Fat-Free Mass Normalization Impacts Cardiorespiratory Fitness in Overweight Adolescents

Abstract

Accurate assessment of cardiorespiratory fitness (CRF) in adolescents is critical. However, normalizing oxygen consumption (VO₂) to body mass (BM) may underestimate CRF in overweight (OW) youth by including metabolically inactive fat mass. This study examined differences in VO₂ normalized by BM and fat-free mass (FFM) between normal weight (NW) and OW adolescents. Thirty-eight participants (19 NW, 19 OW; 12–17 years) underwent anthropometric, body composition, and cardiorespiratory fitness assessments. VO₂ at the aerobic threshold (VO₂AerT), anaerobic threshold (VO₂AnT), and peak exercise (VO₂peak) were measured and expressed in absolute terms and relative to BM and FFM. Group differences in the main outcomes were analyzed using a one-way ANOVA, and Pearson correlation was used to examine associations between VO₂, BM and FFM. When normalized by BM, NW adolescents showed significantly higher VO₂AerT (18.7 ± 3.6 vs. 14.5 ± 2.3), VO₂AnT (28.8 ± 6.3 vs. 23.6 ± 4.7), and VO₂peak (37.7 ± 6.7 vs. 29.1 ± 7.0) compared to OW peers (p < 0.05). No significant differences were found when VO₂ values were normalized by FFM. A group difference was observed in the VO₂peak vs. BM slope (p = 0.03) but not in the VO₂peak vs. FFM slope. FFM normalization provides a more accurate assessment of CRF by accounting for differences in body composition, underscoring the importance of evaluating the aerobic capacity of metabolically active tissue rather than total body weight in youth populations.

1. Introduction

Over the past three decades, the prevalence of adolescent overweight and obesity in the U.S. has risen dramatically, presenting both a public health crisis and economic burden [1]. Early interventions targeting cardiovascular and metabolic health are essential to reduce long-term disease risk in this population [2,3,4].

While weight management is important, cardiorespiratory fitness (CRF) is independently associated with favorable health outcomes in youth [5]. CRF is typically assessed during a graded exercise test using peak oxygen uptake (VO₂peak), often normalized to body mass (BM) to account for individual differences in size. However, this method may underestimate CRF in overweight (OW) individuals, as it incorporates metabolically inactive fat mass (FM), thereby lowering VO₂ values regardless of true aerobic capacity [6,7,8]. To address this limitation, researchers have explored alternative normalization strategies, such as adjusting VO₂ to lean body mass, fat-free mass (FFM), or standardizing against predicted BM at the 50th BMI percentile [9,10,11]. FFM, defined as total body mass minus FM, is an accessible measure in adolescents and is directly related to CRF, as it excludes metabolically inactive fat [12]. Thus, FFM is a practical and physiologically meaningful variable for assessing CRF.

In adolescents, submaximal markers of CRF such as VO₂ at the aerobic threshold (VO₂AerT) and anaerobic threshold (VO₂AnT) are frequently used for individuals who cannot achieve maximal effort during a graded exercise test [13]. These submaximal indicators reflect metabolic efficiency and endurance capacity and are reliable proxies of CRF [3,13,14,15]. To our knowledge, few studies have investigated how normalizing submaximal parameters and VO₂peak by FFM affects the interpretation of CRF in OW versus normal weight (NW) adolescents.

This study aims to determine whether normalization to FFM provides a more accurate and equitable assessment of CRF in these populations compared to the traditional normalization to body BM, both at maximal and submaximal intensities. We hypothesized that OW adolescents would demonstrate lower VO₂peak, VO₂AerT, and VO₂AnT than their NW peers when expressed relative to BM, and that these differences would be largely attenuated when expressed relative to FFM.

