Body composition is a valuable indicator of health status and a key determinant of athletic performance (6,23). Although body composition demands vary considerably between sports, excess body fat is typically associated with impaired athletic performance, whereas increased fat-free mass (FFM) is associated with improvements in physical performance (2). Unsurprisingly, measures of body composition and kinanthropometry have been shown to influence parameters of athletic performance such as strength, anaerobic power, and absolute power in a variety of athletic populations (14,16,19,20,26). Owing to the excellent predictive quality of body composition in athletes, the use of effective body composition classification metrics is of utmost importance to researchers and practitioners alike. However, because absolute measures of body composition such as fat mass (FM) or FFM do not allow for direct comparisons between athletes of different heights or body sizes, height-normalized metrics of body composition have been proposed as an effective means to quantify relative fatness or muscularity in athletic populations (24).
Fat-free mass index (FFMI) is a height-adjusted measure of FFM, which was initially proposed as a means to assess the presence of protein-energy malnutrition in subjects undergoing experimental semistarvation (27) and later suggested as a screening tool to detect the use of anabolic-androgenic steroid use in resistance-trained men (10). Calculated by dividing a person's absolute FFM by height squared, FFMI has been suggested as a superior alternative to body mass index as a means to classify disease risk and overall health status (1,3,9,11,13,18,21). Recent investigations have proposed that FFMI can be used to assess an athlete's capacity for FFM accretion, guide training and nutritional practices, recommend competition weight in weight-class sports, suggest optimal position assignments, aid athlete recruitment, and identify athletes' potential for future success (22,24). To date, naturally attainable upper limits of FFMI have been identified in National Collegiate Athletic Association (NCAA) Division I and II football players (24) and in resistance-trained men (10). In addition, although FFMI is inherently a height-adjusted metric, Trexler et al. (24) noted that additional height adjustment through linear regression may be necessary because increases in body thickness are disproportional to increases in body height. The application of this correction to all persons would allow for better comparison between individuals of varying height.
Currently, more research is needed regarding the applications of FFMI in athletic populations. Although earlier investigations have provided normative FFMI data for football (24) and baseball (12) athletes, more information is required in additional sports and competition levels (24). To date, height-corrected FFMI has only been assessed at the collegiate level in NCAA football athletes. Therefore, the purpose of this study was to expand on the pre-existing literature surrounding FFMI in athletes by providing sport-specific normative FFMI data in male collegiate athletes, reporting upper limits of FFMI when possible, and determining the presence of any between-sport differences in the sample. In addition, this study aimed to determine whether height corrections through linear regression are necessary when assessing FFMI in male collegiate athletes. It was hypothesized that average FFMI would differ significantly between sports.
Experimental Approach to the Problem
This IRB-approved, cross-sectional investigation was conducted to determine sport-specific FFMI values in male collegiate athletes. Over the course of 3 years, the body composition of male collegiate athletes from 10 different drug-tested sports was assessed using dual-energy x-ray absorptiometry (DEXA). Results of the body composition assessments were used to determine each subject's FFMI. These values were then further adjusted for height through linear regression, as described by Trexler et al. (24). Sport-specific normative values were calculated, and the naturally attainable upper limit for FFMI was identified by calculating the 97.5th percentile values for the entire cohort. Between-sport differences were identified using appropriate parametric or nonparametric methods.
Male collegiate athletes (n = 209) from 10 drug-tested sports participated in this study (mean ± SD; age: 20.7 ± 1.9 years [age range: 18-31 years], height: 182.9 ± 6.7 cm, body mass: 90.8 ± 16.8 kg, and percent body fat [BF%]: 15.6 ± 5.3%). This study included athletes from the following sports: baseball (n = 41), cross country (n = 6), football (n = 29), golf (n = 2), ice hockey (n = 3), weightlifting (n = 14), rugby (n = 54), swimming and diving (n = 36), track and field (n = 12), and water polo (n = 12). Absolute FM and FFM data for each sport can be found in Table 1. Written informed consent was given by all subjects before any data collection, and all consent forms and procedures were approved by the Lindenwood University institutional review board (IRB-19-L0032) before any data collection.
