Relationship between Physical Activity and Variables of Body Composition among Saudi Female Medical Students Taking into Account the Body Mass Index

Introduction

Regular physical activity (PA) is one of the most important ways in which people can improve their physical and mental health1. The minimum level of PA recommended by the World Health Organization (WHO) per week is 150 min moderate- and/or 75 min vigorous-intensity physical activity, but that was not met by 27.5% of adults aged 18 years worldwide in 20222. A systematic review by Ferreira et al.,3 reported that the prevalence of physical inactivity had increased among university students, which might refer to physical and mental changes and university being a very competitive environment3, 4. In Saudi Arabia, a recent cross-sectional study found that approximately three in ten undergraduate students were physically inactive5. Another study on a sample of 375 Saudi female university students found that 62% did not meet the WHO recommendation for vigorous activity and 70% of the participants did not meet that for moderate activity6.

Insufficient levels of PA can be associated with the appearance of non-communicable diseases7. In addition, a recent systematic review reported that a low level of PA is a major contributing factor to the high prevalence of obesity8. Obesity is often a progressive, chronic disorder that affects health negatively and leads to stigma9. There are several complications related to obesity. For example, people who are overweight or obese often face social prejudice. They are also at higher risk of various life-threatening diseases, including cardiovascular disease, and an increased mortality rate10. The 2019 Global Burden of Disease analysis found that obesity-related disability-adjusted life years (DALYs) rose 0.48% annually from 2000 to 2019 and were predicted to increase by 39.8% from 2020 to 203011. In Saudi Arabia specifically, the prevalence of obesity (35%) is higher than the global average (13%)8. Unfortunately, with such a global increase in the prevalence of obesity, and the well-known consequences of this for population health and well-being, there is no clear solution, despite numerous reports and recommendations9.

Body mass index (BMI) is, due to its simplicity, routinely utilized as a robust measure to determine obesity. It is used to classify body mass into underweight, normal weight, overweight, and obese and has a modest correlation with adiposity12. Obesity is associated with excessive body fat mass and the BMI reflects the relationship between height and weight and should serve as an indicator for pre-estimating abnormal weight13. However, healthy individuals with high muscle mass can be misclassified as overweight or even obese, which makes BMI an inaccurate measure of body fat mass14-17. Despite a strong correlation between BMI and body fat % found in a large sample of the US adult population, BMI has limited diagnostic accuracy for diagnosing obesity, especially in individuals in the intermediate BMI ranges18.

Body composition assessment is used to describe body fat %, fat mass, and muscle mass. Numerous approaches can be taken to assess human body composition, which include analyzing bioelectrical impedance and body density, dual-energy X-ray absorptiometry, and total body water estimates19. Moreover, the risk of many chronic conditions associated with high body fat can be reduced by increasing PA20.

This study was undertaken in response to the high prevalence of both obesity and physical inactivity among young Saudi females and the need to promote PA among this population. The study assessed the variation among the participants' BMI groups in terms of their PA levels and determined the relationship between PA and body composition, taking BMI classification into account.

Materials and methods

Study Design, Participants, and Sample Size

This was an observational comparative study conducted for 3 months at the College of Applied Medical Sciences (Female campus), King Saud University (KSU). It was designed and reported in line with the STrengthening the Reporting of OBservational studies in Epidemiology (STROBE) guidelines21. Eighty female medical students aged 18­–25 were recruited. Participants were divided into four sub-groups according to their BMI, with n = 20 participants in each group: BMI < 18.5 = underweight; BMI 18.5–24.9 = normal weight; 25–30 = overweight; and BMI > 30 kg/m² = obese. Students were excluded from the study if they reported any of the following: if they suffered from a cardiovascular, metabolic, rheumatic, or vestibular disease22; if they had had problems with their lower limbs, such as injuries, in the 6 months prior to the study or had previous surgery on the legs, clinical instability in the knee or ankle, any amputation, contracture or deformity of the knee, hip replacements, or arthritis22, 23; a history of chronic conditions (such as musculoskeletal disorders and type II diabetes mellitus); any kind of visual impairments24; if they were smokers or recent ex-smokers (abstinent for less than 6 months); or pregnant23,24. A convenience sampling method was used. The sample size was calculated based on previous studies24, 25. It was found that 80 students needed to be included in the study, with a confidence interval of 95% and a margin of error of 5%. Eighty students were recruited and then divided into four groups: 20 students in each group: underweight, normal weight, overweight and obese group.

