Professional Documents
Culture Documents
Reinhardt 2015
Reinhardt 2015
Lower core body temperature and greater body fat are components of a human thrifty phenotype
Martin Reinhardt1,2; Mathias Schlgl1,3; Susan Bonfiglio1; Susanne B. Votruba1; Jonathan Krakoff1; Marie
S. Thearle1
1
Obesity and Diabetes Clinical Research Section, Phoenix Epidemiology and Clinical Research Branch
National Institute of Diabetes and Digestive and Kidney Diseases, National Institutes of Health,
Department of Health and Human Services, 4212 N. 16th Street, Phoenix, Arizona 85016
2
Department of Diagnostic and Interventional Radiology, University Hospital Leipzig, Liebigstrasse 20,
04103 Leipzig, Germany
3
Department of Geriatrics and Aging Research, University Hospital Zurich, Raemistrasse 101, 8091,
Zurich, Switzerland
Funding. This research was supported by the Intramural Research Program of the National Institutes of
Health, National Institute of Diabetes and Digestive and Kidney Diseases.
Abstract
energy expenditure (EE) decrease with fasting, is associated with less weight loss during caloric
restriction. In rodents, models of diet-induced obesity often have a phenotype including a reduced EE and
decreased core body temperature. We assessed whether a thrifty human phenotype associates with
Subjects/Methods: Data for this cross-sectional analysis were from 77 individuals participating in one of
two normal physiology studies while housed on our clinical research unit. 24-hour EE using a whole-
room indirect calorimeter and 24-hour core body temperature were measured during 24 hours each of
fasting and 200% overfeeding with a diet consisting of 50% carbohydrates, 20% protein and 30% fat.
Body composition was measured by dual X-ray absorptiometry. To account for the effects of body size on
EE, changes in EE were expressed as a percentage change from 24-hour EE (%EE) during energy
balance.
Results: A greater %EE decrease with fasting correlated with a smaller %EE increase with overfeeding
(r=0.27, p=0.02). The %EE decrease with fasting was associated with both fat mass and abdominal fat
mass, even after accounting for covariates (=0.16 [95% CI: -0.26, -0.06] %EE per kg fat mass,
p=0.003; =0.0004 [-0.0007, -0.00004] %EE per kg abdominal fat mass, p=0.03). In males, a greater
%EE decrease in response to fasting was associated with a lower 24-hour core body temperature, even
after adjusting for covariates (=1.43 [0.72, 2.15] %EE per 0.1C, p=0.0003).
Conclusion: Thrifty individuals, as defined by a larger EE decrease with fasting, were more likely to
have greater overall and abdominal adiposity as well as lower core body temperature consistent with a
Introduction
The energy expenditure response to both caloric restriction (CR) and overfeeding can vary
substantially between individuals (15); however, the meaningfulness of these differences have only
recently begun to be elucidated. In a group of 14 men, 24-hour energy expenditure (EE) changes during
fasting and overfeeding were associated such that those individuals with a greater EE decrease during
fasting had a smaller EE increase with overfeeding defining a potential thrifty phenotype. Conversely,
those individuals with a larger EE increase during overfeeding were more likely to maintain EE during
fasting consistent with a more spendthrift phenotype (1). We recently found that more thrifty obese
individuals defined by a greater EE decrease with fasting lost less weight in a carefully controlled
inpatient weight loss study (6). There are a number of rodent models of diet induced obesity with a
common, coalescing phenotype including decreased core body temperature (CBT) and/or cold
intolerance, and reduced EE (710). A human thrifty phenotype may have similar attributes.
sweating, non-shivering and shivering thermogenesis, and insulation, help maintain CBT within tightly
regulated limits. On average, variance of more than 0.2C from the 37C set point activates the
thermoregulatory system (11). It is unclear how much these thermoregulatory mechanisms may contribute
to an increased propensity for obesity (12). Since adipose tissue has an insulating effect in many large
mammals (13), it is possible that accumulation of adipose tissue is part of the compensatory strategy of
the human body to prevent heat loss in some individuals. Some authors have reported a relationship
between increased body insulation due to increased adipose tissue and a decreased EE response to both
cold exposure (14) or a mixed meal (14,15), and propose that decreased heat loss due to increased
insulation contributes to a lower energy cost of maintaining CBT (14). A small number of studies have
investigated the relationship of human obesity to 24-hour CBT and found no difference in CBT between
lean and obese individuals at thermoneutrality (16,17), during and after exercise, after a meal, or during
sleep (17). While individual variation in temperature during low-energy diets has been proposed as a
potential contributor to a thrifty metabolism (12), the relationship of CBT to body composition,
including fat and lean mass and abdominal adipose tissue, has not fully been explored. Nor is it known
how CBT and these body composition measures relate to the EE responses to fasting or overfeeding.