Each subject reported to the laboratory for a single study visit, during which all anthropometric and body composition assessments were collected. Owing to the high volume of scans required, data collection took place over a 3-year period during all parts of the competitive season. Subjects were instructed to arrive to the laboratory in a euhydrated state between the hours of 05:30–10:00 after refraining from exercise and caffeine consumption for at least 24 hours and from food or drink for 8 hours. Subjects wore light athletic clothing and were instructed to remove shoes as well as any plastic, metal, and easily removable jewelry from their body. Body mass was measured to the nearest 0.1 kg with a digital scale (BWB-627A; Tanita, Tokyo, Japan). Height was measured with a stadiometer (HR-200, Tanita Stadiometer; Tanita) to the nearest 0.1 cm. All demographic information (height, body mass, age, and ethnicity) was entered in the software provided by the DEXA manufacturer (Hologic APEX Software, Version 4.5.3). Subjects were positioned in a supine manner on the DEXA scanning bed in accordance with the manufacturer's instructions and were instructed to remain still during the duration of each assessment. All whole-body DEXA scans were completed using a single Hologic QDR Discovery A (Bedford, MA, USA; Serial # 85063) unit by a trained research associate according to manufacturer guidelines. Before all scans, internal calibration procedures were completed as outlined by the manufacturer. After the completion of all scans, a trained researcher (B.S.C.) analyzed all scans using the aforementioned software and the TBAR 1209 (Pre-NHANES) calibration method. Total BF%, total FM, total FFM, and total bone mineral content were then recorded for every subject. These procedures have yielded excellent test-retest reliability data in our laboratory for FM (95% confidence interval [CI] of the intraclass correlation coefficient [ICC] = 0.996–0.999), lean mass (ICC = 0.999 to >0.999), and bone mineral content (ICC = 0.996–0.999).
Unadjusted FFMI was calculated using equation 1, as described by VanItallie et al. (27).
Statistical correction of FFMI values for differences in height was determined according to the procedures of Kouri et al. (10) and Trexler et al. (24). Unadjusted FFMI values in all subjects above the median FFMI were regressed against height because this subgroup has been previously suggested to represent individuals with higher levels of FFM accretion (24). The regression slope (β = −0.014, p = 0.631) and the average height of all 209 subjects (182.9 cm) were then used in equation 2 to calculate height-adjusted FFMI (FFMIAdj) for each subject, using procedures similar to Trexler et al. (24).
All data were analyzed with IBM SPSS Statistics Version 25 (IBM, Armonk, NY, USA) and are presented as mean ± SD. Data were assessed for normality using the Shapiro-Wilks test, and log transformations were computed when assumptions of normality were violated. Measures of central tendency were calculated for FFMI and FFMIAdj. A paired-samples t-test was completed to identify differences between FFMI and FFMIAdj. The Levene test was used to assess homogeneity, and a one-way analysis of variance (ANOVA) with appropriate post hoc comparisons using log-transformed data was used to determine between-sport differences in FFMIAdj. Percentile classifications were also calculated. Finally, the 97.5th percentile was calculated to determine the natural upper limit for FFMIAdj, as described by Trexler et al. (24). An alpha level of 0.05 was determined to represent the threshold of statistical significance.
