Metabolic adaptation after combined resistance and aerobic exercise training in older women

Abstract

Objective:

To investigate if combined aerobic and resistance training in older women leads to metabolic adaptation.

Methods:

80 women (64 whites, BMI: 30.0±4.4 kg/m², age: 64.8±3.5 years) followed 32 weeks of aerobic and resistance training. Body weight/composition (dual-energy X-ray absorptiometry) and RMR (indirect calorimetry) were measured at baseline, weeks 16 and 32. Metabolic adaptation was defined as significantly lower measured vs predicted RMR. A regression model to predict metabolic adaptation was developed including race, age, baseline fat-free mass (FFM), RMR and respiratory quotient (RQ), and changes in net submaximal oxygen consumption (VO₂) after different tasks.

Results:

There was significant metabolic adaptation at week 16 (−59±136 kcal/day, P=0.002), following a 640 kcal/week energy loss (−0.7±2.6 kg weight loss). In 53 women with complete data, metabolic adaptation was seen both at weeks 16 (−64±129 kcal/day, P=0.001) and 32 (−94±127 kcal/day, P<0.001). Metabolic adaptation at week 16 was predicted by race, age, baseline FFM, RMR and RQ, and change in net VO₂ of walking (R² adjusted = 0.90, P<0.001). Similar results were seen at week 36.

Conclusion:

In older women with overweight and obesity, minimal energy deficit induced by aerobic and resistance exercise is associated with metabolic adaptation at the level of RMR.

Introduction

The phenomenon of metabolic adaptation, also known as adaptive thermogenesis, in response to weight loss, has been one of the most controversial issues in the obesity field (1–7). It has now been established that metabolic adaptation does occur after weight loss and that its magnitude is modulated by the energy balance (EB) status of the individual (8), being minor (50 kcal/day) or non-existent when measurements are taken under conditions of weight stability (8, 9). Even though the role of metabolic adaptation as a risk factor for weight regain has been dismissed (8–11), metabolic adaptation has been shown to be associated with less weight and fat mass (FM) loss after low-energy diets (12), and a longer time to achieve weight loss goals (13).

The majority of the studies on metabolic adaptation have used diet alone interventions (8, 9, 14–18), or more seldom a combination of diet and exercise (10, 19). From our knowledge only two studies have looked at exercise alone and both reported no metabolic adaptation at the level of resting metabolic rate (RMR) in response to supervised exercise (20, 21). However, both studies included aerobic exercise only and were performed in relatively young populations (20, 21).

Aging is associated with decreases in RMR and fat free mass (FFM), as well as changes in the RMR-FFM relationship, with a greater contribution of low-metabolic rate (skeletal mass and bone) compared with high-metabolic rate (heart, liver, kidneys and brain) organs and tissues (22, 23). Given the increased prevalence of obesity, and aging, it is important to understand if exercise-induced weight loss in older adults, including resistance training, known to help to preserve FFM and optimize physical function in this patient group (24), is associated with metabolic adaptation. Therefore, the aim of this secondary analysis was to investigate the existence (or not) of metabolic adaptation, at the level of RMR, following combined aerobic and resistance training in older women, and to explore potential variables predictive of that response. We hypothesized that even a small energy deficit induced by exercise leads to metabolic adaptation in older women.

Methods

Participants

Eighty women, 60–74 yr old, participated in a 32-week combined aerobic and resistance exercise training program at the Department of Nutrition Sciences, University of Alabama at Birmingham (UAB). All participants were healthy, free of any metabolic disorders, not taking medications that may affect energy expenditure, nonsmokers, and self-reported to exercising less than once per week for the past year). Institutional review board approved informed consent was obtained before participation in the study in compliance with the Department of Health and Human Services Regulations for Protection of Human Research Subjects.

