Title:Body adiposity index performance in estimating body fat in a sample of severely obese Brazilian patients

Transcription

1 Author's response to reviews Title:Body adiposity index performance in estimating body fat in a sample of severely obese Brazilian patients Authors: Giliane Belarmino (giliane85@hotmail.com) Lilian M Horie (mikahorie@gmail.com) Priscila C Sala (sala.priscila@gmail.com) Raquel SMM Torrinhas (torrinhas@uol.com.br) Steven B Heymsfield (steven.heymsfield@pbrc.edu) Dan L Waitzberg (dan.waitzberg@gmail.com) Version:2Date:11 November 2015 Author's response to reviews: see over

2 1 To Nutrition Journal Editorial Ref: Body adiposity index performance in estimating body fat in a sample of severely obese Brazilian patients. Giliane Belarmino, Lilian M. Horie, Priscila C. Sala, Raquel S. Torrinhas, Steven B. Heymsfield and Dan L. Waitzberg (MS: ) Dear Editor, Thank you for the opportunity to answer the remarks of the Reviewers 1 and 2 concerning our manuscript "Body adiposity index performance in estimating body fat in a sample of severely obese Brazilian Patients. The Reviewers comments demonstrate their critical in-depth analysis of our study and also a strong scientific background, which have undoubtedly contributed to improve the scientific quality of our manuscript. The text of our manuscript was carefully reviewed, aiming to answer all the reviewer s questions and attend their suggestions, as described below. During the manuscript revision, our statistician (LD) actively provided intellectual contribution to attend some of these questions and, therefore, we would like to include him as a co-author. REVIEWER 1 Comment 1: The question is well defined to evaluate the capacity for BAI to estimate body fat percent for severely obese Brazilian patients. This study uses Air displacement techniques to compare the BAI method in obese Brazilian men and women. The question is not new as there have been previous studies attempting to validate the BAI method. Additionally, there was a previous study of the correlation between BAI and DXA in Brazilian obese women. I recommend adding more justification for why it is important to evaluate BAI for such a narrow target population as this was not the original intension of the development of the method bye Bergman et al. (2011) though it does have value.

3 2 Answer: Thank you for the opportunity to discuss the relevance of our study question. The adequate identification of fat amount is relevant to estimate the risk to develop different complications potentially associated with overweight, such as metabolic syndrome. In this issue, the body adiposity index (BAI) was found to be sensitive in estimating body fat (BF) %, compared to the reference dual-energy X-ray absorptiometry (DXA). Since BAI development by Bergman et al. its application has been addressed in several studies aiming to test whether this simple, practical and inexpensive tool can have a good performance to estimate BF% in other population than that where BAI was developed and validated. This effort is because most public hospitals from countries with limited financial resources for health spending are eager for simple and affordable methods for assessing body fat of their patients. Particularly, severe obese patients make up a population at risk of complications associated with excess body fat and whose BF% estimation is constantly limited due to the unavailability of methods which happen to contain the size and weight of patients with this profile in clinical practice. Therefore, we tested whether BAI could be adopted as a tool with a good performance to estimate BF% in severely obese Brazilian patients. As the Reviewer 1 corrected point out, another study has addressed this question in obese Brazilians. However, in fact this early study did not focuses on obese but overweight patients (mean BMI = 26.9) and our study focuses only on obese patients (BMI 30 kg/m 2, mean BMI = 47.45), by considering this population the most susceptible for complications associated with excessive body fat and lacking in simple methods that enable the analysis of their BF%. The choice of obese population implied in using air displacement plethysmography (ADP) as the gold standard method for estimating body fat in our study because the dual energy xray absorptiometry (DXA) used for this purpose in Bergman's study can analyze only patients with limited weight and size. ADP has been validated in a number of studies for normal weight populations 1,2, and mostly in obese populations 3,4 by being suitable and producing realistic body composition data for patients with BMI over 40 kg/m² 5,6. In an obese population similar to that we presently studied, Hames et al (2014) compared DXA and ADP and found strong agreement on fat free mass (FFM) estimation (r = 0.95, P < 0.001) 3. In addition, ADP was able to estimate a higher FFM than that estimated by DXA