The current study was designed to identify whether differences in 24-hour CBT and body
composition are associated with the EE responses to overfeeding or fasting. We hypothesized that a
human thrifty phenotype defined by a greater EE decrease with fasting and lower EE response to
overfeeding would also have lower 24-hour CBT and greater adiposity.
Subjects
We screened 213 individuals, and 130 individuals met the initial eligibility criteria for admission
into one of two ongoing studies investigating the EE responses to caloric restriction and overfeeding
(NCT00687115), NCT00523627). These initial criteria included evidence of good health and no medical
conditions as assessed by history, physical, EKG and basic laboratory measures. Both studies included
admission to the clinical research unit (CRU) of the Obesity and Diabetes Clinical Research Section of
the National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK) in Phoenix, AZ and
measurement of EE during energy balance, overfeeding with identical diets and fasting while the subjects
were weight stable. A further 53 subjects were excluded during the inpatient phase because they either
left the study for personal reasons (n=13), were diagnosed with impaired glucose tolerance during the 75g
oral glucose tolerance test (OGTT) that occurred on day 4 of admission after an overnight fast (an
exclusion criterion for study NCT00523627, n=20), were diagnosed with diabetes mellitus during the
OGTT (an exclusion criterion for both studies, n=4), developed an acute illness (n=2), or did not have
available data for all three assessments, i.e. energy balance, fasting and overfeeding, included in this
analysis (n=14). Glucose tolerance status was assessed according to the American Diabetes Association
criteria (18). The majority (n=74) of subjects had normal glucose regulation. Data from 77 subjects
(NCT00687115 (n=17), NCT00523627 (n=60)) were included in the final analyses. Thyroid stimulating
hormone (TSH) was determined by the laboratory of the Phoenix Indian Medical Center on the day of
admission to the CRU (Dimension TSH Flex ELISA Immunoassay Module, Siemens Healthcare
Diagnostics, Inc, East Walpole, MA). Subjects represented a wide BMI range (17.8 50.0 kg/m2). All
subjects reported weight stability for the 6 months prior to admission. Subjects were weighed daily upon
first awakening and were asked not to engage in vigorous exercise for the duration of their stay. Upon
admission, volunteers were provided a standard, weight maintaining diet (WMD) with 50% carbohydrate,
30% fat, and 20% protein content. Initial weight maintaining energy needs were determined using unit
specific equations based on weight, height and sex (19). The caloric content of the WMD was adjusted as
needed to maintain a stable weight (1%). Volunteers were fully informed of the nature and purpose of
the study they participated in, and written informed consent was obtained before admission. Both
described (2). Breakfast was provided at 07:00, and the subjects entered the calorimeter about 30 minutes
after completion of breakfast. Three other meals were provided within the calorimeter at 11:00, 16:00 and
19:00. To more closely achieve energy balance during the baseline EE assessment, EE was measured
twice during eucaloric conditions. Energy intake during the first eucaloric respiratory chamber was based
on unit specific calculations (20), which is approximately 80% of the WMD to account for the reduction
in physical activity within the whole-room calorimeter. The 24h caloric content of the meals fed during
the second eucaloric measurement was equivalent to the EE calculated from the first assessment.
Because short-term adaptations to fasting or overfeeding are associated with both the amount of
voluntary weight loss during dieting (6) and involuntary weight change during free living conditions in
adults (21) we phenotypically typed individuals EE and CBT changes during 24h of overfeeding and 24h
of fasting. The second eucaloric measurement was considered to be the baseline assessment and was
doubled to determine the caloric intake of the overfeeding diet. The overfeeding diet contained an
identical macronutrient composition as the eucaloric assessments (50% carbohydrate, 30% fat, 20
protein), and was given for only 24h, during which EE was measured. All subjects also had an EE
assessment while fasting with no caloric intake from 19:00 the night before entering the calorimeter.