Study design

This is a secondary analysis of prior work designed to evaluate the cardiometabolic effects of 16 and 32 weeks of combined aerobic and resistance training. Participants were randomly assigned to one of three training groups, with different exercise frequency of resistance and aerobic training/week (1 time/week, 25 subjects; twice/week, 32 subjects; or three times/week, 23 subjects) for 32 weeks and asked not to change their diet while in the study. For detailed information about the study, see Hunter et al (25). For the purpose of this study all groups were analyzed together.

Briefly, the exercise program consisted of: 1) 40 minutes of cycle ergometer or treadmill exercise at 80% maximum heart rate and 2) resistance training with strength exercises including leg press, squats, leg extension, leg curl, elbow flexion, lateral pull-down, bench press, military press, lower back extension, and bent leg sit-ups (two sets of 10 repetitions with a 1.5- to 2-min rest between sets, starting at 60% of the maximum weight the subject could lift at one time (one-repetition maximum [1RM]) and was then gradually increased to 80% of 1RM at week 8. All exercise sessions were supervised, and participants had to wear heart rate monitors in order to ensure that endurance exercise was performed at the right intensity. Participants randomized to the once and twice/week exercise could choose to do the aerobic and resistance training on the same day or on different days of the week. In the group randomized to exercise 3 times/week, the order of the exercise training (endurance and resistance) was alternated between sessions.

Data collection

The following measurements were conducted at baseline, after 16 wk, and after 32 wk of the combined resistance and aerobic training program.

Body weight and composition

Body weight, FM and FFM were measured with dual-energy X-ray absorptiometry (DXA) (Lunar DPX-L densitometer; LUNAR Radiation, Madison WI), with the use of the Software version 1.33, (Lunar Corp). We converted losses of FM and FFM (from DXA) to energy lost (i.e., kilocalories lost), using energy coefficients of 9.3 kcal/g for FM and 1.1 kcal/g for FFM lost (kilocalories lost = [9.3 kcal × ΔFM(g)] + [1.1 kcal × ΔFFM(g)]). For participants accruing FFM, an energy coefficient of 1.8 kcal/g was used (26, 27): total kilocalories lost when FFM was gained = [9.3 kcal × ΔFM (g)] + [1.8 kcal × ΔFFM(g)].

Resting metabolic rate

RMR was measured, after an overnight fast (between 6:30–7:30AM), at least 48h after the last exercise session. O₂ consumption and CO₂ production were measured continuously for 40 minutes through open-circuit, indirect calorimetry (DeltaTrac II: Sensor Medics, Yorba, CA) and the final 20 minutes used to calculate RMR.

VO_2max and exercise economy

To measure VO_2max, a graded exercise test consistent with a modified-Balke protocol (Max-1 Cart; Physio-Dyne Instrument Corporation, Quogue, NY) was performed on a treadmill to voluntary exhaustion. Heart rate was continuously measured using a 12-lead electrocardiogram. The highest 20-second VO₂ was considered VO_2max (mL/kg/min). On a separate day, submaximal VO₂ was obtained in the steady state during the third and fourth minutes of 4 standardized exercise tasks. The 4 tasks, selected to reflect typical activities of adult women under free-living conditions, were level walking (0% grade, 4.8 km/h, 4 min), grade walking (2.5% grade, 4.8 km/h, 4 min), stair climbing (17.8-cm step, 60 steps/min, 4 min), and level walking carrying a loaded box (0% grade, 3.2 km/h, 4 min). The weight of the box was equivalent to 30% of the subject’s maximal isometric elbow flexion strength and was intended to simulate carrying a small load. Net VO₂ was calculate by subtracting resting VO₂ from steady-state VO₂. For more details see Weinsier et al 2000 (18).

Statistics

Statistical analysis was performed with SPSS version 22 (SPSS Inc., Chicago, IL), data presented as mean ± SD and statistical significance set at P<0.05. Changes in body weight/composition and RMR between baseline and week 16 were assessed with paired sample t-tests. Changes in body weight/composition and RMR over time (baseline, weeks 16 and 32), in a subgroup of participants with data at all timepoints, were assessed with a repeated-measures ANOVA, using Bonferroni correction for multiple comparisons. The presence of metabolic adaptation was tested by paired t-tests, comparing measured RMR (RMRm) and predicted RMR (RMRp) at the same time points.

