body composition
Body composition was measured by DXA by using the Hologics QDR 4500A whole-body scanner. The scans were analyzed with QDR software version 11.1 (Hologics).
Estimated energy intake calculation
The estimated EI during the overfeeding period was calculated according to the intake-balance method with the baseline energy requirement summed with the estimated thermic effect of food (TEF) of the overfeeding diet and changes in energy deposition in fat-free mass (FFM) and fat mass (FM).
First, to account for the increase in TEF with consumption of the overfeeding diet, we computed the overfeeding TEF as 10% of the overfeeding kilocalories for all participants.
 |
Second, the daily change in ES in FFM and FM was calculated from DXA and metabolic weight measurements and divided by the number of days of overfeeding. The coefficients used for tissue gain were assumed to be 13.1 kcal/g of FM gained and 2.2 kcal/g of FFM gained, which reflect the energy content of the tissue and the energy cost of tissue synthesis (4).
 |
Estimates of EI were then calculated as the sum of the baseline energy requirement, overfeeding TEF, and changes in body energy stores (ΔES) over the 56-d study.
 |
Estimated EI (kcal/d) was then compared with the actual EI provided (kcal/d) during overfeeding. The actual EI includes kilocalories attributed to foods returned to the kitchen (weigh back) or, in the case of the free-living individuals, additional foods consumed as reported by the participants.
Although 7.4 kcal/g is a commonly used energy coefficient for weight (4), estimation of the change in energy storage from weight change requires an assumption of energy partitioning. To account for this, we calculated an energy coefficient for each participant:
 |
The average energy coefficient for weight across the participants, 8.4 kcal/g, also was used to estimate EI from the change in energy stores computed from the daily change in metabolic weight.
 |
Given that 7.4 kcal/g is a commonly used energy coefficient for ΔESweight, estimated EI was also calculated by using 7.4 kcal/g as the energy coefficient, and the results can be found in Supplementary Table 1 and Supplementary Figure 1 (under “Supplemental data” in the online issue).
Statistical analysis
All calculations and data analysis were performed by a biostatistician (HH) with SAS version 9.3 (SAS Institute). The Student's t test for paired samples along with Bland-Altman regression analysis (23, 24) were used to determine differences between estimated and actual EI. The Student's t test for paired samples was used to evaluate the validity of estimated EI compared with actual EI during overfeeding. Bland-Altman regression analysis was used to evaluate whether the mean difference varied across different amounts of actual EI (proportional bias). Data are reported as means ± SDs, and P < 0.05 was considered significant.
RESULTS
Validation study I: estimated EI with weight gain induced by overfeeding participants on a metabolic ward
Baseline characteristics
Although sex was balanced (5 men and 3 women), the participants were predominantly black. The mean BMI was 26.0 ± 2.9 and ranged from 22.2 to 29.6 (Table 1). The average weight gained during overfeeding was 6.3 ± 2.4 kg but varied considerably among the participants (4.1–11.0 kg). Most (56.0 ± 13.7%) of the excess weight was deposited in FM (3.5 ± 1.6 kg), but the FFM was also expanded (2.8 ± 1.5 kg).
Calorie intake and EE
EI required for baseline weight maintenance was 2443 ± 510 kcal/d and was significantly higher than total daily EE measured at baseline by DLW (2146 ± 557 kcal/d, P = 0.036). The actual average EI consumed during overfeeding was 3490 ± 729 kcal/d. The overfeeding diet thereby provided an additional 1027 ± 190 kcal/d or 139.5% of the baseline energy requirement.
Validation of estimating EI during overfeeding
The mean baseline energy requirement (2443 ± 510 kcal/d), when summed with the overfeeding TEF (105 ± 23 kcal/d) and ΔES from FFM and FM (913 ± 395 kcal/d), derived an estimated EI of 3461 ± 848 kcal/d during overfeeding (Table 2). Estimated EI during overfeeding by using ΔES from FFM and FM was not significantly different from actual intake during overfeeding (−29 ± 273 kcal/d or 0.8%; P = 0.78; limits of agreement: −564, 505 kcal/d). The Bland-Altman plot shows (Figure 1A) that the proportional bias was not significant across different amounts of actual intake (P = 0.27).