The mean unadjusted FFMI for the entire sample was 22.8 ± 2.8 kg·m−2 and the median was 22.7 kg·m−2. The mean unadjusted FFMI of the subgroup used for regression analysis was 25.0 ± 1.9 kg·m−2. The mean FFMIAdj for the entire sample was also 22.8 ± 2.8 kg·m−2. However, a paired-sample t-test revealed significant differences between adjusted and unadjusted FFMI values (p < 0.001). The Levene test demonstrated heterogeneous variance between groups (p = 0.001), so a one-way ANOVA with Games-Howell post hoc comparisons was computed and revealed significant differences (p < 0.05) in FFMIAdj between sports. As described in Table 2, average FFMIAdj for rugby athletes (24.28 ± 2.39 kg·m−2) was significantly greater than baseball (22.15 ± 1.61 kg·m−2, p < 0.001; 95% CI: [−3.371 to −0.891]) and swimming (20.94 ± 1.61 kg·m−2, p < 0.001; 95% CI: 2.069–4.613) athletes. Average FFMIAdj for football athletes (24.94 ± 2.42) was significantly greater than baseball (22.15 ± 1.61 kg·m−2, p < 0.001; 95% CI: 1.199–4.376), swimming (20.94 ± 1.61 kg·m−2, p < 0.001; 95% CI: 2.387–5.609), and water polo (20.68 ± 3.56 kg·m−2, p = 0.022; 95% CI: 0.480–8.041) athletes. In addition to being significantly less than rugby and football, average FFMIAdj for baseball athletes (22.15 ± 1.61) was significantly greater than swimming (20.94 ± 1.61 kg·m−2, p = 0.025; 95% CI: 0.096–2.325) athletes. Percentiles for the entire sample and each sport were calculated and are reported in Table 3. As mentioned previously, Trexler et al. (24) used the 97.5th percentile to report upper limits, and in this study, the upper limit was reported to be 29.1 kg·m−2 in rugby athletes, 25.5 kg·m−2 in baseball athletes, and 28.4 kg·m−2 across all sports. Owing to an inadequate sample size, upper limits were not computed for the remaining sports. Figure 1 depicts all FFMIAdj values plotted against height in relation to the upper limit identified for the entire cohort.
The purpose of this study was to provide sport-specific normative FFMIAdj data in male collegiate athletes, to determine the presence of between-sport differences in FFMIAdj, and to identify the upper limits of FFMI in this population. This study included the most diverse sample of male collegiate athletes to date and reported data in several sports for the first time. This investigation also found significant differences in FFMIAdj between sports. In addition, this study reported an estimation of the natural upper limit for height-adjusted FFMI for all athletes in the cohort and well-represented sports. Finally, the results of this study suggest that further height correction of FFMI data through linear regression may be necessary to account for the effect of increasing body thickness on relative muscularity, as significant differences (p < 0.001) were reported between unadjusted and height-adjusted values. Thus, any direct comparisons between the adjusted data reported by this study and previously reported unadjusted data must be made with caution.
Fat-free mass index has been shown to effectively distinguish between playing position (24), level of competition (24), and type of sport (17). Previous investigations (22,24) have suggested that normative FFMI data can be used to inform recruiting decisions and playing position assignments, guide training and nutritional practices, and predict future athletic achievement. As hypothesized, significant (p < 0.05) differences in FFMIAdj were identified between several sports in the sample (Table 2). Rugby athletes had significantly higher (p < 0.05) FFMIAdj values than athletes competing in baseball and swimming. Similarly, mean height-adjusted FFMI values in football players were found to be significantly higher (p < 0.05) than baseball, swimming, and water polo. These results are not surprising, given the physiological demands of rugby and football that favor increased body size and lean mass accretion (4,25), and agree with the previous literature indicating between-sport body composition differences (7,17). The between-sport differences found in the present investigation are similar to those reported by Santos et al. (17), who observed higher FFMI values in rugby athletes relative to triathletes, basketball players, and swimmers. However, comparisons between the results of the current study and those reported by Santos et al. must be made with caution because the researchers did not use height correction and did not adjust for multiple comparisons when analyzing between-sport differences. In addition, the percentiles reported by Santos et al. were calculated using a Bayesian estimation technique because of a limited number of athletes in several sports.