4 3 (P < 0.001), suggesting a higher sensibility of ADP to estimate FFM in obese patients with high BMI 3. Furthermore, there was a strong agreement between the estimations of fat content provided by these both methods regardless of the expression unit (kg and %) 3. Based on Reviewer 1 comments we realize the consistent need to highlight the characteristics of our population as the main differential our study. For this purpose we included some information at the introduction (page 5, lines here, highlighted in red), as follows: ( ) both of the databases that were used for the development and validation of the BAI included groups of subjects with varying degrees of BF% (BMIs average: ~ 30 kg/m², ranging from normal to obese) 8. It is worth noting that more severely obese patients are at higher complications risk associated with excess body fat. Their large size and weight limits de performance of simple (such as skin fold) and more sophisticated (such as DXA) methods for the clinical BF% estimation. Therefore, we tested whether BAI could have a good performance to estimate BF% in severely obese Brazilian patients. References for this answer: 1. Fields DA, Goran MI, McCrory MA (2002) Body-composition assessment via air-displacement plethysmography in adults and children: a review. Am J Clin Nutr 75: McCrory MA, Gomez TD, Bernauer EM, Molé PA (1995) Evaluation of a new air displacement plethysmograph for measuring human body composition. Med Sci Sports Exerc 27: Hames KC, Anthony SJ, Thornton JC, Gallagher D, Goodpaster BH (2014) Body composition analysis by air displacement plethysmography in normal weight to extremely obese adults. Obesity (Silver Spring) 22: Bedogni G, Agosti F, De Col A, Marazzi N, Tagliaferri A, et al. (2013) Comparison of dual-energy X-ray absorptiometry, air displacement plethysmography and bioelectrical impedance analysis for the assessment of body composition in morbidly obese women. Eur J Clin Nutr 67:

5 4 5. Ginde SR, Geliebter A, Rubiano F, Silva AM, Wang J, et al. (2005) Air displacement plethysmography: validation in overweight and obese subjects. Obes Res 13: Petroni ML, Bertoli S, Maggioni M, Morini P, Battezzati A, et al. (2003) Feasibility of air plethysmography (BOD POD) in morbid obesity: a pilot study. Acta Diabetol 40 Suppl 1: S Comment 2: The methods are appropriate and well described; however, the authors who are attempting to describe the level of performance of BAI in predicting percent body fat from ADP are comparing this to the results of the original study by Bergman et al. (2011) that used DXA as the reference method. Therefore, the authors need to provide the correlation coefficients of %BF as calculated between ADP and DXA from other studies, so that the readers can make the study results translatable between ADP and DXA. Answer: We thank the Reviewer 1 for this relevant suggestion. As explained above, the inclusion of severe obese patients in our sample implied in using air displacement plethysmography (ADP) as the gold standard method for estimating body fat in our study because the dual energy xray absorptiometry (DXA) used for this purpose in Bergman's study can analyze only patients with limited weight and size. In an obese population similar to that we presently studied, Hames et al (2014) compared DXA and ADP and found strong agreement on fat free mass (FFM) estimation (r = 0.95, P < 0.001) 1. In addition, ADP was able to estimate a higher FFM than that estimated by DXA (P < 0.001), suggesting a higher sensibility of ADP to estimate FFM in obese patients with high BMI 1. Furthermore, there was a strong agreement between the estimations of fat content provided by these both methods regardless of the expression unit (kg and %) 1. Based on the Reviewer 1 suggestion we realize the consistent need to add some information showing that our study results are translatable between ADP and DXA. This information was added at discussion section (page 12, lines here highlighted in red), as follows:

6 5 A limitation of the present study was the relatively small sample of Brazilian obese patients with a BMI of 30 kg/m ² (range: kg/m²). In addition, the inclusion of severe obese patients in our sample implied in using ADP, and not DXA, as the gold standard method for BF% estimation. While ADP has been shown to be suitable to adequately access body composition in patients with BMI over 40 kg/m², the use of DXA usually is applied only in patients with limited weight and size. 26,27 In an obese population similar to our study, Hames et al (2014) found a strong agreement between DXA and ADP performance in estimating fat content regardless of its expression unit (kg and %). 28 Reference for this answer: 1. Hames KC, Anthony SJ, Thornton JC, Gallagher D, Goodpaster BH (2014) Body composition analysis by air displacement plethysmography in normal weight to extremely obese adults. Obesity (Silver Spring) 22: Comment 3: The authors need to include a table with descriptive characteristics of the patients (demographic age, ethnicity - and anthropometric variables measured) in the study, with a breakdown by sex to aid in replicating the study. Answer: Thank you for this great suggestion that was promptly attended in results section (page 8, line ), as follows: Table 1 provides the baseline demographic and anthropometric data of the 72 obese patients assessed. Table 1- Demographic and anthropometric data of obese patient sample Gender Female (n = 53) Male (n = 19) Total (n = 72) p value¹ n % n % N % Age (years) 44,7 11,8 36,7 12,2 42,6 12,3 0,015 (1) Body weight (kg) 118,2 20,7 154,08 27,37 127,68 27,5 <0,001 (1) Height (m) 1,6 0,1 1,8 0,1 1,6 0,1 <0,001 (1) BMI (kg/m²) 46,8 6,4 49,4 7,4 47,4 6,7 0,146 (1)

7 6 Abdominal circumference (cm) 132,1 14,6 152,3 16,1 137,5 17,4 <0,001 (1) Hip circumference (cm) 134,3 13,8 137,6 13,9 135,2 13,8 0,382 (1) Waist/hip ratio 1,0 0,1 1,1 0,1 1,0 0,1 <0,001 (1) Race* white , , ,9 0,644 (2) Race* black/brown ,8 7 36, ,1 BF% ADP 53,4 4,7 48,6 5,9 52,1 5,4 0,001 (1) BF% - BAI 49,2 6,5 40,8 5,4 47,0 7,2 <0,001 (1) (1) Student t-test.(2) Chi square. BMI, body mass index; ADP, air displacement plethysmography; BF%, body fat percentage; BAI, body adiposity index. *Note for Reviewer 1: Ethnicity was substituted by race, because the different subgroups that compound Brazilian population are nationally classified according to color or race, while ethnicity includes cultural variables, as religion adopted. Comment 4 and 5: In the Subjects and methods section under patients, the authors should describe at greater length information on the ethnicity of the patients, especially since they provide a figure of the relationship between skin color and %BF from ADP. Skin color is not the same as ethnicity. On figure 2, the authors should change the labels from white vs black/brown to the Ethnic classification congruent with the individuals as mentioned in the discussion (Afro-Brazilian, Amerindian, etc.). Answer: Thank you for this important comment and suggestion. Due the diversity of the Brazilian population, which includes a mixture of Amerindian, European, and African genetic backgrounds is very difficult to classify its ethnicity. For instance, is common a Brazilian black subject to have Ameridian and/or Europian genetics mixed with his African genetic background. Therefore, the classification provided by the Brazilian Institute of Geography and Statistics (IBGE, the national organ responsible to classify population origin) is performed according to color or race into the following categories: white, black/brown, yellow and indigenous, while ethnicity includes cultural variables, as the religion adopted. ¹ Indeed, in our study we now opted to describe sample ethnicity as race by considering skin color for classification. Our sample consisted of 36.1% black/brown and 63.9% white, as adequately provided in the new Table 1.

8 7 In order to adequately address the ethnicity classification in our study, we added a new sentence at Methods section (page 6, lines ), as follows: According to the Brazilian Institute of Geography and Statistics (IBGE), sample ethnicity was self-reported as race by considering skin color for classification into the following categories: white, black/brown, yellow and indigenous 14. In addition, we changed the term ethnicity by race when describing our results, including in Figure 2 (now, the new Figure 3), as follows: Figure 3. Scatter plot of hip circumference vs. BF% by ADP among white and black/brown obese Brazilian patients. Solid line represents the Local Polynomial Regression fitting that is bounded by the 95% confidence band. Finally, we also changed the sentence addressing the ethnicity issue at the discussion section of our manuscript (page 9-10, lines ), as follows: From: In addition, the lack of a significant correlation between hip circumference and ADP-determined BF% with respect to ethnicity may be due to the diversity of the Brazilian population which includes a mixture of Amerindian, European, and