Subjects only consumed non-caloric, non-caffeinated beverages during this assessment. The overfeeding
diet and the fasting assessments were done a minimum of 3 days apart from one another. All energy
expenditure responses to overfeeding and fasting are expressed as the percent change from the EE during
energy balance (EEDiet of Interest EEenergy balance / EEenergy balance x 100) to control for EE differences due to
Whole body composition, including percent body fat, fat mass (FM) and fat free mass (FFM) was
previously described (16). Total abdominal fat mass (AF) and %AF (percent AF from whole body
weight) was determined from a region of interest (ROI), which was defined from the top of the liver down
to the intervertebral disc level of L5/S1, and segmented manually using the enCore 2003 software (GE
Ingestible sensors (CorTemp, HQ Inc., Palmetto, FL) that transmit temperature values every
minute to a data recorder worn around the volunteers waist were used to measure CBT. Data was
available from a subset of 59 subjects with valid CBT measures during the EE assessments. In female
subjects, the CBT assessments were not consistently synchronized with the follicular phase of the
menstrual cycle. Missing CBT data was due to capsules passing through the gastro-intestinal tract in less
than 16 hours or technical failure of either the sensors or the data recorder. Diurnal CBT included the
CBT between 08:00 to 23:00, which represents the time period when the thermic effect of food can be
observed. Twenty-four hour CBT and diurnal CBT values are expressed as the mean temperature during
the respective time periods. As CBT demonstrated a high intra-individual consistency, and to maximize
the use of all CBT measures acquired, we also created the combined mean CBT during all three diets for
each individual.
Statistical Analysis
Normally distributed data was described with means and standard deviations. Skewed data (AF,
BMI) was described with medians and interquartile ranges. Pearson correlation coefficients were used to
assess initial associations between continuous, normally distributed variables, and Spearman rank
correlation coefficients were used for skewed variables. Two sample t-tests were used to assess
differences between groups, and paired t-tests were used to assess for differences between diets within
individuals. The contributions of FM and AF to the percent EE changes during the dietary interventions
were assessed in separate linear regression models including age, sex, race, ambient temperature, and
whole body fat free mass as covariates. Effect sizes of multivariate linear regression models are given as
mean standard deviation and [95% confidence limits]. Because of differences in CBT between men and
women and because of menstrual cycle temperature variation in women, all CBT analyses after the initial
comparison by sex were limited to men only. The associations between plasma TSH concentrations and
CBT with the percent EE changes during overfeeding and fasting were assessed with linear regression
models as above but also including FM or AF as a covariate. In order to better understand the differences
between individuals, we categorized them into thrifty or spendthrift phenotypes defined by a median split
of the EE response to fasting (Table 2), as has been done previously (6). The assessments of differences
between FM and AF in thrifty and spendthrift individuals were adjusted for sex.
Results
Characteristics of the study population are shown in table 1. No differences were noted between
ethnicities (Native American, Hispanic, White, Black) in any of the measured variables.
Compared to EE during energy balance, the mean percent EE (%EE) decrease during fasting EE was
84.3% (p<0.0001), and the %EE increase with overfeeding was 9.75.2% (p<0.0001). Individuals with
a smaller %EE decrease in response to fasting had a greater %EE increase in response to overfeeding
(r=0.27, p=0.02) (Figure 1). When individuals were categorized as thrifty v spendthrift (Table 2), thrifty
individuals had greater FM (30.6 v 22.6 kg, p=0.01) and AF (7.72 v 5.52 kg; p=0.02) compared to
spendthrift individuals; these results were similar after limiting the dataset to men only (26.6 v 19.6 kg,
p=0.02; (6.61 v 4.78 kg; p=0.04). There was a negative correlation between FM, %body fat, AF, %AF
and the %EE response to fasting (r=0.36, p=0.001; r=0.35, p=0.002; =0.34, p=0.003, r=0.30,
p=0.009, respectively), and the %EE response to overfeeding (r=0.27, p=0.02; r=0.36, p=0.002;
=0.25, p=0.03, r=0.32, p=0.005, respectively), i.e. individuals with more total and abdominal fat mass
showed a greater decrease in EE during fasting and a smaller EE increase during overfeeding. After
accounting for covariates (age, sex, race, ambient temperature, and whole body fat free mass), FM and
AF were both independent determinants of the %EE decrease with fasting (=0.16%EE per kg, p=0.003,
[-0.26, -0.06]; =0.0004%EE per kg, p=0.03, [-0.0007, -0.00004], respectively) (Table 3, Figure 2).
Neither FM nor AF were independent predictors of the increase in %EE with overfeeding after
Associations between average CBT, the %EE responses to fasting and overfeeding, and body composition
Men (n=34) had a lower mean CBT than women (n=12) during energy balance (36.82 vs 37.17 oC,
p=0.0002). Similar differences were seen during fasting and overfeeding. Because of these differences
and because of differences in temperature in women due to the menstrual cycle, all further CBT analyses
were limited to men. In men, the mean CBT was greater during overfeeding compared to energy balance
(36.960.18 v 36.820.27C, p=0.02), and fasting (36.960.18 v 36.860.24C, p=0.003). There was no
difference between CBT during fasting and energy balance. The %EE decrease with fasting was
correlated with all measures of CBT including during fasting (r=0.36, p=0.03), energy balance (r=0.42,
p=0.01), overfeeding (r=0.56, p=0.0008), and the combined average CBT (r=0.59, p<0.0001) (Figure 3).