Outcome Measures

Body Composition

Measurements of selected body mass and body composition variables were performed for all participants. The participants were measured in the morning, at least 3 hours after waking up and eating or drinking. Bioelectrical impedance, using the Tanita Body Composition Analyzer BC- 418 (Tanita, Tokyo, Japan), was employed to assess body mass and body composition. The Tanita BC 418-MA is a reliable system with good test–retest reliability and excellent limits of agreement (< 1%). It is equivalent to skinfold thickness assessment and provides a non-invasive alternative that does not require operator training26.

All the guidelines/recommendations for the analysis of bioelectrical impedance were considered. The device was calibrated to account for clothing weight (0.2 kg). Then, data regarding the age, sex, and height of each participant were entered. A Seca weight and height scale (Seca GmbH & Co., Hamburg, Germany) was used to measure weight and height. Afterwards, each participant was asked to stand on the scale with bare feet and to place her hands on the places marked. The body composition analyzed by the device was based on differences in the ability of the body tissues (varying resistances) to conduct an electrical current due to the different water content. The variables measured included body weight (kg), body mass index (BMI, kg/m2), fat mass (FM, kg and %), fat-free mass (FFM, kg and %), and total body water (TBW, kg and %).

Physical Activity

The PA level was measured using the International Physical Activity Questionnaire – Short Form (IPAQ-SF). The IPAQ is a self-reported questionnaire that asks participants about the time they spend on various intensities of PA (vigorous, moderate, walking) for a minimum of 10 min at a time in the last 7 days27. The IPAG demonstrates moderate to good validity, as evidenced by correlation coefficients typically ranging from 0.50 to 0.80 when compared to objective measures of PA across 12 countries. Moreover, the IPAQ exhibits satisfactory reliability, with intraclass correlation coefficients (ICC) generally falling within the same range, supporting its consistent performance in assessing PA levels among diverse populations27. The total PA score is calculated using a metabolic equivalent (MET) task scored in minutes/week (MET-min/week)27, 28. PA was reported in this study as a continuous score (https://sites.google.com/site/theipaq/ ) and as a categorical score: < 600 = inactive; > 600 and < 3000 = sufficiently active; and > 3000 MET-min/week = very active.

Data Collection/Data Source

After receiving ethical approval for this study, an advertising banner with the researcher’s contact information (telephone number and email) was distributed in the corridors and in the main hall of the college. Students contacted the researcher by phone and/or email to ascertain their eligibility to be included in the study based on the inclusion criteria. Following this stage, the researcher booked an appropriate appointment for one session with each student separately. The researcher then explained the study procedure to each participant and asked her to sign a consent form and complete the IPAQ-SF. After that, height and weight were measured to calculate the BMI and body composition was assessed to identify the body mass, FM, and fat % of each participant.

Statistical Analysis

All statistical analysis was performed using the SAS statistical software package version 9.4 (SAS Institute Inc., Cary, NC, USA). Anthropometric and body composition measures were assumed a priori to follow a normal distribution and were summarized by providing means and standard deviations (SDs). The PA measures were summarized using median, minimum, and maximum values as they were not normally distributed. One-way analysis of variance (one-way ANOVA) was used to assess the differences among the four groups (underweight, normal body weight, overweight, and obese) in terms of anthropometric and body composition using an F-test. For PA measures, non-parametric one-way ANOVA was used to assess the differences between the corresponding Wilcoxon scores of the PA measures among the four groups using the Kruskal–Wallis test.

The correlation between body composition measures and PA was determined using the Spearman’s rho correlation test. In this study, the level of significance was accepted as P ≤ .05.

Results

Characteristics of the Participants

A total of 80 participants were included in this study. The demographics of the participants, classified by their BMI, are presented in Table 1. The mean (SD) age of all participants was 20.2 (1.15) years. All the participants (100%) were female.

Anthropometric and Body Composition Measures

There was a statistically significant difference among the groups in terms of anthropometric measures (weight and BMI) and all body composition measures of P < .0001. No statistically significant differences were seen among the groups for age and height (Table 1).

For the PA analysis, the Kruskal–Wallis test showed a significant difference among groups for vigorous PA and total PA score of P < .05. However, there was no significant difference among groups in respect of moderate PA, walking, and sitting time (Table 2).