9 8 African genetic backgrounds. Moreover, this ethnic diversity can lead to less accurate determinations of ethnicity based on skin color. 20 To: In addition, the lack of a significant correlation between hip circumference and ADP-determined BF% with respect to race may be due to the ethnic background diversity of the Brazilian population which includes a mixture of Amerindian, European, and African genetic backgrounds and implies in a less accurate determination of ethnicity by race based on skin color. 22 Reference for this answer: 1. Notas técnicas: Histórico da investigação sobre cor ou raça nas pesquisas domiciliares do IBGE (2008). Available in estatistica/populacao/características_raciais/notas_tecnicas.pdf. Last access at November Comments 6 and 7: The discussion is well balanced and mostly supported by the data. In the Bergman (2011) original study they show the biases of the method as it overestimates %BF in persons with low BF and underestimates it in persons with high %BF. In light of the narrow target population of the current study (severely obese), it is unsurprising that the authors found that the BAI underestimated %BF similarly to the original study. Additionally, the correlation in this study between BF% from BAI and BF% from ADP which was vs 0.85 in the Bergman et al original study is not surprising because the original paper showed that BAI is not as good at predicting %BF among individuals at the extremes of the distribution. On page 10 the authors state a significant, but relatively low correlation between ADP-determined BF% and hip circumference was observed However, the pearson correlation coefficient (0.59) in this study is almost identical to the original Bergman study (0.602). I suggest revising this statement. The true difference is that height in this study was not associated with %BF from ADP whereas in the original study it was highly inversely correlated. Their conclusion that BAI may not be adequate for estimating BF in severely obese Brazilian patients follows from their results and

10 9 this is important for future studies to know that in general BAI can give a good approximation of body fat percent when it s applied to a normal ranges of body fat, but that in the extremes of the distribution this method has problems. I would like to have more discussion on why height in this study was not associated with BAI. Answer: Thank you for this relevant comment regarding the interpretation of our data. As properly referred by the Reviewer 1, the Person correlation for hip circumference (HC) and DXA-estimated BF% at developing phase of BAI by Bergman et al. was similarly low to that found in our study (0.59 vs 0.66). The Person correlation for hip circumference (HC) and DXA-estimated BF% improved to 0.85 only after the proper adjustments of the BAI formula. In fact, the main difference of our findings and those obtained from Bergman et al. was a lack of a significant inverse correlation between ADP-estimated BF% and height, which may be due our sample profile. BAI was developed and validated on samples highly heterogenic but with a very similar mean of BMI values (approximately 30, ranging from normal to obese), while our sample was composed only by severe obese subjects with BMI values > 30 (mean 47.4). In those subjects with BMI near to normal values it seems consistent that height can be relevant for BF% determination. We believe that taller subjects have a higher area for adiposity distribution than those shorter, which may be responsive for the significant inverse correlation between height and DXA observed by Bergman et al. However, the excessive fat amount and its general main concentration in abdominal area in severe obese patients may neutralize this relation between height and BF%. Consequently, other variables (such as abdominal circumference) seem more relevant than height to estimate BF% in this population. As suggested by the reviewer, we realize the need in change (page 11, lines here highlighted in red) and adding (page 11, lines ) some information to better address this issue in our discussion section, as follows: Here, similarly to Bergman et al. study a significant, but relatively low correlation between BF% determined by the reference method and hip circumference was observed. In addition, an inverse correlation between BF% determined by the reference method and height was also presently found but it

11 10 was not significant, differently of Bergman et al. findings. The lack of a significant inverse correlation between ADP-estimated BF% and height observed by us may be because we included only severe obese patients, while Bergman et al. included subjects with BMI ranging from normal to obese values during the BAI development and validation. Taller subjects are likely to have a higher area for adiposity distribution than those shorter, but in severe obese patients the excessive fat amount and its general main concentration in abdominal area may neutralize this effect. Consequently, other variables (such as abdominal circumference) seem more relevant than height to estimate BF% in this population. Comments 8: There are only minor grammatical revisions necessary. Minor Essential Revisions a) Page 4, line 74 adult should be changed to adults. b) Page 4-5, lines contain a run on sentence. c) Page 10, line 224 remove comma between low and correlation Answer: Thank you for these observations. All the suggested corrections were adequately performed in our manuscript. REVIEWER 2 Comment 1: Introduction - ADP and its benefits are discussed but almost all of the studies evaluating BAI compare it to DXA. A paragraph explaining how ADP and DXA are similar and/or different in terms of estimating BF% in your population of interest is necessary. Answer: Thank you for this relevant suggestion. The choice of obese population implied in using air displacement plethysmography (ADP) as the gold standard method for estimating body fat in our study because the dual energy xray absorptiometry (DXA) used for this purpose in Bergman's study can analyze only patients with limited weight and size. ADP has been validated in a