When individuals were categorized as thrifty or spendthrift, thrifty men had a lower combined average
CBT (36.80 v 36.96C; p=0.004) compared to spendthrift men. Even after adjusting for covariates, the
combined average CBT (=1.43%EE per 0.1C, p=0.0003, [0.72, 2.15]) (Table 3) was an independent
predictor of the %EE response to fasting. Results were similar if the individual CBT values during the
dietary interventions were used. The %EE increase with overfeeding was not associated with any measure
of CBT including during fasting (r=0.22, p=0.2), energy balance (r=0.24, p=0.16), overfeeding
(r=0.13, p=0.5), or combined average CBT (r=0.03, p=0.86), and CBT was not an independent predictor
of the %EE response to overfeeding after adjusting for covariates (all p>0.1). We found no independent
associations between FM, AF and CBT during any of the diets, nor between FM, AF and diurnal CBT.
There were no correlations between the %EE changes during overfeeding and fasting and the
corresponding changes in CBT. Plasma TSH levels did not correlate with the %EE response to fasting,
the response to overfeeding, nor with any CBT measures, in either men or the whole group.
Discussion
In a larger and more diverse population, this study confirmed the previous observation that
humans with a greater EE decrease in response to fasting have a smaller EE increase in response to
overfeeding and vice versa, i.e. thrifty and spendthrift phenotypes (1). Those individuals who were
classified as thrifty, based on a larger decrease in 24h EE with fasting, have increased adiposity (FM
adjusted for body size, i.e. FFM) as well as decreased CBT during all dietary interventions, consistent
There are a number of rodent models with a similar thrifty phenotype including relatively lower
energy expenditure, low CBT (or cold intolerance) and an increased propensity for obesity (8,9,2225).
Interestingly, these models are produced by a variety of genetic defects (8,9,2325). An association
between low body temperature (although not all have been 24 hour assessments) and human obesity has
been indicated in some (26,27), but not all, (16,17,28) studies. Our results add to this prior literature by
confirming that there are human individuals with a lower CBT and a more energy conserving phenotype
during fasting similar to phenotypes observed in the above mentioned rodent models. Twelve obese
individuals included in this analysis were also participants in a caloric restriction study (caloric restriction
occurred after these initial phenotypic profiling assessments) which demonstrated that more thrifty obese
individuals were less successful at losing weight during 50% caloric restriction over 6 weeks of carefully
controlled conditions (6). The integration of these subjects (BMI range: 30.050.0 kg/m2) into a
population with a wider BMI range (17.850.0 kg/m2) demonstrates that these individuals with obesity
were in fact positioned more towards the middle and thrifty end of the spectrum of EE responses to
fasting. Nevertheless, even within this small obese population, there was sufficient intra-individual
difference in the EE response to fasting to characterize more spendthrift individuals who lost more weight
To put our main findings into context, a 0.1C lower CBT was associated with a 1.4% greater
decrease in EE during fasting, the trait we used to define the thrifty phenotype. Fat mass and abdominal
fat were each negatively related to the EE decrease with fasting. On average, thrifty men had a 0.16C
lower CBT, 7.0 kg more FM and 1.7 kg more AF compared those with a spendthrift phenotype. However,
the cross-sectional study design does not allow conclusions about the directionality of our findings.
Whether more efficient EE responses to fasting and overfeeding predisposes a person to increased
adiposity, or whether adiposity modulates the EE responses to dietary extremes through its insulating
properties remains to be established. However, evidence demonstrating that the amount of voluntary
weight loss over longer periods in obese individuals (6) and future weight change during free living
conditions in adults (21) may be reflected in the short term energy expenditure adaptations to fasting or
overfeeding may support the former possibility. Because CBT is tightly controlled (11) through
homeostatic mechanisms, it seems plausible that inter-individual variations in CBT may impact the EE
responses to fasting rather than the other way around. Data from the Baltimore Longitudinal Study of
Aging indicate that men with body temperature below the median live longer (29), which makes it
tempting to speculate that a more energy-conserving (thrifty) phenotype can result in increased longevity.
Such a phenotype would also likely promote increased survival in times of food scarcity; however, with
the current food abundance associated with modern living conditions, a thrifty phenotype is more likely to
contribute to an increased predisposition to weight gain. The speculation that a more energy conserving
phenotype may increase lifespan is supported by prior reports of an association between increased 24h EE
Although they were both significant, the estimate of the relationship between the %EE decrease
with fasting and abdominal fat volume was multifold smaller than the estimate for whole body fat mass.
This is most likely because abdominal fat represents a smaller absolute amount of adipose tissue.