The normative height-corrected FFMI data for several sports in the present sample are in accordance with sport-specific FFMI data reported by earlier investigations, although differences in height correction methods make direct comparisons challenging. For example, Trexler et al. (24) reported that the average FFMIAdj in NCAA Division I and Division II football players was 24.1 ± 2.0 kg·m−2 and 23.1 ± 2.0 kg·m−2, respectively, whereas the NCAA Division II football players in this study were found to have an FFMIAdj of 24.9 ± 2.4 kg·m−2. Similarly, Loenneke et al. (12) reported a mean unadjusted FFMI of 21.1 ± 1.6 kg·m−2 in collegiate baseball players, whereas the NCAA Division II baseball players in this study were found to have an FFMIAdj of 22.2 ± 1.6 kg·m−2. A potential reason for these differences is the use, or lack thereof, of height-adjusted FFMI values in the Loenneke's article. To this point, Trexler et al. (24) used a similar height correction approach as the present investigation, whereas Loenneke et al. (12) did not report height-adjusted FFMI data. Finally, differences in sport-specific FFMI between this study and those reported by earlier investigations may be explained by differences in the positional makeup of the athletes in the sample. For example, if a high percentage of the football athletes recruited in the present sample were offensive or defensive linemen, the average FFMI data would likely be elevated compared to a sample with fewer linemen (24). Although FFMI data were not collected, Fields et al. (7) demonstrated the importance of position-specific data. In male swimmers, Fields et al. (7) reported similar overall BF percentage (14.2 ± 3.5%) and FM (11.1 ± 3.1 kg) to swimmers in the present investigation (13.2 ± 3.8% and 10.9 ± 4.1 kg, respectively). However, among their male swimmers, Fields et al. (7) reported significant differences in both BF percentage and FM between sprinters and distance swimmers, exhibiting the variation that can be present within a single sport.
The determination of appropriate sport-specific upper limits of FFMIAdj allows practitioners to establish and track lean mass accretion goals (24). This study reported the upper limit (97.5th percentile) of height-adjusted FFMI to be 28.4 kg·m−2 across all sports in the sample. In addition, the naturally attainable upper limits for several well-represented sports (n > 40) were calculated and were found to be 29.1 kg·m−2 for rugby athletes and 25.5 kg·m−2 for baseball athletes. These results are in accordance with earlier investigations, as Kouri et al. (10) reported an upper limit of 25 kg·m−2 in male athletes free from steroid use, and Trexler et al. (24) reported an upper limit of 28.1 kg·m−2 in NCAA Division I and NCAA Division II collegiate football players, with an upper limit of 28.8 kg·m−2 in the NCAA Division I football players. The results of this study suggest that similar upper limits of relative muscularity exist between NCAA Division I football players and collegiate rugby players, as well as between NCAA Division II baseball players and the male resistance-trained athletes recruited by Kouri et al. (10). It is reasonable for rugby athletes to have such a high FFMIAdj average and upper limit because of the high power to body mass ratio and the level of lean mass favored by the sport (4,5). Figure 1 provides an overview of all subjects' FFMIAdj values in relation to the 28.4 kg·m−2 upper limit proposed in this study and indicates that few subjects were above this threshold.
The FFMIAdj data reported in this study may differ from previous investigations for numerous reasons. First, the method of body composition measurement is not consistent between studies. In contrast to this study, which used a single DEXA unit for all assessments and analyses, several previous investigations have used various measures of body composition such as skinfold measurements (10) and bioelectrical impedance (12). Furthermore, Loenneke et al. (12) demonstrated that bioelectrical impedance analysis significantly underestimated athletes' FFMI compared with DEXA. Similarly, data collected with air plethysmography may underestimate FFM in larger individuals and overestimate FFM in smaller individuals (8). In addition, differences between DEXA units and between correction factors, such as the TBAR 1209 (Pre-NHANES) correction factor used in the present investigation, may further amplify disparities between data sets. Moreover, changes in pretesting procedures such as diet or exercise can impact DEXA results (15). Finally, variations may also be caused by a smaller sample size and if studies measure body composition at different times of athletes' competitive season.