12 11 number of studies for normal weight populations 1,2, and mostly in obese populations 3,4 by being suitable and producing realistic body composition data for patients with BMI over 40 kg/m² 5,6. In an obese population similar to that we presently studied, Hames et al (2014) compared DXA and ADP and found strong agreement on fat free mass (FFM) estimation (r = 0.95, P < 0.001) 3. In addition, ADP was able to estimate a higher FFM than that estimated by DXA (P < 0.001), suggesting a higher sensibility of ADP to estimate FFM in obese patients with high BMI 3. Furthermore, there was a strong agreement between the estimations of fat content provided by these both methods regardless of the expression unit (kg and %) 3. Based on the Reviewer 2 suggestion we realize the consistent need to add some information showing that our study results are translatable between ADP and DXA. This information was added at discussion section (page 12, lines here highlighted in red), as follows: A limitation of the present study was the relatively small sample of Brazilian obese patients with a BMI of 30 kg/m ² (range: kg/m²). In addition, the inclusion of severe obese patients in our sample implied in using ADP, and not DXA, as the gold standard method for BF% estimation. While ADP has been shown to be suitable to adequately access body composition in patients with BMI over 40 kg/m², the use of DXA usually is applied only in patients with limited weight and size. 26,27 In an obese population similar to our study, Hames et al (2014) found a strong agreement between DXA and ADP performance in estimating fat content regardless of its expression unit (kg and %). 28 References for this answer: 1. Fields DA, Goran MI, McCrory MA (2002) Body-composition assessment via air-displacement plethysmography in adults and children: a review. Am J Clin Nutr 75: McCrory MA, Gomez TD, Bernauer EM, Molé PA (1995) Evaluation of a new air displacement plethysmograph for measuring human body composition. Med Sci Sports Exerc 27:

13 12 3. Hames KC, Anthony SJ, Thornton JC, Gallagher D, Goodpaster BH (2014) Body composition analysis by air displacement plethysmography in normal weight to extremely obese adults. Obesity (Silver Spring) 22: Bedogni G, Agosti F, De Col A, Marazzi N, Tagliaferri A, et al. (2013) Comparison of dual-energy X-ray absorptiometry, air displacement plethysmography and bioelectrical impedance analysis for the assessment of body composition in morbidly obese women. Eur J Clin Nutr 67: Ginde SR, Geliebter A, Rubiano F, Silva AM, Wang J, et al. (2005) Air displacement plethysmography: validation in overweight and obese subjects. Obes Res 13: Petroni ML, Bertoli S, Maggioni M, Morini P, Battezzati A, et al. (2003) Feasibility of air plethysmography (BOD POD) in morbid obesity: a pilot study. Acta Diabetol 40 Suppl 1: S Comment 2: A sample size calculation based on a study that looked at BAI in overweight/obese women and determined that an adequate sample size would be 62. While 72 people were tested in this study, only 53 were female. Please provide a paragraph discussing how BAI is not impacted by sex as the importance (or lack of importance of sex) is not adequately discussed in the text. Answer: Thank you for this pertinent suggestion that enables us to discuss our impression regarding the universal use of BAI in female and male populations. In fact our sample was composed by both genders included at non-equivalent proportion, by considering that BAI would not need further adjustments for gender to improve its performance. Although male and female are largely different in terms of body composition, BAI was not impacted by gender during its development and validation. In their article Bergman et al. did not properly discuss this issue, but we believe that the variable height in BAI formula could be responsible to neutralize such differences between genders. Because men are on average taller than women, they have a larger body area than women where fat can be distributed. Therefore, by considering height for calculation, BAI may neutralize this difference between genders. Thanks to the reviewer suggestion, we realize the need in adding some information to better address this issue in our discussion section, that was