Analysis of the IPAQ showed that approximately 46% of the participants were determined as being inactive. There was also a lack of a high level of PA among the participants in the normal weight and obese groups (0%). Only 10% of those in the underweight group and 15% of the overweight group were very active. Sixty percent of participants in the underweight group were sufficiently active, and 70% of those in the normal weight group were inactive (Figure 1).

Figure 1 Physical Activity Level of Participants.

The Spearman’s rho test detected a moderate negative correlation between total MET-min/week and FM as well as fat % in both the underweight and normal weight groups. Moderate positive correlations were found between total MET-min/week and BMI, FFM, and TBW in the obese group. P-values for these results ranged from .09 to .11 (Table 3).

Discussion

This study aimed to assess the PA levels among Saudi female medical students with varying BMI and determine the relationship between PA and body composition, taking into account BMI classification. About 46% of the participants were determined as being physically inactive. This result supports the findings of Alzamil et al.,29, who found that 46% of females in Saudi health sciences colleges were inactive, and Al-Drees et al.,30, who reported that 55% of the Saudi female medical students in their study were inactive. All these findings confirm the earlier results of a systematic review that reported a high prevalence of physical inactivity in Saudi youth31. In contrast, Aljehani et al.,6 showed that almost one third (29%) of a large sample (N = 375) of female students were inactive. All these studies were conducted at KSU. However, Aljehani et al.,6 included students from different specialties, whereas the others recruited students from medical/health sciences colleges. In practice, students from different specialties may have more time to participate in PA, whereas female medical students may suffer from lack of time due to academic load. However, the most commonly reported barrier by medical students to engaging in PA is lack of time 29,30,32,33.

However, the prevalence of physical inactivity in this study exceeds that of physical inactivity among medical students worldwide. For example, in Brazil, the prevalence of physical inactivity among N = 294 medical students was 34.4%34. Another study, in Bangalore, India, showed that only 15.4% of N = 259 medical students were inactive35. The lower level of physical inactivity among medical students could be explained by the trend in the students’ awareness of the importance of a healthy lifestyle36. The lower PA level among Saudi medical students found in this study might relate to several factors, such as environmental barriers, cultural differences, and lifestyle habits. For example, the climate in Saudi Arabia was reported as a barrier factor to PA6,31,37. Furthermore, incongruous urban designs, lack of footpaths, cycle ways, and safety issues force most students to use cars instead of walking to and from the the university6, 31.

The current study found variation among groups in terms of their PA levels, as participants with a low BMI had higher levels of PA. The findings showed a significant statistical difference in total PA score among the groups, those in the overweight group having the highest score with a median of 1386 (49.5, 12456). Those in the normal weight and obese groups did not participate in a vigorous level of PA at all. It was determined that more than half of those in the underweight group (60%) and half (50%) of those in both the overweight and obese groups participated in a level of moderate PA. In addition, about two thirds of those in the normal weight group (70%) and just under half (42.5%) of the overweight and obese participants were inactive. This finding is in contrast to that in Shahadan et al.,38 who found only 18.6% of the participants were inactive in a study that included N = 102 overweight and obese female Malaysian students38. Again, this variation could relate to environmental and cultural factors6,31,27.

There was no significant statistical difference detected in this study among groups in relation to moderate PA, walking, and sitting time. A significant difference was only determined among groups in respect of vigorous PA. However, walking was detected as the most common form of PA, especially for those in the underweight and overweight groups, with a maximum MET-min/ week of 5544, and this finding is consistent with the literature6,39,40. This can be justified by students walking between colleges to meet their friends, as well as walking in shopping malls. However, students might be walking out of necessity and not due to a preference for walking as a form of PA6.

The findings show a significant but weak negative correlation between fat % and PA for the whole sample regardless of their BMI. This result is in line with the literature41-44. With regard to BMI classification, a moderate negative but not significant correlation was detected between PA, FM, and fat % in the underweight and normal weight groups only. This supports findings of a study in China that reported that the inverse association between PA and total body fat % was found only in a particular sub-population who had normal BMI but a high body fat content in a sample of middle-aged Chinese adults42. The findings of the current study also show a moderate positive but not significant correlation between PA and BMI, FFM, and TBW in the obese group. This finding is in contrast to the hypothesis that when PA increases, the BMI value decreases. It is also in contrast to the literature41,43. For example, in a large sample of the middle-aged British general population, PA was inversely associated with BMI and body fat %. More PA was associated with a lower BMI and a lower body fat %, although even men and women who engaged in high-level PA were, on average, overweight43.