2. Materials and Methods

Thirty-eight North Carolina adolescents (age 14 ± 2 years) were recruited for this cross-sectional, observational study. To ensure representation across socioeconomic levels and ethnic groups, participants were recruited from a state-funded program serving both rural and urban areas. Recruitment also targeted the local rural community in North Carolina, USA, through flyers, word of mouth, and email outreach. Participants were assigned to an NW (6 female and 13 male, 19 total) or OW (6 female and 13 male, 19 total) group based on body mass index (NW, BMI <85th percentile and OW, BMI ≥ 85th percentile for age and sex) and body fat percentage (BF%) (NW < 24% for males and <30% for females; OW ≥ 24% for males and ≥30% for females) [14,16]. Written informed consent was obtained from parents or legal guardians. Assent was obtained from all adolescent participants using a simplified assent form, with clarifications provided as needed. Consent and assent were provided in person upon arrival at the laboratory under the supervision of the principal investigator. Inclusion criteria included an age between 12 and 17 years. Exclusion criteria were age ≤11 or ≥18 years, or the presence of established renal, metabolic, or cardiovascular disease. Participants were instructed to avoid strenuous physical activity on the day of testing and to refrain from consuming caffeine, dietary supplements, or other stimulants for at least four hours beforehand. To ensure consistency in data collection while accommodating participant availability, tests were scheduled from 8:00 to 11:00 AM and from 4:00 to 6:00 PM.

2.1. Body Composition and Anthropometric Measurement

Height and BM were measured using a stadiometer and a calibrated scale to the nearest 0.1 cm and 0.1 kg, respectively. FM and FFM in absolute and relative values (FM%, FFM%) were assessed using BOD POD (COSMED USA, Concord, CA, USA) using the Siri equation and the predicted thoracic gas volume. Participants were given the opportunity to use the restroom prior to testing and were required to wear tight-fitting clothing and a swimmer’s cap, with shoes and jewelry removed during the assessment. Before each test, a warm-up, autorun, scale calibration, and volume calibration were performed following manufacturer’s recommendations. The BOD POD has been shown to obtain acceptable [17] and reliable FM% measurements [18] while providing a minimally invasive and non-operator-dependent assessment.

2.2. Cardiorespiratory Fitness Test

A graded exercise test to volitional exhaustion was used to assess CRF. Participants exercised on an electronically braked pediatric cycle ergometer (Lode Corrival, Groningen, The Netherlands) while respiratory gas exchange was monitored breath-by-breath using a portable metabolic cart (K5, Cosmed, Chicago, IL, USA). The cycle ergometer allowed precise control of the workload and minimized the risk of injury, while the reduced dimensions of the portable metabolic cart improved ease of exercise. During the graded test, participants rested for 2 min and then completed a 15 Watts per minute protocol at a cadence of 65–70 rpm, starting from unloaded pedaling at 0 Watts. The test ended when participants pedaled at a cadence slower than 50 rpm or decided to stop. Prior to each test, the metabolic cart was calibrated according to the manufacturer’s instructions, including turbine calibration with a 3-L syringe, gas calibration with reference concentrations of O₂ (16.00% and 20.93%) and CO₂ (5.0% and 0.04%), as well as CO₂ scrubber and delay calibrations [19]. Raw breath-by-breath data were processed using a 6-point moving average and subsequently averaged over 10-s intervals. Heart rate (HR) was continuously monitored with a GARMIN chest strap device (GARMIN, Olathe, KS, USA). The ventilatory equivalent method was used as the primary criterion for the graphical determination of VO₂AerT and VO₂AnT, with the V-slope method as a secondary criterion [15]. Two independent investigators determined the thresholds, and results were confirmed by an experienced exercise physiologist if they differed by more than 30 s. The values of VO₂AerT, VO₂AnT, and VO₂peak were calculated as 30-s averages, with VO₂peak defined as the average of the highest VO₂ values at the end of the test. The VO₂AerT, VO₂AnT, and VO₂peak were expressed in absolute terms (mL/min), relative to BM (mL/kgBM/min), and relative to FFM (mL/kgFFM/min).