However, a past study has shown evidence of greater postprandial heat leakage across the abdominal wall
in un-insulated lean men compared to obese or lean men with artificial insulation placed around the
abdomen. These results were based on CBT measurements for two hours following the ingestion of a
large meal (15). In our study, neither whole body fat mass nor abdominal fat volume was associated with
CBT nor daytime CBT, with daytime CBT representing the period when diet induced thermogenesis is
apparent. Of note, we measured CBT for >16h versus the 2 postprandial hours evaluated in this previous
study (15). While more abdominal adipose tissue may provide an insulating layer that decreases heat
leakage across the abdominal wall acutely and especially during the immediate postprandial period, given
time, obese individuals likely redistribute the heat to augment heat release from the extremities, thereby
helping to maintain the individual temperature set point over 24 hours (3135). To fully understand the
presumably complex interplay of CBT and obesity in metabolically thrifty and spendthrift individuals,
Whether the combination of energy conserving traits in the thrifty phenotype is the result of
genetics (36), sympathetic nervous system responses (3739), prior weight loss attempts (40), differing
levels of physical fitness or individual hormonal responses is unclear. Plasma TSH levels did not correlate
with the EE responses to fasting and overfeeding or with CBT, however, we did not have measures of
other hormones including thyroxine or triiodothyronine. It is possible that our subjects had variable
amounts of brown adipose tissue (BAT), which may influence both human energy expenditure and
maintenance of CBT (41). The inpatient nature of the study combined with the high demand and limited
availability of the whole room indirect calorimeter made it impractical to synchronize the multiple CBT
measurements with the follicular phase of the menstrual cycle in women. However, the number of men in
the study provided a large enough sample size to explore the relationships between CBT and other
variables. Although we found lower CBT and increased FM in thrifty individuals, we did not observe a
correlation between FM and CBT. The relationship between CBT and FM is likely complex and probably
includes modifying factors as the accumulation of FM will provide insulation, thus potentially increasing
CBT slightly and therefore, obscuring a clear correlation between CBT and FM, which would have been
expected and could have possibly been shown with a larger sample size. Since our measurements of CBT
were limited to a subset of our subjects, additional studies will be needed to verify our results.
In conclusion, in this diverse population, we found that individuals with a larger decrease in
energy expenditure during fasting had a lower CBT and also had greater adiposity. Thus, body
temperature may be a further defining feature of the human thrifty phenotype and offer insight into
restriction or excess. It is possible that in the future CBT may be useful to identify an individuals
phenotype (thrifty or spendthrift), and thus, be harnessed to direct weight loss or weight maintenance
Acknowledgments. The authors thank the dietary, nursing, and technical staff of the National Institutes
of Health Clinical Unit in Phoenix, AZ, for their assistance. Most of all, the authors thank the volunteers
Authors Contributions. M.R. wrote the manuscript. M.R. and M.S.T. analyzed the data. M.S.T. and
S.B.V. designed the study. M.S. and S.B. contributed to the study design and collected data. M.R.,
M.S.T., M.S., S.B., J.K., and S.B.V. contributed to the interpretations of findings and commented on and
edited the drafts. M.S.T. is the guarantor of this work and, as such, had full access to all the data in the
study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
References
2. Thearle MS, Pannacciulli N, Bonfiglio S, Pacak K, Krakoff J. Extent and determinants of thermogenic
responses to 24 hours of fasting, energy balance, and five different overfeeding diets in humans. J
Clin Endocrinol Metab. 2013 Jul;98(7):27919.
3. Garrow JS, Warwick P, Blaza S, Ashwell M. PREDISPOSITION TO OBESITY. The Lancet. 1980 May
24;315(8178):11034.
4. Miller DS, Mumford P, Stock MJ. Gluttony 2. Thermogenesis in Overeating Man. Am J Clin Nutr.
1967 Nov 1;20(11):12239.
6. Reinhardt M, Thearle MS, Ibrahim M, Hohenadel MG, Bogardus C, Krakoff J, et al. A Human Thrifty
Phenotype Associated With Less Weight Loss During Caloric Restriction. Diabetes. 2015 May 11;
7. Klaus S, Mnzberg H, Trloff C, Heldmaier G. Physiology of transgenic mice with brown fat
ablation: obesity is due to lowered body temperature. Am J Physiol - Regul Integr Comp Physiol.
1998 Feb 1;274(2):R28793.
8. Lateef DM, Abreu-Vieira G, Xiao C, Reitman ML. Regulation of body temperature and brown
adipose tissue thermogenesis by bombesin receptor subtype-3. Am J Physiol - Endocrinol Metab.
2014 Mar 15;306(6):E6817.
9. Trayhurn P, Thurlby PL, James WPT. Thermogenic defect in pre-obese ob/ob mice. Nature. 1977
Mar 3;266(5597):602.