TABLE 2.
Calculations of EI1
| Validation study I (inpatient; n = 8) | Validation study II (outpatient; n = 35) | |
| Baseline energy requirement (kcal/d) | 2443 ± 510 (1906–3310) | 3046 ± 393 (2186–4078) |
| Actual overfeeding EI (kcal/d) | 3490 ± 729 (2760–4694) | 4286 ± 488 (3130–5183) |
| TEF (kcal/d) | 105 ± 23 (85–138) | 124 ± 17 (86–157.2) |
| ΔWeight (kg) | 6.3 ± 2.4 (4.1–11.0) | 7.2 ± 1.9 (2.3–10.7) |
| ΔESweight (kcal/d) | 935 ± 371 (604–1644) | 957 ± 259 (312–1428) |
| ΔFM (kg) | 3.5 ± 1.6 (2.3–7.2) | 4.1 ± 1.3 (0.8–7.0) |
| ΔFFM (kg) | 2.8 ± 1.5 (0.9–5.7) | 3.1 ± 1.4 (0.5–6.0) |
| ΔESFFM/FM (kcal/d) | 913 ± 395 (646–1830) | 954 ± 286 (215–1537) |
| Estimated overfeeding EI (kcal/d) | ||
| From ΔESFFM/FM | 3461 ± 848 (2754–4973) | 4123 ± 500* (3017–5454) |
| From ΔESweight | 3482 ± 874 (2596–4837) | 4126 ± 492* (3244–5409) |
Values are means ± SDs; ranges in parentheses. Student's t test was used for paired samples to evaluate the validity of estimated EI compared with actual EI during overfeeding. *P < 0.05. EI, energy intake; ES, energy stores; FFM, fat-free mass; FM, fat mass; TEF, thermic effect of food; Δ, change from baseline to day 56 of overfeeding.
FIGURE 1.
Comparison of actual EI consumed compared with EI calculated by the intake-balance method by using changes in energy stores from fat mass/fat-free mass (A) and body weight (B) in participants (n = 8) confined to a metabolic ward during a controlled overfeeding study for 56 d. Black circles indicate females, and gray circles indicate males. The dashed lines represent the mean and the mean ± 1.96 SD. A: The proportional bias was not significant across different amounts of actual intake (P = 0.27). B: The proportional bias was significant across different amounts of actual intake (P = 0.03). EI, energy intake.
Estimated EI during overfeeding by using ΔES from weight was not significantly different from actual intake during overfeeding (−7 ± 193 kcal/d or 0.2%; limits of agreement: −386, 370 kcal/d; P = 0.9). The proportional bias was significant across different amounts of actual intake (P = 0.03; Figure 1B).
Validation study II: overfeeding and weight gain in free-living participants
Baseline characteristics
This study population was predominantly male and white. The mean BMI was 25.6 ± 2.3 and ranged from 22.5 to 31.4 (Table 1). The average weight gained during overfeeding was 7.2 ± 1.9 kg but ranged considerably between the participants (2.3–10.7 kg). Similar to the inpatient study (validation study I), most (57.2 ± 14.8%) of the excess weight was stored in FM (4.1 ± 1.3 kg; 0.8–7 kg), but the FFM was also expanded (3.1 ± 1.4 kg; 0.5–6 kg).
Calorie intake and EE
The baseline energy requirement was 3046 ± 393 kcal/d and was not significantly different from the total daily EE measured during baseline by DLW (3054 ± 567 kcal/d, P = 0.42). The actual average EI provided during overfeeding was 4286 ± 488 kcal/d. This provided an additional 1240 ± 170 kcal/d during overfeeding or 140.3% of the baseline energy requirement.