An equation to predict RMR was derived from baseline data of all the study participants (n=103).

Model: RMRp (kcal/day) = 774.122 − (9.212 × Age (years)) − (96.263 × Race (0 for whites and 1 for blacks)) + (4.5451 × FM (kg)) + (24.593 × FFM (kg)).

R² = 0.55; P<0.001

This small R² is due to the study design of the study in which a very narrow range of age, and only women, were included. Exercise group was not a significant predictor of baseline RMR and therefore was not included in the final model. There was no multicollinearity among the independent variables included in the model (variance inflation factor <1.3). The residuals of the regression model were normal distributed.

Correlation analysis was performed between metabolic adaptation (both at weeks 16 and 32) and age, weight, FM at baseline, as well as weight, FM and FFM loss using Spearmen correlation, as the variables were non-normal distributed. The association between metabolic adaptation (at both weeks 16 and 32) and FFM at baseline was tested using Person correlation, as the variables were normal distributed. A regression model to predict metabolic adaptation, at both week 16 and 32, was constructed using race, age, baseline RMR, FFM, and RQ, and change in net VO₂ during walking. There was no multicollinearity among the independent variables included in the model (variance inflation factor <2.5). The residuals of the regression model were normal distributed.

A one-way ANOVA showed no differences in metabolic adaptation between exercise groups, at either wks 16 or 36, and exercise group was not a significant predictor of metabolic adaptation.

Results

Eighty women (64 whites) with an average BMI of 30.0±4.4 kg/m², an age of 64.8±3.5 years and a maximum aerobic capacity of 23.6±4.7 ml/kg/min were included in the present analysis. Body weight/composition, and RMR at baseline and week 16 can be seen in Table 1. There was a significant reduction in body weight (−0.7±2.6 kg, P=0.018) and FM (−1.2±2.6 kg, P<0.001) and a significant increase in FFM (0.7±1.4 kg, P>0.001) between baseline and week 16. There was also a significant reduction in RMRm between baseline and week 16 (−44±130 kcal/day) and RMRm at week 16 was significantly lower than RMRp (−59±136 kcal/day, P=0.002). There was a cumulative average energy loss of 10262±23549 kcal (641 kcal/week) at the end of the first 16 weeks of the exercise program.

Table 1.

Anthropometrics, cardiovascular fitness and RMR at baseline and after 16 weeks of exercise

	Baseline (n=80)	Week 16 (n=80)	P value
Weight (kg)	73.6±11.7	72.9±11.0	0.018
FM (kg)	31.2±8.8	30.0±8.2	<0.001
FFM (kg)	38.8±4.2	39.5±4.4	<0.001
VO2max (ml/kg/min)	22.8±4.5	23.8±4.8	0.005
RMRm (kcal/day)	1251±186	1208±174	0.004
RMRp (kcal/day)	1255±132	1267±134	0.003
RMRm-p (kcal/day)		−59±136 ***

Data shown as mean±SD. FM: fat mass; FFM: fat-free mass; RMR: resting metabolic rate; RMRm: RMR measured; RMRp: RMR predicted.

^***P=0.002 for the comparison between RMRm-p.

Anthropometrics, cardiovascular fitness, RMRm, and RMRp, as well as metabolic adaptation (RMRm – RMRp), over time in those women with data at all three time points (n = 53) can be seen in Table 2. Women in this subsample had an average weight loss (compared with baseline) of −1.0±2.7 kg (P=0.032) and −1.7±4.0 kg (P=0.011), at wks 16 and 32, respectively. No significant changes in body weight were seen between wks 16 and 32 (−0.7±, P=0.064). A similar pattern was seen for FM (kg), while FFM (kg) was significantly higher both at wks 16 and 32, compared with baseline. RMRm decreased by approximately 45±103 kcal/day at wks 16 (P=0.007), and 74±111 kcal/day at wks 32 (P<0.001), compared with baseline, with no significant changes seen between wks 16 and 32 (−29±90, P=0.072). RMRm was significantly lower than RMRp both at wks 16 and 32 (−64±129 kcal/day, P=0.001 and −94±127 kcal/day, P<0.001, respectively) and there was a significant increase in metabolic adaptation between weeks 16 and 32 (−30±85 kcal/day, P=0.013). There was a cumulative average energy loss of 17771±35544 kcal (555 kcal/week) at the end of the 32 weeks of the exercise program, representing a significant increase from week 16 (12192±25197 vs 177771±35544 kcal, at week 16 and 32, respectively (P=0.042)).