This investigation has many strengths in contrast to the current body of the FFMI literature. First, a single DEXA unit was used to measure body composition in all subjects, thus minimizing the potential between-instrument error. Moreover, identical pretesting requirements and analysis approaches were used for all assessments. Second, several efforts were made to control for numerous variables that could impact DEXA results, such as diet, exercise, and the hour at which scans were completed. Third, this investigation encompasses the most diverse sample of FFMIAdj data in male collegiate athletes. Finally, these data were further corrected for height with linear regression, a change that was shown to be necessary in this population. Despite this study's strengths, it also has limitations that must be discussed. First, some of the included sports had a relatively small sample size, such as cross country running (n = 6), ice hockey (n = 3), and golf (n = 2). Consequently, these sports were not considered for between-sport comparisons and only had their descriptive data reported. Second, playing position and academic age data were not collected. Owing to the vast range of muscularity between positions within one sport, such as between forwards and backs in rugby or between throwers and sprinters in track and field, practitioners will be able to better apply these results if position-specific FFMI data were available, which the recent literature has demonstrated (7,24). Similarly, although all athletes in the sample were subject to random drug testing during the data collection period, anabolic-androgenic steroid use was not screened in the present investigation. Finally, although subjects were instructed to be hydrated, fasted, and abstained from exercise before reporting to the laboratory, no quantitative measurements were collected to confirm these criteria were met.
The findings of this study suggest that the upper limits of FFMI are higher than those previously reported in male collegiate athletes and support the utility of FFMI to distinguish between athletes competing in different sports. The present investigation determined the naturally attainable upper limit of FFMI for a wide variety of male collegiate athletes and provided several sport-specific upper limits in well-represented sports. This information can be used by coaches and trainers to determine an athlete's potential for further FFM accretion, establish body composition and training goals, and direct appropriate training and nutritional regimens. Furthermore, the normative FFMI data provided by this investigation can be used to evaluate an athlete's relative muscularity compared with collegiate level athletes in this particular sport, thus providing coaches with greater insight into potential athlete recruitment and personnel decisions.
The authors thank all subjects for their contributions to this project.
1. Bahadori B, Uitz E, Tonninger-Bahadori K, Pestemer-Lach I, Trummer M, Thonhofer R, et al. Body composition: The fat-free mass index (FFMI) and the body fat mass index (BFMI) distribution among the adult Austrian population-results of a cross-sectional pilot study. Int J Body Compos Res 4: 123, 2006.
2. Boileau RA, Horswill CA. Body Composition in Sports: Measurement and Applications for Weight Loss and Gain. Philadelphia, PA: Lippincott Williams and Wilkins, 2000. pp. 319–334.
3. Coin A, Sergi G, Minicuci N, Giannini S, Barbiero E, Manzato E, et al. Fat-free mass and fat mass reference values by dual-energy X-ray absorptiometry (DEXA) in a 20–80 year-old Italian population. Clin Nutr 27: 87–94, 2008.
4. Duthie G, Pyne D, Hooper S. Applied physiology and game analysis of rugby union. Sports Med 33: 973–991, 2003.
5. Duthie GM, Pyne DB, Hopkins WG, Livingstone S, Hooper SL. Anthropometry profiles of elite rugby players: Quantifying changes in lean mass. Br J Sports Med 40: 202–207, 2006.
6. Fields J, Merrigan J, White J, Jones M. Seasonal and longitudinal changes in body composition by sport-position in NCAA Division I Basketball Athletes. Sports 6: 85, 2018.
7. Fields JB, Merrigan JJ, White JB, Jones MT. Body composition variables by sport and sport-position in elite collegiate athletes. J Strength Cond Res 32: 3153–3159, 2018.
8. Frisard MI, Greenway FL, Delany JP. Comparison of methods to assess body composition changes during a period of weight loss. Obes Res 13: 845–854, 2005.
9. Hull HR, Thornton J, Wang J, Pierson RN, Kaleem Z, Pi-Sunyer X, et al. Fat-free mass index: Changes and race/ethnic differences in adulthood. Int J Obes (Lond) 35: 121–127, 2011.
10. Kouri EM, Pope HG Jr, Katz DL, Oliva P. Fat-free mass index in users and nonusers of anabolic-androgenic steroids. Clin J Sport Med 5: 223–228, 1995.