Validation of estimating EI during overfeeding
As shown in Table 2, the mean baseline energy requirement (3046 ± 393 kcal/d), when summed with the TEF associated with overfeeding (124 ± 17 kcal/d) and ΔES from FFM and FM (954 ± 286 kcal/d), derived an estimated EI during overfeeding of 4123 ± 500 kcal/d. The estimated EI using ΔES from FFM and FM significantly (P = 0.003) underestimated actual EI by −162 ± 301 kcal/d or 3.8% (limits of agreement: −751, 427 kcal/d). The proportional bias was not significant across different amounts of actual intake (P = 0.70; Figure 2A). Estimating overfeeding EI by using ΔES from weight similarly significantly underestimated actual EI during overfeeding (−159 ± 270 kcal/d or 3.7%; limits of agreement: −688, 370; P = 0.001). The proportional bias was not significant across different amounts of actual intake (P = 0.65; Figure 2B).
FIGURE 2.
Comparison of actual EI consumed compared with EI calculated by the intake-balance method by using changes in energy stores from fat mass/fat-free mass (A) and body weight (B) in free-living participants (n = 35) during a controlled overfeeding study for 56 d. Black circles indicate females, and gray circles indicate males. The dashed lines represent the mean and the mean ± 1.96 SD. The proportional bias was not significant across different amounts of actual intake (A: P = 0.70; B: P = 0.65). EI, energy intake.
DISCUSSION
In this report, we used data derived in a landmark overfeeding study in which participants resided on a metabolic ward for 56 d and consumed 40% above baseline energy requirements (17). In this controlled feeding study, participants consumed all meals under supervision and consumed only foods that were prepared by the metabolic kitchen. The EI calculated from the intake-balance method underestimated actual EI by <1%. The use of weight alone to estimate the change in energy stores also underestimated actual EI but with improved accuracy and precision compared with estimated EI from FM and FFM.
In addition, we used data derived from an overfeeding study in which participants, while consuming foods provided by the metabolic kitchen, were free-living throughout the trial. Even in free-living individuals, the intake-balance method estimated EI within 4% of actual EI consumed. Although estimated EI in free-living individuals is statistically less than actual EI during overfeeding, the intake-balance method provides a far more accurate and objective assessment of EI than many self-report instruments, which underestimate actual EI by 15–20% (8, 9). We assumed that estimated EI of individuals with minimal and repeatable levels of physical activity (and hence EE) on a metabolic ward would be more accurate than those derived in free-living participants with variable activity. Although EI estimates cannot be directly compared because no 2 participants completed both studies, the data support this hypothesis. The EI measured by the intake-balance method more closely approximated food provision during overfeeding in those who were confined to a metabolic ward compared with individuals who were free-living.
Although 40% overfeeding is greater than what the general weight-gaining population consumes habitually, the use of clinical data from short-term inpatient and outpatient studies of 40% overfeeding allows a unique opportunity for researchers to test the effect of weight gain in a well-controlled environment during a short period that would more realistically occur over 6–12 mo. The analysis also indicates that if body weight is the only available measure for estimating changes in ES, estimates of EI can be likely derived within 4% of actual EI consumed, which again outperforms any assessment of EI from self-report (8). In our previous work in dietary restriction, we learned that point estimates of EI can be derived by using the intake-balance method (DLW and DXA) over both shorter (1 mo) and longer (3, 6, or 12 mo) durations (13, 22). This study in overfeeding, which supports the use of the intake-balance method in the same fashion, is a valuable tool for researchers and clinicians to objectively assess EI in their patients at 6- or 12-mo follow-up visits more accurately than self-report assessments, resulting in more reliable research findings and more effective weight management recommendations.
Even though we used gold-standard assessments for estimating the baseline energy requirement and changes in ES, the intake-balance method was not able to account for all calories consumed during the 2 independent overfeeding studies. The intake-balance method indeed outperforms any self-reported instrument. However, there are some limitations to consider, many of which probably involve inherent measurement error. For example, the DLW method, although the most accurate method for quantifying total daily EE in free-living individuals, has a CV of 5–7% in comparison with the gold-standard measure of EE by room calorimetry, which has a CV of 2–3% (21). It is important to note, however, that we did not just rely on the total daily EE measurement from the DLW study to determine the baseline energy requirement. Instead, we used a hybrid of approaches, but most important, the estimated energy requirement was tested for its ability to maintain weight stability in the participants at baseline over several days. Recognizing that many groups will not have the capability for DLW, it is noted that baseline energy requirements can also be calculated from numerous prediction equations for total daily EE (19, 25) or resting metabolic rate multiplied by an activity factor (26). The accuracy and precision of these methods have been tested for different populations and ethnic groups (27–30).