Table 2.

Anthropometrics, cardiovascular fitness and RMR at baseline and after 16 and 32 weeks of exercise in those with complete data at all timepoint

	Baseline (n=53)	Week 16 (n=53)	Week 32 (n=53)
Weight (kg)	73.1±12.6ab	72.1±11.2a	71.4±10.4b
FM (kg)	30.6±9.8cd	29.1±8.7c	28.6±7.8d
FFM (kg)	38.6±4.3cd	39.2±4.6c	39.1±4.1d
VO2max (ml/kg/min)	23.1±4.6a	23.9±4.9	24.3±4.6a
RMRm (kcal/day)	1238±185ce	1192±160c	1164±148e
RMRp (kcal/day)	1246±138	1256±141	1257±125
RMRm-p (kcal/day)		−64±129 **	−94±127 ***

Data shown as mean±SD. FM: fat mass; FFM: fat-free mass; RMR: resting metabolic rate; RMRm: RMR measured; RMRp: RMR predicted. Averages with the same superscript letter are significantly different:

^a,bP<0.05

^cdP<0.01

^eP<0.001.

^**P=0.001 and

^***P<0.001 for the comparisons between RMRm-p.

No significant correlations were seen between metabolic adaptation, at either wks 16 or 32, and age, baseline weight or BMI, and weight, FM and FFM loss at the respective timepoints. Moderate to strong positive correlations were seen between metabolic adaptation at wks 16 and 32 and changes in net submaximal oxygen consumption in different tasks (see Table 3). Scatterplots for the association can be seen in Figure 1 and Figure 2 for metabolic adaptation at weeks 16 and 32, respectively.

Table 3.

Correlation between metabolic adaptation at weeks 16 and 32 and changes in net VO₂ during different submaximal exercise tasks

	Walk	Walk with grade	Steps	Carry load
Metabolic adaptation wk 16
r	−0.545	−0.538	−0.508	−0.570
P	<0.001	<0.001	<0.001	<0.001
n	76	77	75	74
Metabolic adaptation wk 32
r	−0.432	−0.418	−0.351	−0.430
P	0.001	0.001	0.002	0.001
n	52	53	51	49

Wk: week; r: correlations coefficient; P: significance level; n: sample size

Figure 1.

Association between metabolic adaptation at weeks 16 and changes in net VO₂ during different submaximal exercise tasks. VO₂: volume of oxygen; wk: week, B: baseline.

Figure 2.

Association between metabolic adaptation at weeks 32 and changes in net VO₂ during different submaximal exercise tasks. VO₂: volume of oxygen; wk: week, B: baseline.

There were no significant differences in metabolic adaptation between women who lost (n=45, Δweight: −1.8±1.4 kg) and did not lose weight (n=35, Δweight: 1.1±0.9 kg), with the exercise intervention at weeks 12 (−55±111 vs −57±153 kcal/day, P=0.947). Similar results were seen at week 32, with no significant differences in metabolic adaptation (−72±120 vs −99±132 kcal/day, P=0.467) between those who lost weight (n=35, Δweight: −3.0±4.2 kg) and did not lose weight (n=18, Δweight: 1.1±1.4 kg).