10. Abreu-Vieira G, Xiao C, Gavrilova O, Reitman ML. Integration of body temperature into the analysis
of energy expenditure in the mouse. Mol Metab [Internet]. [cited 2015 Apr 10]; Available from:
http://www.sciencedirect.com/science/article/pii/S2212877815000563
11. Kurz A. Physiology of thermoregulation. Best Pract Res Clin Anaesthesiol. 2008 Dec;22(4):62744.
12. Landsberg L, Young JB, Leonard WR, Linsenmeier RA, Turek FW. Is obesity associated with lower
body temperatures? Core temperature: a forgotten variable in energy balance. Metabolism. 2009
Jun;58(6):8716.
13. Young RA. Fat, Energy and Mammalian Survival. Am Zool. 1976 Sep 21;16(4):699710.
14. Jquier E, Gygax PH, Pittet P, Vannotti A. Increased thermal body insulation: relationship to the
development of obesity. J Appl Physiol. 1974 Jun 1;36(6):6748.
15. Brundin T, Thrne A, Wahren J. Heat leakage across the abdominal wall and meal-induced
thermogenesis in normal-weight and obese subjects. Metabolism. 1992 Jan;41(1):4955.
16. Heikens MJ, Gorbach AM, Eden HS, Savastano DM, Chen KY, Skarulis MC, et al. Core body
temperature in obesity. Am J Clin Nutr. 2011 May 1;93(5):9637.
17. Hoffmann ME, Rodriguez SM, Zeiss DM, Wachsberg KN, Kushner RF, Landsberg L, et al. 24-h Core
Temperature in Obese and Lean Men and Women. Obesity. 2012;20(8):158590.
18. American Diabetes Association. Diagnosis and classification of diabetes mellitus. Diabetes Care.
2013 Jan;36 Suppl 1:S6774.
19. Ferraro R, Boyce VL, Swinburn B, De Gregorio M, Ravussin E. Energy cost of physical activity on a
metabolic ward in relationship to obesity. Am J Clin Nutr. 1991 Jun;53(6):136871.
20. Ravussin E, Lillioja S, Anderson TE, Christin L, Bogardus C. Determinants of 24-hour energy
expenditure in man. Methods and results using a respiratory chamber. J Clin Invest. 1986
Dec;78(6):156878.
21. Schlgl M, Piaggi P, Pannacciuli N, Bonfiglio SM, Krakoff J, Thearle MS. Energy expenditure
responses to fasting and overfeeding identify phenotypes associated with weight change.
Diabetes. 2015 Jul 16;
22. Davis TRA, Mayer J. Imperfect Homeothermia in the Hereditary Obese-Hyperglycemic Syndrome of
Mice. Am J Physiol -- Leg Content. 1954 May 1;177(2):2226.
23. Jrgens HS, Schrmann A, Kluge R, Ortmann S, Klaus S, Joost H-G, et al. Hyperphagia, lower body
temperature, and reduced running wheel activity precede development of morbid obesity in New
Zealand obese mice. Physiol Genomics. 2006 Apr 1;25(2):23441.
24. Zhang X, Lam KSL, Ye H, Chung SK, Zhou M, Wang Y, et al. Adipose Tissue-specific Inhibition of
Hypoxia-inducible Factor 1 Induces Obesity and Glucose Intolerance by Impeding Energy
Expenditure in Mice. J Biol Chem. 2010 Oct 22;285(43):3286977.
25. Bal NC, Maurya SK, Sopariwala DH, Sahoo SK, Gupta SC, Shaikh SA, et al. Sarcolipin is a newly
identified regulator of muscle-based thermogenesis in mammals. Nat Med. 2012 Oct;18(10):1575
9.
26. Adam K. Human body temperature is inversely correlated with body mass. Eur J Appl Physiol. 1989
Mar 1;58(5):4715.
28. Rising R, Fontieille AM, Larson DE, Spraul M, Bogardus C, Ravussin E. Racial difference in body core
temperature between Pima Indian and Caucasian men. Int J Obes Relat Metab Disord J Int Assoc
Study Obes. 1995 Jan;19(1):15.
29. Shock NW, Others A. Normal Human Aging: The Baltimore Longitudinal Study of Aging. 1984 Nov
[cited 2015 May 19]; Available from: http://eric.ed.gov/?id=ED292030
30. Jumpertz R, Hanson RL, Sievers ML, Bennett PH, Nelson RG, Krakoff J. Higher energy expenditure in
humans predicts natural mortality. J Clin Endocrinol Metab. 2011 Jun;96(6):E9726.