Although the observed difference between the EI estimated by the intake-balance method and the EI provided is small, some of this discrepancy might be explained by 2 assumptions made in the calculation of EI from the intake-balance method. Without a DLW study during the overfeeding period, the total daily EE component of our equation required some assumptions. First, the increased TEF of the overfeeding diets (∼140% of baseline energy requirements) was assumed to be 10% for all participants but likely varied slightly between individuals. Second, in using the baseline energy requirement, which is the EI for weight stability, including habitual activity (either inpatient in study 1 or outpatient in study 2), we assumed that the level of physical activity of the participants was maintained throughout the overfeeding period. Perhaps an actual measure of total daily EE might be argued to have provided a more accurate and precise estimation of EI in our study, but an improvement greater than 1–4% would be unlikely. This is owed to the importance of establishing the baseline energy requirement for weight stability in these kinds of studies (31). Our previous work in weight loss suggested that limitations for estimating EI are affected by the precision of measurements used to estimate changes in ES (13). For example, the repeatability of DXA to assess percentage of body fat is 1% (32). Therefore, when DXA is used to estimate changes in ES, a change <1% may be due not to true changes in energy stores but to measurement error. Although we are confident there were measureable changes in FM and FFM in all participants, our measurements, and thus calculation of estimated EI during overfeeding, is limited by the precision of the DXA.
It remains a goal of scientists to develop methods that can accurately estimate habitual EI in humans and with potential for widespread scalability. The cost of feeding individuals to energy balance obviously limits the use of the intake-balance method in clinical weight management programs. The recent development and validation of mathematical models from a culmination of studies with gold-standard measures of energy requirements (by DLW or caloric titration) and body composition (by DXA), however, allow estimates of EI in free-living individuals to be more widely available (33–35). These models, centered on the intake-balance method and theoretical estimates of changes in energy stores, have been conveniently extrapolated to user-friendly Web-based platforms and therefore have the potential to be used in clinical settings to provide data-driven counseling to patients on EI.
In summary, the intake-balance method can provide an accurate, precise, and objective assessment of EI during weight gain. This objective measure allows for a more truthful look at the causes of obesity and provides us with evidence-based clinical tools to treat this widespread epidemic.
Supplementary Material
Acknowledgments
We are indebted to Susan Mancuso for study coordination for both of these challenging study protocols, Courtney Brock as the research kitchen director, Jennifer Rood for mass spectroscopy oversight, and of course the study participants for making this body of work possible.
The authors’ responsibilities were as follows—LMR, ER, and GAB: designed the study; HH: conducted the analysis; LAG, LMR, and ER: interpreted the data; and LAG and LMR: prepared the manuscript. All authors edited and approved the final version. No conflicts of interest were reported.
Footnotes
Abbreviations used: DLW, doubly labeled water; DXA, dual-energy X-ray absorptiometry; EE, energy expenditure; EI, energy intake; ES, energy stores; FFM, fat-free mass; FM, fat mass; TEF, thermic effect of food; ΔES, change in energy stores.