2.3. Statistical Analysis

Based on an a priori power analysis using previously published VO₂peak (mL/kgBM/min) data from our laboratory, we calculated that a two-tailed comparison of independent means with an anticipated effect size of 0.910, an alpha level of 0.05, and a desired power of 0.8 would require 20 participants per group (NW and OW) to detect a statistically significant difference. Mean differences in primary (VO₂peak, VO₂AerT, and VO₂AnT) and secondary (anthropometrics) outcomes between NW and OW participants were identified using a one-way ANOVA. Normality of the sample was assessed using the Shapiro–Wilk test and homogeneity of variances with Levene’s test. Effect sizes for between-group differences were calculated using Cohen’s d, with thresholds of 0.2, 0.5, and 0.8 interpreted as small, moderate, and large effects, respectively. Associations between independent (BM and FFM) and dependent (VO₂peak, VO₂AerT, and VO₂AnT) variables were assessed using Pearson correlations and linear regressions. Differences in regression slopes between NW and OW adolescents were tested using Analysis of Covariance (ANCOVA) to determine whether the relationship between the dependent variable (VO₂) and the covariate (BM or FFM) varied significantly between groups (grouping factor: NW, OW). Results were expressed as mean and SD. Statistical significance was defined as p < 0.05. The SPSS Software, IBM version 29.0.1.0 (IBM Corp., Armonk, NY, USA) was used to perform all analyses.

3. Results

Our sample was composed of participants representing different ethnicities including White (n = 24, 63.2%), Indian (n = 7, 18.4%), Asian (n = 4, 10.5%), and Black (n =3, 7.9%).

All anthropometric and fitness measurement raw data are reported in the Supplementary Table S1. Shapiro–Wilk tests indicated that most CRF indices were normally distributed (p > 0.05), except VO₂AerT normalized by BM, VO₂AnT, and VO₂AnT normalized by BM (p < 0.05). Given the robustness of ANOVA in moderating deviations from normality, all variables were included in the analysis.

Table 1. Anthropometric Characteristics of Normal Weight vs. Overweight Adolescents.

Variables	NW	OW
Age (years)	14.5 ± 2.3	14.0 ± 2.3
Height (cm)	164.8 ± 13.5	166.0 ± 11.7
Weight (kg)	52.8 ± 13.1	79.2 ± 24.1 **
BMI	19.1 ± 2.9	28.2 ± 5.7 **
FM (kg)	8.6 ± 4.8	28.6 ± 14.2 **
FFM (kg)	44.2 ± 12.1	50.7 ± 14.2
FM%	16.4 ± 7.9	35.3 ± 10.3 **
FFM%	83.6 ± 7.9	64.7 ± 10.3 **
RERpeak	1.20 ± 0.12	1.16 ± 0.09
PWRpeak (Watt)	164 ± 51	178 ± 49

BMI, body mass index; FM, fat mass; FFM, fat-free mass; FM%, percentage of FM; FFM%, percentage of FFM; RERpeak, peak respiratory exchange ratio; PWRpeak, peak power; NW, normal weight; OW, overweight. ** p < 0.01 compared to NW.

presents the anthropometric profiles of the NW and OW groups. Compared to NW peers, OW adolescents had significantly higher BM (p < 0.01), BMI (p < 0.01), FM (p < 0.01), and FM% (p < 0.01). Conversely, FFM% was significantly lower in OW compared to NW participants (p < 0.01), while absolute FFM did not differ significantly between groups (p > 0.05).

Table 2. Oxygen Consumption Parameters in Normal Weight and Overweight Adolescents.