14 13 promptly attended (pages 11-12, lines main information is here highlighted in red) as follows: Here, similarly to Bergman et al. study a significant, but relatively low correlation between BF% determined by the reference method and hip circumference was observed. In addition, an inverse correlation between BF% determined by the reference method and height was also presently found but it was not significant, differently of Bergman et al. findings. The lack of a significant inverse correlation between ADP-estimated BF% and height observed by us may be because we included only severe obese patients, while Bergman et al. included subjects with BMI ranging from normal to obese values during the BAI development and validation. Taller subjects are likely to have a higher area for adiposity distribution than those shorter, but in severe obese patients the excessive fat amount and its general main concentration in abdominal area may neutralize this effect. Consequently, other variables (such as abdominal circumference) seem more relevant than height to estimate BF% in this population. In addition, it is possible that height could be responsible for the neutral effect of gender on BAI performance, which accounted to a lack of concern to include a similar number of men and women in our study. Men are generally taller, with a consequent higher area for adiposity distribution than women. Therefore, by considering height for calculation, BAI may neutralize this difference between genders. The lack of an effect of height on BF% estimation observed by us suggests that a new formula adjusted by gender may be required for BF% estimation in severely obese subjects. Comment 3: How was abdominal circumference measured? Answer: Thank you for this important question that allowed us to realize that this information was missing in the description of our methodology. This information was promptly added at Methods section (pages 6-7, lines ), as follows: Waist circumference was measured at the trunk midway between the lower costal margin (bottom of the lowest rib) and the iliac crest (top of the pelvic bone) with the subject standing with his/her feet cm apart. The measurement was taken by fitting the tape snugly, without compressing the

15 14 underlying soft tissue. Circumference was measured to the nearest 0.5 cm; at the end of a normal expiration. 16 Comment 4: Why wasn't BMI compared to ADP? Other studies have done this, as indicated in the text. This additional analysis would be an important contribution to this literature and would serve to confirm or contrast the findings of others working in this area. Answer: Thank you for this coherent suggestion. We did not compare BMI to ADP because we thought this could divert from our study focus, which is to evaluate the BAI performance in estimating BF% in severely obese subjects. However, to attend the Reviewer 2 suggestion, we have included this analysis in the new version of our manuscript. In opposition to Geliebter et al., whose studied a severely obese population with a BMI mean similar to our population, we did not find a better performance of BMI to estimate BF%, in relation to BAI. Probably, this distinct finding may be due to the different nationality of our sample. At this point, we would like to highlight that even BMI had a better performance than BAI to estimate BF% in the Geliebter et al. study, the relation between BMI and reference methods was weak, showing that BMI also is not a good method for BF% estimation in severe obese patients. In our point of view, it is important to proper display this information in our manuscript to avoid a wrong interpretation of our findings (BAI was better than BMI to estimate BF% in our population, but both methods have a poor performance for this prupose). Therefore we added some new sentences throughout the new version of our manuscript, as follows: In Methods (page 8, line here highlighted in red): The 95% limits of agreement between the BAI and ADP and BMI and ADP were determined using the Bland-Altman method. 20 In Results (page 8, lines here highlighted in red) The 95% limits of individual agreement between BAI and ADP were 5.77% to 16.04% (range: 21.8%), as shown in Figure 1. These limits of individual agreement were higher than those found between BMI and ADP ( 7.34% to 16.72%, range: 24.1%; r p =0.39), as shown in Figure 2. However, the BAI

16 15 exhibited an average positive bias of 5.13% compared to the reference method. Figure 1 - Bland-Altman plot showing limits of agreement between BF% by BAI vs. ADP. Bold continuous line indicates observed average agreement. Continuous line indicates line of perfect average agreement. Dashed lines indicate 95% limits of agreement. Lin s concordance correlation coefficient (r c ) is shown.