REFERENCES
- 1.Ogden CL, Carroll MD, Kit BK, Flegal KM. Prevalence of obesity among adults: United States, 2011-2012. NCHS Data Brief 2013;(131):–8. [PubMed]
- 2.Hill AL, Rand DG, Nowak MA, Christakis NA. Infectious disease modeling of social contagion in networks. PLoS Comput Biol 2010;6:e1000968. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Swinburn B, Sacks G, Ravussin E. Increased food energy supply is more than sufficient to explain the US epidemic of obesity. Am J Clin Nutr 2009;90:1453–6. [DOI] [PubMed] [Google Scholar]
- 4.Tataranni PA, Harper IT, Snitker S, Del Parigi A, Vozarova B, Bunt J, Bogardus C, Ravussin E. Body weight gain in free-living Pima Indians: effect of energy intake vs expenditure. Int J Obes Relat Metab Disord 2003;27:1578–83. [DOI] [PubMed] [Google Scholar]
- 5.Champagne CM, Bray GA, Kurtz AA, Monteiro JB, Tucker E, Volaufova J, Delany JP. Energy intake and energy expenditure: a controlled study comparing dietitians and non-dietitians. J Am Diet Assoc 2002;102:1428–32. [DOI] [PubMed] [Google Scholar]
- 6.Horner NK, Patterson RE, Neuhouser ML, Lampe JW, Beresford SA, Prentice RL. Participant characteristics associated with errors in self-reported energy intake from the Women's Health Initiative food-frequency questionnaire. Am J Clin Nutr 2002;76:766–73. [DOI] [PubMed] [Google Scholar]
- 7.Schoeller DA, van Santen E. Measurement of energy expenditure in humans by doubly labeled water method. J Appl Physiol 1982;53:955–9. [DOI] [PubMed] [Google Scholar]
- 8.Hill RJ, Davies PS. The validity of self-reported energy intake as determined using the doubly labelled water technique. Br J Nutr 2001;85:415–30. [DOI] [PubMed] [Google Scholar]
- 9.Trabulsi J, Schoeller DA. Evaluation of dietary assessment instruments against doubly labeled water, a biomarker of habitual energy intake. Am J Physiol Endocrinol Metab 2001;281:E891–9. [DOI] [PubMed] [Google Scholar]
- 10.Bandini LG, Schoeller DA, Cyr HN, Dietz WH. Validity of reported energy intake in obese and nonobese adolescents. Am J Clin Nutr 1990;52:421–5. [DOI] [PubMed] [Google Scholar]
- 11.Schoeller DA. Limitations in the assessment of dietary energy intake by self-report. Metabolism 1995;44(suppl 2):18–22. [DOI] [PubMed] [Google Scholar]
- 12.Martin CK, Correa JB, Han H, Allen HR, Rood JC, Champagne CM, Gunturk BK, Bray GA. Validity of the Remote Food Photography Method (RFPM) for estimating energy and nutrient intake in near real-time. Obesity (Silver Spring) 2012;20:891–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.de Jonge L, DeLany JP, Nguyen T, Howard J, Hadley EC, Redman LM, Ravussin E. Validation study of energy expenditure and intake during calorie restriction using doubly labeled water and changes in body composition. Am J Clin Nutr 2007;85:73–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Racette SB, Das SK, Bhapkar M, Hadley EC, Roberts SB, Ravussin E, Pieper C, DeLany JP, Kraus WE, Rochon J, et al. Approaches for quantifying energy intake and %calorie restriction during calorie restriction interventions in humans: the multicenter CALERIE study. Am J Physiol Endocrinol Metab 2012;302:E441–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Schoeller DA. How accurate is self-reported dietary energy intake? Nutr Rev 1990;48:373–9. [DOI] [PubMed] [Google Scholar]
- 16.Schulz LO, Alger S, Harper I, Wilmore JH, Ravussin E. Energy expenditure of elite female runners measured by respiratory chamber and doubly labeled water. J Appl Physiol 1992;72(1):23–8. [DOI] [PubMed] [Google Scholar]
- 17.Bray GA, Smith S, de Jonge L, Xie H, Rood J, Martin C, Most M, Brock C, Mancuso S, Redman L. Effect of dietary protein content on weight gain, energy expenditure, and body composition during overeating. JAMA 2012;307:47–55. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Dériaz O, Tremblay A, Bouchard C. Non linear weight gain with long term over feeding in man. Obes Res 1993;1:179–85. [DOI] [PubMed] [Google Scholar]