31. LeBlanc J. Subcutaneous Fat and Skin Temperature. Can J Biochem Physiol. 1954 Jul 1;32(4):3548.
32. Frim J, Livingstone SD, Reed LD, Nolan RW, Limmer RE. Body composition and skin temperature
variation. J Appl Physiol. 1990 Feb 1;68(2):5403.
33. Aita M, Yoshizumi K. [The effects of environmental thermal condition on transitional skin
temperature of peripheral parts of human hands and feet during exercise]. Ann Physiol Anthropol
Seiri Jinruigaku Kenkykai Kaishi. 1994 Nov;13(6):4217.
34. Claessens-van Ooijen AMJ, Westerterp KR, Wouters L, Schoffelen PFM, van Steenhoven AA, van
Marken Lichtenbelt WD. Heat production and body temperature during cooling and rewarming in
overweight and lean men. Obes Silver Spring Md. 2006 Nov;14(11):191420.
35. Savastano DM, Gorbach AM, Eden HS, Brady SM, Reynolds JC, Yanovski JA. Adiposity and human
regional body temperature. Am J Clin Nutr. 2009 Nov 1;90(5):112431.
36. Bouchard C, Tremblay A, Desprs J-P, Nadeau A, Lupien PJ, Thriault G, et al. The Response to
Long-Term Overfeeding in Identical Twins. N Engl J Med. 1990;322(21):147782.
37. Saad MF, Alger SA, Zurlo F, Young JB, Bogardus C, Ravussin E. Ethnic differences in sympathetic
nervous system-mediated energy expenditure. Am J Physiol. 1991 Dec;261(6 Pt 1):E78994.
38. Tataranni PA, Young JB, Bogardus C, Ravussin E. A low sympathoadrenal activity is associated with
body weight gain and development of central adiposity in Pima Indian men. Obes Res. 1997
Jul;5(4):3417.
39. Rosenbaum M, Hirsch J, Murphy E, Leibel RL. Effects of changes in body weight on carbohydrate
metabolism, catecholamine excretion, and thyroid function. Am J Clin Nutr. 2000 Jun;71(6):1421
32.
40. Leibel RL, Rosenbaum M, Hirsch J. Changes in energy expenditure resulting from altered body
weight. N Engl J Med. 1995 Mar 9;332(10):6218.
41. Saito M. Brown Adipose Tissue as a Regulator of Energy Expenditure and Body Fat in Humans.
Diabetes Metab J. 2013 Feb;37(1):229.
Legend - Figure 1
Filled circles represent subjects with a wide BMI range (17.850.0) participating in study NCT00523627;
open triangles represent 12 obese individuals (BMI range: 30.050.0) that with evaluation prior to their
participation in a caloric restriction weight loss study (NCT00687115; details can be found here: (6)).
Legend - Figure 2
A and B. Correlation of fat mass, abdominal fat and percent change in 24 hour energy expenditure in
response to fasting (r=0.36, p=0.001; =0.34, p=0.003, respectively). C and D. Correlation of fat mass,
abdominal fat, and percent change in 24 hour energy expenditure in response to fasting after adjustment
for fat free mass, age, sex, race, and ambient temperature (=0.16%EE per kg, p=0.003, [-0.26, -0.06];
=0.0004%EE per kg, p=0.03, [-0.0007, -0.00004], respectively). Filled circles represent subjects with a
wide BMI range (17.850.0) participating in study NCT00523627; open triangles represent 12 obese
individuals (BMI range: 30.050.0) evaluated prior to participation in a caloric restriction weight loss
Legend Figure 3
Correlation between the EE response to fasting in men and average 24 hour core body temperature during
fasting (A. r=0.36, p=0.03, N=36), energy balance energy balance (B. r=0.42, p=0.01, N=34), overfeeding
(C. r=0.56, p=0.0008, N=32), and the combined average 24 hour core body temperatures during fasting,
Values are presented as mean SD. Differences between men and women were assessed by two sample
Students t-test *= p < 005; **= p <001; ***= p < 0.0001. Abbreviations: W = White; H = Hispanic;
NA = Native American; B = Black; BMI = Body mass index; FM = Fat mass; FFM = Fat free mass; AF
= Abdominal fat mass; EE = 24-hour energy expenditure; TSH = Thyroid stimulating hormone; CBT =
average 24 hour core body temperature. Combined average CBT are values from average CBT
Thrifty Spendthrift
N 39 (28 men; 11 women) 38 (30 men; 8 women)
Race 12W, 13NA, 7B, 7H 14W, 12NA, 6B, 6H
Age (years) 37.3 9.5 34.6 10.8
Body weight (kg) 87.3 16.8 80.2 18.7
2
BMI (kg/m2) 28.0 [25.4; 33.4] 25.8 [22.9; 27.9]
Body fat (%) 34.0 9.9** 26.8 12.5
FM (kg) 30.6 13.0* 22.6 14.1
FFM (kg) 56.8 10.3 57.6 11.5
2,
AF (kg) 7.1 [5.16; 10.37] * 4.5 [2.05; 7.33]
AF (% of body weight) 8.6 3.5* 6.4 4.0
EE (kcal) 3 2099 [2036; 2162]** 1992 [1930; 2055]
EE response to fasting (%) 1
11.4*** 4.6 2.4
EE response to overfeeding (%) 8.3 5.3* 11.0 4.7
Fasting glucose (mg/dl) 4 94.9 5.3** 90.4 4.8
2h glucose (mg/dl) 112.0 31.3 107.4 24.4
TSH (IU/mL) 1.56 0.77 1.55 0.95
o
Combined average CBT ( C ) 36.80 0.18 N=25** 36.96 0.15 N=19
In order to better understand the differences between individuals, we categorized them into thrifty or
spendthrift phenotypes 1 defined by a median split of the 24h-energy expenditure (EE) response to fasting.