Although some studies suggested that PA had an impact on the participants’ BMI41,43, others reported that PA had an impact on body FM and fat %42,45. Bradbury et al.,43 found that people who more active had a lower body fat % even they have the same BMI. This might be because BMI is not a consistent measure of body fatness45.

As BMI does not consider the percentage of fat and non-fat body mass distribution in the total body weight, it does not always have explanatory power for assessing body weight. The current study only included female students and it is known that women have higher visceral fat % and genetically predisposed to higher body fat storage7. This may contribute to the limited accuracy of BMI as an index of body fatness7,44,45 and general health status, since bodybuilders, for example, have a high BMI, close to overweight, even if they have the lowest percentage of fat45.

Further, a moderate positive correlation between muscle mass and PA was detected only in members of the obese group in this study, 50% of whom were active at a moderate level. Therefore, it can be claimed that this increase in muscle mass might be a physical adaptation in response to moderate PA, as the body responds to a high volume of weekly PA by reducing fat %45. However, the current study only included female students and it is known that females are less physically active, on average, than males30,31.

Possible limiting factors for this study are that only female students were included and assessing PA (weekly training hours and types of activity) was based solely on a self-reported questionnaire. Participants were not given any specific instructions prior to using the bioimpedance device. As body fat values may be affected by some variables, including hydration status, exercise, and food consumption, controlling these factors might show slightly stronger associations between body fat % and PA45,46. Therefore, conducting a larger study that includes both male and female students using an objective outcome measure for assessing PA

and considering standardized factors of using bioimpedance is recommended.

Conclusion

The study found a significant weak negative correlation between PA and fat % among female students regardless of their BMI. There was variation among the groups. For the underweight and normal weight groups in relation to PA, there was a moderate negative correlation between total MET and FM as well as fat %. Moderate positive correlations were found between total MET-min/week and BMI, FFM, and TBW in the obese group. Variation among groups in terms of their PA levels confirms that BMI is not an accurate predictor of a person’s PA level. However, the findings of this study confirmed the need to find strategies to promote PA among female university students, regardless of their BMI.

Acknowledgments

The author would like to acknowledge Dr Hosam Alzahrani, Department of Physical Therapy, College of Applied Medical Sciences, Taif University, Taif, Saudi Arabia for providing feedback. Thanks also go to the MSc students for their assistance in the data collection. Lastly, the author expresses her sincere gratitude to all those who participated in the study.

Conflicts of Interest

The author declares no conflict of interest.

Financial support

The author received no funding for this work.

Ethical approval

The study was approved by the Institutional Review Board, College of Medicine, KSU (Reference no. 21/0924/IRB). A written consent form was taken from each participant.

References

1. Physical activity [Internet]. Who.int. [cited 2023 Dec 11]. Available from: https://www.who.int/news-room/fact-sheets/detail/physical-activity

1. Activity MP. Global status report on physical activity 2022 [Internet]. Who.int. World Health Organization; 2022 [cited 2023 Dec 11]. Available from: https://www.who.int/publications/i/item/9789240059153

1. Silva F, Mendonça RM, Azevedo CR, Memon R, Noll A, Noll PR. Barriers to high school and university students’ physical activity: a systematic review. PloS One. 2022 Apr 4; 17: e0265913.‏

1. Robazzi ML. Promotion of physical and mental health and well-being in the university environment. Rev Eletrônica Saúde Ment Álcool e Drog. 2019; 15:1–3.

1. Samarkandi OA. Prevalence of physical activity among healthcare students in King Saud University, Riyadh, Saudi Arabia. An observational study. INQ. 2022 May 16; 59: 00469580221100157.

1. Aljehani N, Razee H, Ritchie J, Valenzuela T, Bunde-Birouste A, Alkhaldi G. Exploring female university students' participation in physical activity in Saudi Arabia: a mixed-methods study. Front Public Health. 2022 Mar 18; 10: 829296.

1. Petrekova K, Borzenko N, Kovalova M, Strakos J, Klvacova A, Kotouckova I, Krutova J. Assessment of body mass index and body composition with physical activity and dietary preferences in university students. Pre-print, Research Square. 2022 Oct 13; doi:10.21203/rs.3.rs-2142520/v1

1. Salem V, AlHusseini N, Abdul Razack HI, Naoum A, Sims OT, Alqahtani SA. Prevalence, risk factors, and interventions for obesity in Saudi Arabia: asystematic review. Obes Rev. 2022 Mar 26; 23:e13448.