- 19.Redman LM, Heilbronn L, Martin C, de Jonge L, Williamson D, deLany J, Ravussin E. Metabolic and behavioral compensations in response to caloric restriction: implications for the maintenance of weight loss. PLoS ONE 2009;4:e4377. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Acheson KJ, Schutz Y, Bessard T, Ravussin E, Jequier E, Flatt JP. Nutritional influences on lipogenesis and thermogenesis after a carbohydrate meal. Am J Physiol 1984;246:E62–70. [DOI] [PubMed] [Google Scholar]
- 21.Schoeller DA. Measurement of energy expenditure in free-living humans by using doubly labeled water. J Nutr 1988;118:1278–89. [DOI] [PubMed] [Google Scholar]
- 22.Racette SB, Schoeller D, Luke A, Shay K, Hnilicka J, Kushner R. Relative dilution spaces of 2H- and 18O- labeled water in humans. Am J Physiol 1994;267:E585–90. [DOI] [PubMed] [Google Scholar]
- 23.Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1986;1:307–10. [PubMed] [Google Scholar]
- 24.Bray GA, DeLany JP, Volaufova J, Harsha DW, Champagne C. Prediction of body fat in 12-y-old African American and white children: evaluation of methods. Am J Clin Nutr 2002;76:980–90. [DOI] [PubMed] [Google Scholar]
- 25.Pieper C, Redman L, Racette S, Roberts S, Bhapkar M, Rochon J, Martin C, Kraus W, Das S, Williamson D, et al. Development of adherence metrics for caloric restriction interventions. Clin Trials 2011;8:155–64. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Lin PH, Proschan MA, Bray GA, Fernandez CP, Hoben K, Most-Windhauser M, Karanja N, Obarzanek E. Estimation of energy requirements in a controlled feeding trial. Am J Clin Nutr 2003;77:639–45. [DOI] [PubMed] [Google Scholar]
- 27.Hasson RE, Howe CA, Jones BL, Freedson PS. Accuracy of four resting metabolic rate prediction equations: effects of sex, body mass index, age, and race/ethnicity. J Sci Med Sport 2011;14(4):344–51. [DOI] [PubMed] [Google Scholar]
- 28.Anderegg BA, Worrall C, Barbour E, Simpson KN, Delegge M. Comparison of resting energy expenditure prediction methods with measured resting energy expenditure in obese, hospitalized adults. JPEN J Parenter Enteral Nutr 2009;33:168–75. [DOI] [PubMed] [Google Scholar]
- 29.Frankenfield D, Roth-Yousey L, Compher C. Comparison of predictive equations for resting metabolic rate in healthy nonobese and obese adults: a systematic review. J Am Diet Assoc 2005;105:775–89. [DOI] [PubMed] [Google Scholar]
- 30.Patil SR, Bharadwaj J. Comparison of different methods to estimate BMR in adolescent student population. Indian J Physiol Pharmacol 2011;55:77–83. [PubMed] [Google Scholar]
- 31.Hall KD. Modeling metabolic adaptations and energy regulation in humans. Annu Rev Nutr 2012;32:35–54. [DOI] [PubMed] [Google Scholar]
- 32.Baracos V, Caserotti P, Earthman CP, Fields D, Gallagher D, Hall KD, Heymsfield SB, Muller MJ, Rosen AN, Pichard C, et al. Advances in the science and application of body composition measurement. JPEN J Parenter Enteral Nutr 2012;36:96–107. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Thomas DM, Navarro-Barrientos JE, Rivera DE, Heymsfield SB, Bredlau C, Redman LM, Martin CK, Lederman SA, M Collins L, Butte NF. Dynamic energy-balance model predicting gestational weight gain. Am J Clin Nutr 2012;95:115–22. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Thomas DM, Schoeller DA, Redman LA, Martin CK, Levine JA, Heymsfield SB. A computational model to determine energy intake during weight loss. Am J Clin Nutr 2010;92:1326–31. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Hall KD. Metabolism of mice and men: mathematical modeling of body weight dynamics. Curr Opin Clin Nutr Metab Care 2012;15:418–23. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.