Values are presented as mean SD, unless otherwise indicated. 2 Presented as medians and interquartile
range. Differences between men and women were assessed by two sample Students t-test *= p < 005;
**= p <001; ***= p < 0.0001. 3 EE = 24-hour energy expenditure during energy balance, results are
shown as least square means and [95% confidence limits], adjusted for differences in age, sex, race, FM,
FFM, and ambient temperature. 4 Adjusting for differences in percentage of body fat did not alter the
Abbreviations: W = White; H = Hispanic; NA = Native American; B = Black; BMI = Body mass index;
FM = Fat mass; FFM = Fat free mass; AF = Abdominal fat mass; TSH = Thyroid stimulating hormone;
AF -0.0004 (0.03)
FFM (kg) [-0.14, 0.16] [-0.15, 0.13] [-0.08, 0.22] [-0.11, 0.20]
Age (yrs) [-0.13, 0.07] [-0.17, 0.04] [-0.08, 0.19] [-0.08, 0.19]
Race
Ambient Temp (oC) [-0.47, 0.93] [-0.22, 1.22] [-0.77, 0.76] [-0.66, 0.87]
Models 1 and 2 include the whole study population (men and women). Models 3 and 4 include CBT
measurements and were limited to men because of differences in CBT between men and women and
because of menstrual cycle temperature variation in women. Abbreviations: FM = Fat mass; FFM = Fat
free mass; AF = Abdominal fat mass; CBT = Average combined 24 hour core body temperature (values
from average CBT measurements during energy balance, fasting and overfeeding).
-1 4 9 14 19 24
0
Percent change EE during
-5
fasting
-10
-15
-20
Percent change EE during overfeeding
Filled circles represent subjects with a wide BMI range (17.850.0) participating in study NCT00523627;
open triangles represent 12 obese individuals (BMI range: 30.050.0) that with evaluation prior to their
participation in a caloric restriction weight loss study (NCT00687115; details can be found here: (6)).
60
50
20
10
0
-20 -18 -16 -14 -12 -10 -8 -6 -4 -2 0
B Percent change EE during fasting (%)
20
18
16
A and B. Correlation of fat mass, abdominal fat and percent change in 24 hour energy expenditure in
response to fasting (r=0.36, p=0.001; =0.34, p=0.003, respectively). C and D. Correlation of fat mass,
abdominal fat, and percent change in 24 hour energy expenditure in response to fasting after adjustment
=0.0004%EE per kg, p=0.03, [-0.0007, -0.00004], respectively). Filled circles represent subjects with a
wide BMI range (17.850.0) participating in study NCT00523627; open triangles represent 12 obese
individuals (BMI range: 30.050.0) evaluated prior to participation in a caloric restriction weight loss
0
Percent change EE during
-2
-4
-6
fasting
-8
-10
-12
-14
-16
36.3 36.8 37.3
0
Percent change EE during
-2
-4
-6
fasting
-8
-10
-12
-14
-16
35.8 36 36.2 36.4 36.6 36.8 37 37.2 37.4
-8
-10
-12
-14
-16
36.6 36.8 37 37.2 37.4
Average core body temperature during
C
overfeeding (oC)
0
Percent change EE during
-2
-4
-6
fasting
-8
-10
-12
-14
-16
36.3 36.5 36.7 36.9 37.1 37.3 37.5
D Average combined core body temperature during dietary
interventions (oC)
Correlation between the EE response to fasting in men and average 24 hour core body temperature during
fasting (A. r=0.36, p=0.03, N=36), energy balance energy balance (B. r=0.42, p=0.01, N=34), overfeeding
(C. r=0.56, p=0.0008, N=32), and the combined average 24 hour core body temperatures during fasting,