Regression analysis showed that age, baseline FFM, RMR and RQ and changes in net VO₂ after walking were all predictors of metabolic adaptation at week 16, explaining 90% of the variation (P<0.001) (Table 4). Metabolic adaptation (kcal/day) at week 16 = −587.3 + (11.1 * Age (years)) + (79.7 * Race (0 for whites and 1 for Blacks)) − (25.9 * FFM (kg)) + (0.9 * RMR (kcal/day)) − (294.2 * RQ) − (6.5 * change in net VO₂ during walking (ml/kg/min)). All variables were significant predictors of metabolic adaptation (P<0.001). The model did not change when other tasks were used (walking with inclination, steps or carry a load, all R² adj = 0.9, P<0.001). In general, since metabolic adaptation is negative, a lower age, whites more than blacks, the greater the FFM, the lower the RMR, and the higher the RQ (lower fat oxidation) at baseline, and the larger the increase in net VO₂ after different exercise tasks, the greater the metabolic adaptation.

Table 4.

Regression model for predicting metabolic adaptation after16 (model A) and 32 weeks (model B) of combined resistance and aerobic training in older women

Model A	Unstandardized β	Standardized β	R2 adjusted	P
			0.902	<0.001
Intercept	−597.3
Age (years)	11.2	0.289		<0.001
Race	79.7	0.227		<0.001
Baseline FFM (kg)	−25.9	−0.776		<0.001
Baseline RMR (kcal/day)	0.9	1.178		<0.001
Baseline RQ	−294.2	−0.117		0.005
Change net VO2 walk (ml/kg/min)	−6.5	−0.911		<0.001
Model B	Unstandardized β	Standardized β	R2 adjusted	P
			0.889	<0.001
Intercept	−10002.3
Age (years)	12.7	0.355		<0.001
Race	95.5	0.287		<0.001
Baseline FFM (kg)	−25.0	−0.852		<0.001
Baseline RMR (kcal/day)	0.9	1.256		<0.001
Change net VO2 walk (ml/kg/min)	−6.2	−0.769		<0.001

FFM: fat-free mass, RMR: resting metabolic rate; RQ: respiratory quotient; VO₂: volume of oxygen. Race: 0 for whites and 1 for blacks.

Metabolic adaptation at week 32 was also predicted by the same variables, except for baseline RQ, and the model explained 89% of the variation (P>0.001) (Table 4). Metabolic adaptation (kcal/day) at week 32 = −1002.3 + (13.0 * Age (years)) + (95.5 * Race (0 for whites and 1 for Blacks)) − (25.0 * FFM (kg)) + (0.9 * RMR (kcal/day)) − (6.2 * change in net VO₂ during walking (ml/kg/min)). All variables were significant predictors of metabolic adaptation (P<0.001 for all). The model did not change when other tasks were used (walking with inclination, steps or carry a load, all R² adj = 0.89, P<0.001).

Discussion

The present findings represent the first study examining the impact of a combined aerobic and resistance training program on metabolic adaptation. We found significant metabolic adaptation (RMRm-RMRp = −60±136 kcal/day) after an average 640 kcal/week energy deficit (1kg weight loss) at week 16 in older women. Moreover, a subgroup analysis in 53 women with data at all timepoints showed an increase in metabolic adaptation from week 16 to week 32 (−64±129 vs −94±127 kcal/day), in line with the increase in overall energy deficit and weight loss. Additionally, metabolic adaptation was predicted by age, baseline FFM and RMR, and change in FFM and net VO₂ of walking.

