Syeda Sadia Fatima ( Department of Biological and Biomedical Sciences, Aga Khan University, Karachi, Pakistan. )
Rehana Rehman ( Department of Biological and Biomedical Sciences, Aga Khan University, Karachi, Pakistan. )
Bushra Chaudhry ( Department of Biological and Biomedical Sciences, Aga Khan University, Karachi, Pakistan. )
Objective: To compare two methods of classifying obesity based on body mass index and body fat percentage.
Methods: The cross-sectional study was conducted from November 2012 to August 2013 at Jinnah Postgraduate Medical Centre, Karachi.
Male and female volunteers between the ages 15-65 years were selected using simple random sampling. They were classified into different groups for body mass index and body fat percentage measured through bioelectrical impedance scale. The subjects were sub-grouped into underweight, normal weight, overweight and obese. SPSS 11 was used for statistical analysis.
Results: The mean age of the 828 healthy volunteers was 25.67±10.10 years. A total of 552(66.6%) subjects had a higher body fat percentage and were misclassified by body mass index. Only 276(33.3%) subjects had body fat percentage values corresponding to the body mass index classification. The difference in terms of categorising obesity was highly significant (p<0.001).Both body mass index and body fat percentage showed positive correlation with age (r=0.144; p=0.001) (r=0.261; p=0.001) and weight (r=0.578; p=0.001) (r=0.444; p=0.001) respectively. Moreover body fat percentage showed a significant positive association with gender (r=0.109; p=0.027) whereas BMI did not.
Conclusions: Body fat percentage should be incorporated for a better understanding as well as categorising of obesity.
Keywords: BMI, Overweight, Body fat percentage, Obesity. (JPMA 64: 1225; 2014).
Obesity is acknowledged as one of the burning public health problems reducing life expectancy and quality of life.1 Many factors influence the obesity epidemic, including genetic susceptibility, socioeconomic, cultural, behavioural, environmental factors, imbalance between food intake and lack of physical activity.2 Adiposity, however, irrespective of weight and or body mass index (BMI) value, is believed to be a primary risk factor for diabetes and cardiovascular disease,3 providing a rationale for the use of methods which measure body fat directly. The frequently used anthropometric measures such as BMI, waist circumference (WC), waist-to-hip ratio (WHR), hip circumference (HC) have been anticipated to define obesity. However, their limitation to assess degree of fatness in individuals with difference in muscular build is well recognised.4
The Quetelet\'s index is used far more commonly as a surrogate measure of fatness than body fat percentage (BF%) to define obesity5 whereas it is not a measurement of adiposity, but merely an imprecise mathematical estimate.6 Use of BMI alone to classify individuals may result in misclassification because of the varying contributions of bone mass, muscle mass, and fluid to body weight.7 In Europeans, a BMI of 30 correlates with about 25% body fat in males and 30% body fat in females,8 while for the same age, gender and BMI, South Asians have an increased per cent body fat, lesser lean mass, skeletal muscle and bone mineral content along with a higher risk for cardiovascular diseases.9,10
Debate over the value of BMI for the estimation of body fat has recently led investigators to recommend the use of new technologies for the direct measurement of body fat, especially in epidemiological research,11 to account for the differences in body weight. It has been found that amount of body fat rather than excess weight determines health-associated risks.6
The commonly used methods for classifying obesity and overweight fail to appropriately identify the burden of underlying disease, especially in Pakistani population. This study was planned to relate the misclassification of obesity by BMI in contrast with BF%. This might help to provide a working approach to incorporating body fat measurement as a proper obesity indicator.
Subjects and Methods
The cross-sectional study was conducted from November 2012 till August 2013 at Jinnah Postgraduate Medical Centre, Karachi.
In order to achieve 80% power with a 15% estimated prevalence of disease in project area and a two-sided 5% level of significance, the minimum sample size calculated for the study was 260.12 We recruited828 male and female volunteers between the ages 15-65 years by simple random sampling.