Variables	NW	OW	Cohen’s d
VO2AerT (mL/min)	968 ± 238	1117 ± 258	−0.60
VO2AerTFFM (mL/kg/min)	22.4 ± 3.5	22.6 ± 3.2	−0.07
VO2AerTBM (mL/kg/min)	18.7 ± 3.6	14.5 ± 2.3 **	1.39
VO2AnT (mL/min)	1514 ± 510	1842.4 ± 593	−0.59
VO2AnTFFM (mL/kg/min)	34.6 ± 7.1	36.1 ± 4.6	−0.25
VO2AnTBM (mL/kg/min)	28.8 ± 6.3	23.6 ± 4.7 **	0.94
VO2peak (mL/min)	2003 ± 661	2266 ± 746	−0.37
VO2peakFFM (mL/kg/min)	45.1 ± 6.4	44.7 ± 6.6	0.06
VO2peakBM (mL/kg/min)	37.7 ± 6.7	29.1 ± 7.0 **	1.25

VO₂AerT, oxygen consumption at the aerobic threshold; VO₂AerT_FFM, VO₂AerT normalized by fat-free mass; VO₂AerT_BM, VO₂AerT normalized by body mass; VO₂AnT, oxygen consumption at the anaerobic threshold; VO₂AnT_FFM, VO₂AnT normalized by fat-free mass; VO₂AnT_BM, VO₂AnT normalized by body mass; VO₂peak, peak oxygen consumption; VO₂peak_FFM, VO₂peak normalized by fat-free mass; VO₂peak_BM, VO₂peak normalized by body mass; NW, normal weight; OW, overweight. ** p < 0.01. Significance between NW and OW adolescents.

reports the mean differences in oxygen consumption parameters at the aerobic threshold (VO₂AerT), anaerobic threshold (VO₂AnT), and peak exercise (VO₂peak) in NW and OW adolescents along with Cohen’s d statistics. When normalized to BM, VO₂AerT, VO₂AnT, and VO₂peak were significantly lower in OW compared to NW (p < 0.01). No significant differences were observed when normalized to FFM. Between-group comparisons revealed large effects for BM-normalized indices (VO₂AerT, VO₂AnT, VO₂peak; Cohen’s d = 0.94–1.39), while FFM-normalized indices showed negligible to small effects (Cohen’s d = −0.25 to 0.06). Absolute VO₂ demonstrated moderate differences favoring NW adolescents (Cohen’s d = −0.37 to −0.60).

Figure 1 depicts differences in VO₂AerT, VO₂AnT, and VO₂peak values between the NW and OW groups. Significant differences between NW and OW participants were observed in VO₂AerT, VO₂AnT, and VO₂peak when normalized to BM (p < 0.05). Compared to OW participants, NW participants reported higher VO₂AerT (18.7 ± 3.6 vs. 14.5 ± 2.3), VO₂AnT (28.8 ± 6.3 vs. 23.6 ± 4.7), and VO₂peak (37.7 ± 6.7 vs. 29.1 ± 7.0) when normalized to BM, whereas no significant differences in VO₂AerT (22.4 ± 3.5 vs. 22.6 ± 3.2, p = 0.85), VO₂AnT (34.6 ± 7.1 vs. 36.1 ± 4.6, p = 0.83), and VO₂peak (45.1 ± 6.4 vs. 44.7 ± 6.6, p = 0.46) were observed when normalized to FFM.

Figure 2 displays the relationships between VO₂AerT, VO₂AnT, and VO₂peak with BM and FFM in NW and OW adolescents (panels a–f). VO₂AerT showed moderate to very strong correlations with BM (r = 0.673, p < 0.01, NW and r = 0.818, p < 0.01, OW) and FFM (r = 0.779, p < 0.01, NW and r = 0.839, p < 0.01, OW). Similarly, VO₂AnT showed moderate to very strong correlations with BM (r = 0.712, p < 0.01, NW and r = 0.771, p < 0.01, OW) and FFM (r = 0.726, p < 0.01, NW and r = 0.870, p < 0.01, OW). VO₂peak showed very strong correlations with both BM (r = 0.820, p < 0.01, NW and r = 0.819, p < 0.01, OW) and FFM (r = 0.885, p < 0.01 for both groups). A significant difference in the slope of the VO₂peak vs. BM relationship was found between NW and OW adolescents (p = 0.03).