Hopkins and colleagues (20) showed in a one-arm study that, even though 12 weeks of supervised aerobic exercise did not induce metabolic adaptation at the level of RMR in women with overweight and obesity, 43% experienced a greater than expected decline in RMR (−103±78 kcal/day). In another study, Martin et al (21) randomized 171 men and women with overweight or obesity to either a non-exercise control group or 1 of 2 supervised exercise groups: 8 or 20 kcal/kg of body weight/wk for 24 weeks and found no evidence of metabolic adaptation at the level of RMR. In the present analysis, we found metabolic adaptation (−60±136 kcal/day) despite a similar weight loss (approximately 1 kg in the three studies), and that 70% of our women experienced metabolic adaptation at week 16 (−131±73 kcal/day) and 73% at week 32 (−151±88 kcal/day). The studies by Hopkins et al and Martin et al differ from ours in several aspects. First, the program included aerobic exercise alone, while in ours combined aerobic and resistance training was used. Second, even though weight loss was similar (average 1 kg, ranging from 0.4 to 1.7 kg across studies), no changes in FFM were seen in the previous 2 studies, while in the present analysis an increase was observed. Finally, even though average BMI was similar, participants were much younger in the two previous studies compared with the present analysis (41±10 and 48.9±11.4 vs 65±4 years of age). These differences may contribute to the discrepancies found, particularly when it comes to age. Ten Haff and colleagues showed, in a combined analysis of 254 subjects with overweight and obesity, that metabolic adaptation in response to weight loss was only significant in older subjects (defined as above 55 years of age), but not in younger adults, and that the magnitude of metabolic adaptation increased with age, independently of sex and type (diet or diet + exercise) or duration of the intervention (28). This suggests that older adults might be more protected against weight loss and that a greater energy deficit might be needed to induce weight loss, compared with younger people.

We found metabolic adaptation in response to aerobic and resistance training to be predicted by baseline demographic (race and age) anthropometric (FFM at baseline) and metabolic variables (RMR and RQ), as well as changes in net VO₂ during different exercise tasks in this sample of older women. After accounting for race, age, and FFM at baseline, a lower RMR and fat oxidation at baseline were associated with greater metabolic adaptation. Hopkin and colleagues had previously reported a greater metabolic adaptation in response to aerobic exercise to be associated with an attenuated increase in resting fat oxidation (20). We also found that the larger the increase in net VO₂ after different exercise tasks, the greater the metabolic adaptation. This might indicate that the body responds to larger energy deficits, induced by everyday exercise tasks (walking, carrying loads, steps), by downregulation energy metabolism at rest (larger metabolic adaptation at the level of RMR). Overall, these findings suggest that changes in energy expenditure and substrate oxidation could work in concert to favour the defence of body weight.

It must be pointed out that we (29–32), as well as others (33, 34), have previously shown that resistance training results in increases in RMR, probably due to increased lean tissue. In addition, we have also found that aerobic training results in increased RMR at least for 22 hours following either a moderate intensity or high intensity exercise under energy balance conditions (35, 36), probably because of increased sympathetic tone and repair of damaged tissue. The subjects in this study had not exercised for over 48 hours when their RMR was evaluated, presumably too long a time to capture increased energy expenditure due to increased sympathetic activation or muscle repair. Considering that exercise training can have an effect on metabolic adaptation, even after 6 months of training, but that energy expenditure is elevated for at least 22 hours following an exercise bout, these results could be interpreted to further support the concept that exercise is medicine that should be administered regularly.

Opposite several other studies, including diet (9), exercise (20) or a combination of both (10), we did not find an association between weight or FM loss and metabolic adaptation. Several reasons may explain this. First, maybe there was not enough variability in weight and FM loss in the present study to allow for that association to be found. Second, this study was the only one where a significant increase in FFM was observed. It might be that this unique change in body composition that follows resistance training masked the association between weight and metabolic adaptation. We have previously shown that the addition of resistance training to an energy-restricted diet preserves FFM and RMR, despite a 12kg weight loss, while the addition of aerobic exercise fails to do so (32).