All the participants were asked to sign written, informed consent. Subjects were excluded if they had a history of recent acute illness (e.g, pneumonia, myocardial infarction or dehydration), had a chronic condition (e.g, cancer, uncontrolled high blood pressure, or collagen vascular disease), pregnancy and/or menstrual period and vigorous activity (12 hours before) body fat estimation. The study was approved by the ethics review committee of Jinnah Postgraduate Medical Centre\'s Basic Medical Sciences Institute in Karachi.
BMI of the study subjects was calculated by dividing weight with height squared (kg/m2).5 The BF% was measured using Diagnostic Scale BG55 (Beurer Germany) through bioelectrical impedance analysis.
The subjects were classified for BMI and BF% as follows: according to BMI criteria for South Asian population as normal weight (BMI: 18-22.9 kg/m2), overweight (BMI: 23-25.9kg/m2), and obese (BMI >26kg/m2) subjects13 and according to the BF% scale: Males — normal weight (BF%: 12-22%), overweight (BF%: 22.1-27 %), and obese (BF%: = 27.1); Females — normal weight (BF% 17-27%), overweight (BF% 27.1-32%), and obese (BF% > or=32.1).14 Subjects falling below the normal values were classified as underweight for both BMI and BF%.
A descriptive statistical analysis of continuous variables was performed using SPSS 11. Statistical comparisons of categorical variables (BMI and BF%) were computed using Pearson chi square test. Pearson correlation coefficient was applied to check the correlation of BMI and BF% with study parameters. Continuous variables were presented as Mean±SD and percentages and compared by student\'s t-test. In all statistical analysis, p<0.05 was considered significant.
Of the 828 subjects in the study, 426(51.44%) were females and 402(48.55%) were males. The overall mean age was 25.67±10.10 years, BMI was 27.79±8.57kg/m2, and BF% was 24.61%±7.61. According to BMI classification, 42(5%) subjects were underweight, 238(29%) normal weight, 150(18%) overweight and 398(48%) obese. According to BF% classification, 68(8%) subjects were underweight, 160(19%) normal weight, 242(29%) overweight and 358(43%) obese. However, only 276(33%) participants could be correctly identified in similar categories by both the BMI and BF% criteria (Table-1).
The difference in terms of categorising obesity between the two was highly significant (2 df(1)=9/43.47; p<0.001).
Both BMI and BF% showed positive correlation with age (r=0.144; p=0.001) (r=0.261; p=0.001) and weight (r=0.578; p=0.001) (r=0.444; p=0.001) respectively. Moreover, BF% showed a significant positive association with gender (r=0.109; p=0.027) whereas BMI did not (Table-2).
The prevalence of overweight and obesity in developing countries, especially in Pakistani population, has been reported to be 25% and 10% respectively13 with an increased trend of obesity in youngsters. The recognition of true obesity is thus important to identify potential threats of associated health disorder that bear out economic burden on society.15-18
BMI in this regards is considered to be a gauge of obesity and fitness in various cultures andnarrates incidence and prevalence of obesity with regard to mortality and morbidity rates in ethnic populations.19 The BMI cut-off values for the detection of obese, however, have changed from >30kg/m2 to >26 kg/m2 for South Asian population specifically,20 with fewer risks at BMI less than 18·5kg/m2 and increased risk with BMI 23-27·5kg/m2, and maximum risks when values exceed 27·5kg/m2. Besides, the classification of obesity on the basis of BMI is subjective to diversity with respect to the variation of population. Here we propose and agree that Asian populations need to be evaluated by their own cut-off values in terms of BMI, BF%, and associated health risks.13
The results of this study showed 29% subjects to be normal weight by BMI category. However, lesser number of individuals, 19%, fell into the same category by BF%. Only 14% subjects were deemed normal weight by both. An interesting finding was that in this group of normal weight for BMI individuals, 3% turned out to be under-weight, 12% as overweight and 9% were obese when their body fat was measured. This indicated a false positive result for 28% subjects who may be left unnoticed for detection of disorders, if BF% was not measured simultaneously. This means that in order to define normal weight, both criteria should be taken into account. A recent large-scale study21 on UK adults has shown that the association between BMI and BF% is not applicable, particularly when BMI is less than 25kg/m2. Studies22 found that high BF% was associated with increased cardiovascular risk regardless of BMI whose categorisation resulted in an underestimation of subjects with cardiovascular risk factors.23,24
Identification of true underweight is nonetheless important to recognise nutritional deficiencies, immune disorders, brittle bones, arthritic changes and compromised fertility. In this study, true underweight when both criteria were taken into account were 3%, while 5% were underweight by BMI and 8% by BF%. This shows that BF% is also a better predictor of underweight who were misclassified by BMI alone. People with BMI below 18.5kg/m2 are found to be associated with the above-mentioned risks along with higher death rates.19
Maximum number of subjects in our study were declared obese by both methods of estimation. The true obese declared by both BMI and BF% were 72%. Among them, 48%had a BMI >26kg/m2. However, when BF% was measured, the number decreased to 43%; concurring with our hypothesis that by taking BMI alone into consideration, more individuals can be marked obese erroneously since BMI is not a measuring factor for the muscle mass.