4. Discussion

The purpose of this study was to evaluate the impact of normalizing VO₂ by FFM versus BM on the assessment of CRF in OW and NW adolescents. When VO₂AerT, VO₂AnT, and VO₂peak were normalized to BM, OW adolescents exhibited significantly lower values compared to their NW counterparts. However, these differences were no longer evident when VO₂ parameters were expressed relative to FFM. These findings align with the group differences in anthropometric characteristics reported in Table 1. OW adolescents had significantly greater BM, BMI, FM, and FM%, and lower FFM% than NW peers, while absolute FFM did not differ. This indicates that excess fat mass, rather than reduced lean tissue, explains the lower BM-normalized VO₂ in OW adolescents. In contrast, comparable FFM across groups supports the use of FFM-based normalization, which reflects metabolically active tissue and provides a fairer assessment of aerobic capacity.

Our results are consistent with previous studies showing that normalizing VO₂ to BM underestimates aerobic capacity in individuals with excess adiposity, due to the inclusion of metabolically inactive FM [10,15,20,21]. Our results also support the work of Ekelund et al. (2004), who reported that VO₂peak normalized to FFM is a more accurate indicator of CRF in obese adolescents [22]. However, the literature is less consistent regarding submaximal VO₂. In contrast to Ekelund et al. (2004), we observed significant group differences in submaximal VO₂ (both VO₂AerT, VO₂AnT) expressed relative to body mass [22]. This discrepancy may stem from methodological differences; for example, they tested obese adolescents using treadmill protocols, where excess fat mass can elevate the metabolic cost of weight-bearing exercise. Indeed, higher fat mass increases energy expenditure during locomotion without contributing to oxygen delivery or utilization, leading to deceptively high BM-relative VO₂ values in obese individuals [6,23]. Previous research indicates that the muscles of overweight/obese children and adolescents often show impaired oxidative capacity, including reduced type I fiber area, slower VO₂ kinetics, diminished mitochondrial efficiency, and microvascular dysfunction [24]. These alterations may manifest as reduced VO₂ even when normalized to FFM. In contrast, our findings showed no such differences, as VO₂ normalized to FFM was similar between NW and OW adolescents. Given the negative association between type I fiber proportion and BMI, we speculate that the absence of group differences may be explained by the fact that our participants were overweight but not obese [25]. Thus, the OW group may not yet have developed the muscle fiber and mitochondrial alterations typically reported in obesity.

Regression analyses revealed moderate to very strong correlations between VO₂peak and both BM and FFM, with slightly stronger associations observed when VO₂peak was expressed relative to FFM. Although correlations were consistently high across all conditions, the slopes of VO₂peak versus BM differed significantly between NW and OW adolescents, whereas no such differences were observed for VO₂peak versus FFM (Figure 2). These findings suggest that normalizing VO₂peak to BM in OW adolescents may lead to inaccurate interpretations of CRF, while FFM normalization provides a more consistent and physiologically meaningful assessment across weight categories. These findings suggest that BM is a confounding variable in CRF assessments, as it includes metabolically inactive FM, which disproportionately affects overweight individuals [26]. Using FFM as a normalization factor provides information about the CRF of the working muscles, removing biases introduced by variations in adiposity.

These findings carry important implications for both clinical practice and public health. Accurate assessment of CRF in OW adolescents is crucial for identifying those at increased risk for obesity-related cardiometabolic conditions. Relying solely on BM-based normalization may underestimate fitness in OW individuals, potentially leading to misclassification and inappropriate or insufficient interventions. Normalizing VO₂ by FFM provides a more physiologically meaningful evaluation of aerobic capacity, enabling healthcare professionals to tailor exercise prescriptions more effectively [12,27].