Other recent studies have likewise found that regular activity is associated with lower energy expenditure. Careau and colleagues (37) have recently showed, in an analysis of 1754 adults with paired measurements of RMR and total energy expenditure (TEE), a strong compensation between activity energy expenditure and RMR, with more than 25% of the extra energy spent on physical activity not translating into extra energy burned due to reduced RMR (37). Kevin Hall has also suggested, in a reinterpretation of the results of “The Biggest Loser” contest, that increased physical activity might lead to a compensatory reduction in RMR (38). However, the impact of physical activity on TEE and their components is a complex issue and their association in likely not linear. We have previously shown in this population of older women with overweight or obesity that aerobic training twice/week (both aerobic and resistance exercise) induced larger increases in TEE, activity EE (AEE), and non-exercise activity thermogenesis (NEAT) than training less frequently (1/week) or more frequently (3 times/week). In fact, training 3 days/week reduced NEAT, with no overall impact on TEE, probably because of the large time obligation and/or fatigue. Moreover, REE did not change in any of the groups (25). Therefore, the reduction in RMR below predicted values seen in the present analysis could be due to exercise alone (due to a compensatory reduction in RMR following increased physical activity levels), acute weight loss, as we have previously shown in response to energy restricted diets (8, 9), or potentially a combination of both. Even though metabolic adaption might disappear after participants discontinue exercising, more research in needed to understand the impact of exercise on RMR, TEE and body weight homeostasis. It needs also to be taken into account that aging in itself is already associated with metabolic adaptation, when compared with younger subjects (23). This is due to changes in the association between RMR and FFM, with a lower contribution of high-metabolic rate organs and tissues versus skeletal mass and bone with aging (22).

Our study has both strengths and limitations. Gold standard procedures were used for the measurement of RMR (indirect calorimetry) and body composition (DXA) and the exercise training program was supervised. However, this study also suffers from some limitations. First, it includes a very homogenous sample of older women (60–74 yr old), mainly white and with overweight and obesity. This prevents the generalisation of our results to men, other BMI groups and younger subjects. Moreover, this also explains why our regression model had an R² of only 55%, i.e. a truncated range for age and only women. Second, our study, like the vast majority of the literature, examined the presence of adaptive thermogenesis after adjusting for changes in FFM, assuming that this compartment is an homogenous tissue, that losses in FFM are uniform across its components (skeletal muscle and organs), and that tissue hydration remains constant. Reductions in organ masses during weight loss (39) have been reported, as well as with greater age (40) and it may be that after accounting for changes in organ tissue composition, adaptive thermogenesis becomes negligible (41). Third, the magnitude of metabolic adaptation found in the present study might be below the precision of the methods used to measure RMR (indirect calorimetry) and body composition (DXA). However, this is unlikely to have affected the accuracy of our findings, as there is no reason to believe that our measurements were biased, and data “errors” probably behaved at random. Fourth, data collection at weeks 16 and 32 was done with participants in negative EB. We have previously shown that metabolic adaptation following weight loss is significantly reduced, or even abolished after a short period (4 weeks) of weight stabilization, when the measurement of body composition and RMR is done under conditions of EB (8). Finally, our study did not have a control group and, as such, it cannot be ascertained if metabolic adaptation in the present study was caused by exercise per se, weight loss or a combination of both.

In conclusion, in older women, most with overweight or obesity, a small energy deficit induced by combined aerobic and resistance training program is associated with metabolic adaptation. Further research should confirm these findings in a population of men and women with different ages and BMIs, after controlling for EB status and for changes in the anatomical and molecular composition of FFM.

What is already known?

• Diet-induced weight loss is accompanied by metabolic adaptation and its magnitude modulated by the energy balance status (EB) of the individual when the measurements are taken.

• Metabolic adaptation does not seem to be a risk factor for weight regain.

• Metabolic adaptation is associated with less weight and fat mass loss after low-energy diets, and a longer time to achieve weight loss goals.

What does this study add?

• A combined aerobic and resistance exercise program was associated with metabolic adaptation at the level of resting metabolic rate (RMR), in older women with overweight and obesity.

• The lower the RMR and fat oxidation at baseline, and the larger the increase in net VO₂ after different exercise tasks, the greater was the metabolic adaptation.

How might these results change the direction of the research or the focus of clinical practice?

• Recommendations for exercise volume for older adults might change based on the results of this study. However, future research needs to confirm these findings in a population of men and women with different ages and BMIs, after controlling for EB status and for changes in the anatomical and molecular composition of FFM.

Data sharing:

Data described in the manuscript will not be made available because that would violate IRB and HIPAA rules.