1. Kleinert S, Horton R. Obesity needs to be put into a much wider context. Lancet. 2019 Jan 27; 393:724–726.‏

1. Romero-Corral A, Montori VM, Somers VK, Korinek J, Thomas RJ, Allison TG, Lopez-Jimenez F. Association of bodyweight with total mortality and with cardiovascular events in coronary artery disease: a systematic review of cohort studies. Lancet. 2006 Aug 19; 368: 666–678.

1. Chong B, Jayabaskaran J, Kong G, Chan YH, Chin YH, Goh R, Chew NW. Trends and predictions of malnutrition and obesity in 204 countries and territories: an analysis of the Global Burden of Disease Study 2019. EClinicalMedicine. 2023 Feb 15; 57:101850.‏

1. Volgyi E, Tylavsky FA, Lyytikainen A, Suominen H, Alen M, Cheng S. Assessing body composition with DXA and bioimpedance: effects of obesity, physical activity, and age. Obesity (Silver Spring). 2008 Sept 6; 16:700–705.

1. Lebiedowska A, Hartman-Petrycka M, Blonska-Fajfrowska B. How reliable is BMI? Bioimpedance analysis of body composition in underweight, normal weight, overweight, and obese women. Ir J Med Sci. 2021 Oct 21;190:993–998.‏

1. Lee, W.S. Body fatness charts based on BMI and waist circumference. Obesity (Silver Spring). 2016 Jan; 24:245–249.

1. Marques-Vidal P, Bochud M, Mooser V, Paccaud F, Waeber G, Vollenweider P. Obesity markers and estimated 10-year fatal cardiovascular risk in Switzerland. Nutr Metab Cardiovasc Dis. 2009 Sept; 19:462–468.

1. Calling S, Hedblad BO, Engstrom G, Berglund G, Janzon L. Effects of body fatness and physical activity on cardiovascular risk: risk prediction using the bioelectrical impedance method. Scand J Public Health. 2006 Dec; 34:568–575.

1. Lahmann PH, Lissner L, Gullberg B, Berglund GA. prospective study of adiposity and all-cause mortality––the Malmo diet and cancer study. Obes Rev. 2002 May; 10:361–369.

1. Abel RC, Somers VK, Sierra-Johnson J, Thomas RJ, Bailey KR, Collazo-Clavell ML, Allison TG. Accuracy of body mass index to diagnose obesity in the US adult population. Int J Obes (Lond). 2008 Feb 19; 32:959–966.

1. Duren DL, Sherwood RJ, Czerwinski SA, Lee M, Choh AC, Siervogel RM, Chumlea W C. Body composition methods: comparisons and interpretation. J Diabetes Sci Technol. 2008 Nov 1; 2:1139–1146.‏

1. Buchowski MS, Townsend KM, Chen KY, Acra SA, Sun M. Energy expenditure determined by self-reported physical activity is related to body fatness. Obes Res. 1999; 7:23–33.

1. Von Elm E, Altman DG, Egger M, Pocock SJ, Gotzsche PC, Vandenbroucke JP. STrengthening the Reporting of OBservational studies in Epidemiology (STROBE) statement: guidelines for reporting observational studies. BMJ. 2007 Oct 20; 335:806–808.

1. Greve J, Alonso A, Bordini AC, Camanho GL. Correlation between body mass index and postural balance. Clinics. 2007; 62:717–720.

1. Delfa-de la Morena JM, Castro EA, Rojo-Tirado MA, Bores-Garcia D. Relation of physical activity level to postural balance in obese and overweight Spanish adult males: a cross-sectional study. Int J Environ Res Public Health. 2021 Aug 05; 18:8282.

1. Mondal R, Banik PC, Ritu RB, Mashreky SR, Zaman MM. Associations of obesity with balance and gait among young adults in Bangladesh. J Xiangya Med. 2021; 6:1–8.

1. Sarkar A, Singh M, Bansal N, Kapoor S. Effects of obesity on balance and gait alterations in young adults. Indian J Physiol Pharmacol. 2011; 55:227–233.

1. Kelly JS, Metcalfe J. Validity and reliability of body composition analysis using the Tanita BC418-MA. J Exerc Physiol Online. 2012 Dec; 15:74–83.

1. Craig CL, Marshall AL, Sjostrom M, Bauman AE, Booth ML, Ainsworth BE, Oja P. International physical activity questionnaire: 12-country reliability and validity. MMSE. 2003; 35:1381–1395.