Our findings indicate that normalization to FFM reduces intergroup differences in VO₂, including values obtained at the VO₂AerT and VO₂AnT. This observation underscores the utility of submaximal indicators of CRF, particularly in OW adolescents who may be unable to achieve valid maximal exertion during graded exercise testing [13,21]. Moreover, the use of individualized ventilatory thresholds, rather than fixed percentages of VO₂peak, provides a physiologically grounded assessment of fitness at true metabolic transition points and minimizes the risk of misestimating submaximal CRF. Given that FFM is the primary determinant of oxygen consumption during exercise, FFM-based normalization serves as a valuable moderator not only of CRF but also of cardiovascular risk [11,22]. CRF normalized to FFM has shown stronger associations with metabolic health outcomes than BM-based measures [7,28].

These findings contribute to the understanding of the “fat-but-fit” paradigm, wherein OW individuals with high CRF exhibit cardiometabolic profiles comparable to or better than normal weight but sedentary individuals [5,29,30]. Active, insulin-sensitive muscle mass appears to buffer the metabolic effects of adiposity, enhancing glucose control, lipid profiles, and vascular function [31]. Recognizing and quantifying this distinction is critical for reducing stigma and ensuring appropriate care pathways for OW youth with high fitness levels.

Psychologically, FFM-based normalization can serve as an important motivational tool for adolescents and their families as it reflects the performance of metabolically active tissue, offering a more accurate measure of aerobic capacity. In contrast, BM-normalized VO₂ often underestimates fitness in OW adolescents due to the influence of non-functional fat mass. This can lead to misclassification and reinforce negative perceptions about their physical abilities. FFM-based normalization may also support healthier self-perception in OW youth. BM-relative VO₂ values can stigmatize adolescents by labeling them as unfit, regardless of true physiological capacity. FFM-normalized CRF offers a fairer comparison with NW peers, emphasizing function over size and encouraging a more empowering narrative around physical activity [32,33]. This approach can highlight the child’s aerobic potential, rather than framing health solely in terms of weight loss. As youth grow and reduce fat mass, BM-normalized CRF will improve, but FFM-relative VO₂ provides a stable, developmentally appropriate measure of aerobic function. Long-term engagement may be further supported by combining exercise programs that promote aerobic capacity with nutritional strategies that foster healthy weight management. This new approach emphasizes the importance of physical activity and exercise in adolescents regardless of their body weight and may help promote active lifestyles throughout life.

This study has certain limitations that should be considered. The relatively small sample size (n = 38) and the absence of observed sex differences limit the generalizability of our findings. Moreover, because the OW group included adolescents classified as overweight but not obese, the conclusions may not fully extend to youth with obesity or to those with chronic medical conditions. The recruitment of participants from a single geographic region may also restrict the applicability of these results. Future studies should include larger and more diverse cohorts to better assess potential sex- and ethnicity-based differences in CRF normalization methods. Furthermore, factors such as training status, habitual physical activity, and dietary intake were not controlled, which may have influenced CRF measures. Similarly, participants’ diet, hydration status, and activity in the hours prior to testing were not recorded and relied on their adherence to our recommendations. Future research should consider these variables to refine the assessment of aerobic fitness in adolescent populations. Lastly, pubertal status was estimated rather than directly assessed, which could have impacted body composition and metabolic variables.

5. Conclusions

This study demonstrates that normalizing CRF indices by FFM rather than BM provides a more accurate representation of aerobic capacity in adolescents. The strong correlations between VO₂ parameters and FFM underscore the role of lean tissue in determining CRF, reinforcing the necessity of FFM-based normalization for evaluating fitness across weight categories. Additionally, distinguishing between aerobic capacity and aerobic fitness through FFM-based normalization offers a more equitable assessment of fitness potential, reducing the discouragement associated with weight-based evaluations. FFM-based normalization can attenuate intergroup differences not only at peak effort but also at AerT and AnT, thereby enhancing the validity of submaximal CRF assessments in overweight adolescents who may be unable to achieve true maximal exertion during graded exercise testing. This approach has important implications for public health, as it can help adolescents and families adopt a more positive outlook on physical activity and weight management, ultimately fostering long-term engagement in health-promoting behaviors.