11. Kyle UG, Morabia A, Schutz Y, Pichard C. Sedentarism affects body fat mass index and fat-free mass index in adults aged 18–98 years. Nutrition 20: 255–260, 2004.
12. Loenneke JP, Wilson JM, Wray ME, Barnes JT, Kearney ML, Pujol TJ. The estimation of the fat free mass index in athletes. Asian J Sports Med 3: 200–203, 2012.
13. Lu Y, Shu H, Zheng Y, Li C, Liu M, Chen Z, et al. Comparison of fat-free mass index and fat mass index in Chinese adults. Eur J Clin Nutr 66: 1004–1007, 2012.
14. Miller TA, White ED, Kinley KA, Congleton JJ, Clark MJ. The effects of training history, player position, and body composition on exercise performance in collegiate football players. J Strength Cond Res 16: 44–49, 2002.
15. Nana A, Slater GJ, Hopkins WG, Burke LM. Effects of daily activities on dual-energy X-ray absorptiometry measurements of body composition in active people. Med Sci Sports Exerc 44: 180–189, 2012.
16. Potteiger JA, Smith DL, Maier ML, Foster TS. Relationship between body composition, leg strength, anaerobic power, and on-ice skating performance in Division I Men's Hockey Athletes. J Strength Cond Res 24: 1755–1762, 2010.
17. Santos DA, Dawson JA, Matias CN, Rocha PM, Minderico CS, Allison DB, et al. Reference values for body composition and anthropometric measurements in athletes. PLoS One 9: e97846, 2014.
18. Schutz Y, Kyle UU, Pichard C. Fat-free mass index and fat mass index percentiles in Caucasians aged 18–98 years. Int J Obes Relat Metab Disord 26: 953–960, 2002.
19. Siders WA, Lukaski HC, Bolonchuk WW. Relationships among swimming performance, body composition and somatotype in competitive collegiate swimmers. J Sports Med Phys Fitness 33: 166–171, 1993.
20. Silvestre R, West C, Maresh CM, Kraemer WJ. Body composition and physical performance in men's soccer: A study of a National Collegiate Athletic Association Division I team. J Strength Cond Res 20: 177–183, 2006.
21. Strugnell C, Dunstan DW, Magliano DJ, Zimmet PZ, Shaw JE, Daly RM. Influence of age and gender on fat mass, fat-free mass and skeletal muscle mass among Australian adults: The Australian diabetes, obesity and lifestyle study (AusDiab). J Nutr Health Aging 18: 540–546, 2014.
22. Takai Y, Kai T, Horio K, Nakatani M, Haramura M, Aoki T, et al. Lean body mass index is an indicator of body composition for screening prospective young adult soccer players. Football Sci 14: 8–14, 2017.
23. Thomas DT, Erdman KA, Burke LM. Position of the Academy of Nutrition and Dietetics, Dietitians of Canada, and the American College of Sports Medicine: Nutrition and athletic performance. J Acad Nutr Diet 116: 501–528, 2016.
24. Trexler ET, Smith-Ryan AE, Blue MNM, Schumacher RM, Mayhew JL, Mann JB, et al. Fat-free mass index in NCAA Division I and II Collegiate American Football Players. J Strength Cond Res 31: 2719–2727, 2017.
25. Trexler ET, Smith-Ryan AE, Mann JB, Ivey PA, Hirsch KR, Mock MG. Longitudinal body composition changes in NCAA Division I college football players. J Strength Cond Res 31: 1–8, 2017.
26. Ugarkovic D, Kukolj DMM. Standard anthropometric, body composition, and strength variables as predictors of jumping. J Strength Cond Res 16: 227–230, 2002.
27. VanItallie TB, Yang MU, Heymsfield SB, Funk RC, Boileau RA. Height-normalized indices of the body's fat-free mass and fat mass: Potentially useful indicators of nutritional status. Am J Clin Nutr 52: 953–959, 1990.