Another interesting finding in our study was that greater proportion of subjects 29% were declared overweight by BF% compared to 18% by BMI (p=<0.001). Interestingly, the misclassification of obesity on the basis of BMI was found to affect males more, which is contradictory to the results of an earlier study3 which found that BMI-defined obesity (BMI >30kg/m2) was present in 21% of men and 31% of women, but BF%-defined obesity was found in 50% of men and 62% of women. It also found that BMI failed to discriminate between BF% and lean mass in the overweight, or intermediate, range of BMI (25-29.9kg/m2).We believe that the more serious complications from increased adiposity are implicated early in South Asian men and hence detection of extent of adiposity is extremely important for them.
These findings support our concerns that typically normal BMI may conceal underlying excess adiposity characterised by an increased percentage of fat mass and reduced muscle mass. Thus we suggest that the accuracy of BMI in diagnosing obesity is limited, particularly for individuals in the intermediate BMI ranges.
The study emphasises the need to measure BF% together with BMI and catalogue misclassified persons especially for categorisation. Early detection of obesity by simple, quick, safe, low-cost measures of body fatness by bioelectrical impedance analysis (BIA) is thus required to address the related metabolic risk association with underlying disease burden. There is also a need to develop provisional, population based cut-off values for BF% in order to fill information gap because no comparable percentage body fat ranges that exist for evaluation of potentially misclassified subjects referred to body-composition analysis. The limitation of the current study is that since this is the first study conducted in local population, it could not verify the validity of the sampled population.
We recommend that awareness about the impact of higher BMI and BF% as risk factors with early commencement of disease and disorders should be generated among the masses and periodic assessment of body weight and BF% in schools, colleges, universities and workplaces should be reinforced to prevent obesity.
To limit the discrepancy among classification of false negative and false positive values in our population, body fat measurement should be incorporated for a better understanding and classification of obesity. This would be helpful in lowering the disease burden.
1. Rehman R, Ahmed S, Syed S. Exercise induced physiological changes in medical students with different BMI. JIAR 2010 ; 10: 10-5.
2. Sheikh S, Rehman R, Ezdi L. Selection of active life styles by male / female healthy young medical students. Med Channel 2011; 18: 9-12.
3. Romero-Corral, Somers VK, Sierra-Johnson J, Thomas RJ, Collazo-Clavell MI, Korinek J, et al. Accuracy of Body Mass Index In Diagnosing Obesity In The Adult General Population. Int J Obes (London) 2008; 32: 959-66.
4. Garrow JS, Webster J. Quetelet\'s index (W/H2) as a measure of fatness. Int J Obes 1985; 9: 147-53.
5. Shah NR, Braverman ER. Measuring Adiposity in Patients: The Utility of Body Mass Index (BMI), Percent Body Fat, and Leptin. PLoS ONE 2012; 7: e33308.
6. WHO expert consultation. Appropriate body-mass index for Asian populations and its implications for policy and intervention strategies. Lancet 2004; 363: 157-63.