1. Haskell WL, Lee IM, Pate RR, Powell KE, Blair SN, Franklin BA, Bauman A. Physical activity and public health: updated recommendation for adults from the American College of Sports Medicine and the American Heart Association. Circulation. 2007; 116: 1081–1093.

1. Alzamil HA, Alhakbany MA, Alfadda NA, Almusallam SM, Al-Hazzaa HM. A profile of physical activity, sedentary behaviors, sleep, and dietary habits of Saudi college female students. J Family Community Med. 2019 Jan-Apr; 26:1–8.‏

1. Al-Drees A, Abdulghani H, Irshad M, Baqays AA, Al-Zhrani AA, Alshammari SA, Alturki NI. Physical activity and academic achievement among the medical students: a cross-sectional study. Medical Teacher. 2016 Mar 17; 38:S66–S72.‏

1. Al-Hazzaa HM. Physical inactivity in Saudi Arabia revisited: a systematic review of inactivity prevalence and perceived barriers to active living. Int J Health Sci (Qassim). 2018 Nov-Dec; 12:50–64.‏

1. Alfhaid F, Alzahrani M, Almansour M, Alghamdi T, Ansari T, Sami W, Alhugail FS. Lifestyle of health sciences students at Majmaah University, Saudi Arabia. Australas Med J. 2017; 10:111–116.

1. Alissa EM, Alsawadi H, Zedan A, Alqarni D, Bakry M, Hli NB. Knowledge, attitude and practice of dietary and lifestyle habits among medical students in King Abdulaziz University, Saudi Arabia. Int J Nutr Food Sci. 2015 Oct 28; 4:650–655.

1. Schlickmann DW, Kock KS. Level of physical activity knowledge of medical students in a Brazilian university. J Lifestyle Med. 2022 Jan 31; 12:47–55.‏

1. Padmapriya K, Krishna P, Rasu T. Prevalence and patterns of physical activity among medical students in Bangalore, India. Electron Physician 2013 Feb 1; 5:606–610.‏

1. Bergier B, Tsos A, Bergier J. Factors determining physical activity of Ukrainian students. Ann Agric Environ Med. 2014; 21:613–616.

1. Qahwaji DM. Lifestyle behaviours, dietary habits, physical activity and biochemical measurements in Saudi University students. Med Sci. 2023 Apr 24; 27:e198ms2940.‏

1. Shahadan SZ, Bolhan NS, Ismail MF. The association between physical activity status and mental wellbeing among overweight and obese female university students.‏ Mal J Med Health Sci. 2022 Dec; 18:80-86.

1. Al Zahib YH, Baarimah H. Physical activity profile among Saudi adults in Abha City, Saudi Arabia. Middle East J Family Med. 2020 Aug 1;7:37.

1. Al-Hazzaa HM. Health-enhancing physical activity among Saudi adults using the International Physical Activity Questionnaire (IPAQ). Public Health Nutr. 2007 Jan 1; 10:59–64.

1. You HW, Tan PL, Mat Ludin AF. The relationship between physical activity, body mass index and body composition among students at a pre-university centre in Malaysia. IIUM Med J Malays. 2020 Jul; 19:83–88.‏

1. Zou Q, Su C, Du W, Ouyang Y, Wang H, Wang Z, Zhang B. The association between physical activity and body fat percentage with adjustment for body mass index among middle-aged adults: China health and nutrition survey in 2015. BMC Public Health. 2020 May 19; 20:1–12.‏

1. Bradbury KE, Guo W, Cairns BJ, Armstrong ME, Key TJ. Association between physical activity and body fat percentage, with adjustment for BMI: a large cross-sectional analysis of UK Biobank. BMJ Open. 2017; 7:e011843.‏

1. Zanovec M, Lakkakula AP, Johnson LG, Turri G. Physical activity is associated with percent body fat and body composition but not body mass index in White and Black college students. Int J Exerc Sci. 2009 Jul 15; 2:175–185.‏

1. Zaccagni L, Barbieri D, Gualdi-Russo E. Body composition and physical activity in Italian university students. J Transl Med. 2014 May 9; 12:1–9.‏

1. Khalil SF, Mohktar MS, Ibrahim F. The theory and fundamentals of bioimpedance analysis in clinical status monitoring and diagnosis of diseases. Sensors. 2014 June 3; 14: 10895–10928.