7. Mascie-Taylor CG, Goto R. Human variation and body mass index: a review of the universality of BMI cut-offs, gender and urban-rural differences, and secular changes. J Physiol Anthropol 2007; 26: 109-12.
8. Rush EC, Freitas I, Plank LD. Body size, body composition and fat distribution: comparative analysis of European, Maori, Pacific Island and Asian Indian adults. The Br J Nutr 2009; 102: 632-41.
9. Lear SA, Birmingham CL, Chockalingam A, Humphries KH. Study design of the Multicultural Community Health Assessment Trial (M-CHAT): a comparison of body fat distribution in four distinct populations. Ethin Dis 2006; 16: 96-100.
10. Roubenoff R, Dallal GE, Wilson PW Predicting body fatness. The body mass index vs. estimation by bioelectrical impedance. Am J Public Health 1995; 85: 726-8.
11. Arroyo M, Rocandio AM, Ansotegui L, Herrera H, Salces I, Rebato E. Comparison of predicted body fat percentage from anthropometric methods and from impedance in university students. Br J Nutr 2004; 92: 827-32.
12. Bland JM. An Introduction to Medical Statistics, 3rd ed. Determination of sample size. Oxford: Oxford University Press, 2000; pp 335-47.
13. Jafar TH, Levey AS, White FM, Gul A, Jessani S, Khan AQ, et al. Ethnic differences and determinants of diabetes and central obesity among South Asians of Pakistan. Diabet Med 2004; 21: 716-23.
14. Gallagher D, Heymsfield SB, Heo M, Jebb SA, Murgatroyd PR, Sakamoto Y. Healthy percentage body fat ranges: An approach for developing guidelines based on body mass index. Am J Clin Nutr 2000; 72: 694-701.
15. Rehman R, Hussain Z, Fatima SS. Effect of weight status on pregnancy outcome in intra cytoplasmic sperm injection. Iran J Repro Med 2013; 11: 717-24.
16. Fatima SS, Bozaoglu K, Rehman R, Memon AS, Alam F. Elevated chemerin levels in Pakistani men: an interrelation with metabolic syndrome phenotypes. PLoS ONE 2013; 8: e57113.
17. Fatima SS, Rehman R, Alam F. Life style trends in a group of known hypertensive: a questionnaire base survey. Pak Armed Forces Med J 2012; 62: 296-300.
18. Fatima SS, Memon AS, Rehman R, Zuberi NA, ParveenT. Chemerin levels and body fat percentage in Pakistani males. Pak J Biochem Mol Biol 2012; 45: 189-91.
19. Flegal KM, Graubard B I, Williamsen D F, Mitchell G H. Excess deaths associated with underweight, overweight, and obesity. JAMA 2005; 293: 1861-67.
20. Klein S, Allison DB, Heymsfield SB, Kelley DE, Leibel RL, Nonas C, et al. Waist Circumference and Cardiometabolic Risk: a Consensus Statement from Shaping America\'s Health: Association for Weight Management and Obesity Prevention; NAASO, the Obesity Society; the American Society for Nutrition; and the American Diabetes Association. Am J Clin Nutr 2007; 15: 1061-7.
21. Meeuwsen S, Horgan GW, Elia M. The relationship between BMI and percent body fat, measured by bioelectrical impedance, in a large adult sample is curvilinear and influenced by age and sex. Clin Nutr 2010; 29: 560-6.
22. Zeng Q, Dong SY, Sun XN, Xie J, Cui Y. Percent body fat is a better predictor of cardiovascular risk factors than body mass index. Braz J Med Biol Res 2012; 45: 591-600.
23. Gómez-Ambrosi J, Silva C, Galofré JC, Escalada J, Santos S, Millán D, et al. Body mass index classification misses subjects with increased cardiometabolic risk factors related to elevated adiposity. Int J Obes (Lond) 2012; 36: 286-94.
24. Jackson AS, Stanforth PR, Gagnon J, Rankinen T, Leon AS, Rao DC, et al. The effect of sex, age and race on estimating percentage body fat from body mass index: the heritage family study. Int J Obes Relat Metab Disord 2002; 26: 789-96.