17 16 Figure 2 - Bland-Altman plot showing limits of agreement between BF% by BMI vs. ADP. Bold continuous line indicates observed average agreement. Continuous line indicates line of perfect average agreement. Dashed lines indicate 95% limits of agreement. Lin s concordance correlation coefficient (r c ) is shown. In Discussion (page 10, lines here highlighted in red): In a study by Geliebter et al. 13, the limits of agreement between BAI and the reference method (either DXA or ADP) were wider than those between BMI and the reference method. Furthermore, the range of the confidence limits between BAI and ADP in their study ( 18.47% to 8.49%, range: 27.0%) are similar to the confidence limits in the present study. The authors concluded that although BAI appears to be a reasonable index of BF%, it is not an appropriate substitute for BMI in evaluating severely obese women. Although in our severely obese Brazilian patients the BAI performance was discreetly better than the BMI performance to estimate BF%, our findings support the conclusion of Geliebter et al. by not displaying similar results of this measuring between BAI and ADP. Comment 5: Significance is set at P < 0.05, but the only significantly correlated variables listed in Table 1 are BMI, hip circumference, and BAI. Other variables have p-values of less than what am I missing? Did you correct for multiple comparisons? If so, what is the corrected p-value? Answer: Thank you for this important question that enabled us to realize that the description of our data regarding variables significantly correlated with ADPestimated BF% was subject to misinterpretation. In fact, we intended to say that despite the p value was significant for variables other than BMI, hip circumference and BAI, the correlation value for these variables was very low (R < 0.50). However, we recognize that, the way we wrote, this information was transmitted distortedly. Therefore, we are grateful for this observation and inform that the referred text has been changed (page 8, lines ), as follows: From: BMI, hip circumference, and BAI showed significant correlations (r p = 0.529, 0.59, and 0.664, respectively) with the BF% estimates by ADP, although the values of these coefficients were relatively low. No significant correlations

18 17 between the ADP-determined BF% estimates and age, body weight, height, abdominal circumference, or waist/hip ratio were observed. To: Except for age and height, all evaluated variables showed significant correlations with the BF% estimates by ADP, although these coefficients were low with reasonable values (> 0.5) only for BMI, hip circumference, and BAI (r p = 0.529, 0.59, and 0.664, respectively). Comment 6: How was race/ethnicity determined? Was this by self-report or - as I interpret from the text - by visual determination of the researcher? It is not clear in the manuscript. If it is the latter, please remove this analysis as this is an unreliable method of assessment. Answer: Thank you for this pertinent question. First, we would like to inform that in this reviewed version the ethnicity was substituted by race over the manuscript, because the different subgroups that compound Brazilian population are nationally classified according to color or race, while ethnicity includes cultural variables, as religion adopted. The race/ethnicity was determined by self-report 1. In order to adequately address the ethnicity classification in our study, we added a new sentence at Methods section (page 6, lines ), as follows: According to the Brazilian Institute of Geography and Statistics (IBGE), sample ethnicity was self-reported as race by considering skin color for classification into the following categories: white, black/brown, yellow and indigenous. 14. Reference for this answer: 1. Baker DW, Cameron KA, Feinglass J, Thompson JA, Georgas P, Foster S, et al. A system for rapidly and accurately collecting patients race and ethnicity. Am J Public Health Mar;96(3): Comment 7: Why is the only sub-group analysis by race/ethnicity that between hip circumference and ADP-determined BF%? If others were not significant, please say so.

19 18 Answer: Thank you for this pertinent question. In fact, no variable was correlated with ADP-estimated FM%, as showed above: We had presented this observation only to hip circumference, because this was the only variable that was significant in relation to ADP in our study and also significant in relation to DXA during the development of BAI by Bergman et al. However, we fully agree with the reviewer 2 that, in our study, this issue must be better explained and extended to other variables. For this prupose, we replaced the sentence addressing these results in the methods section (pages 8-9, lines here highlighted in red), as follows: Except for age and height, all evaluated variables showed significant correlations with the BF% estimates by ADP, although these coefficients were low with reasonable values (> 0.5) only for BMI, hip circumference, and BAI (r p = 0.529, 0.59, and 0.664, respectively). In addition, no significant correlations between the ADP-determined BF% estimates and these variables were observed with respect to race, even for hip circumference, which was presently correlated to ADP and also correlated to DXA during the BAI development (Figure 3).

20 19 Figure 3. Scatter plot of hip circumference vs. BF% by ADP among white and black/brown obese Brazilian patients. Solid line represents the Local Polynomial Regression fitting that is bounded by the 95% confidence band. Comment 8: Line 205: why do we need this study if it has more or less already been done? Distinguish your study from this one. Answer: Thank you for the opportunity to discuss the relevance of our study question. The adequate identification of fat amount is relevant to estimate the risk to develop different complications potentially associated with overweight, such as metabolic syndrome. In this issue, the body adiposity index (BAI) was found to be sensitive in estimating body fat (BF) %, compared to the reference dual-energy X-ray absorptiometry (DXA). Since BAI development by Bergman et al. its application has been addressed in several studies aiming to test whether this simple, practical and inexpensive tool can have a good performance to estimate BF% in other population than that where BAI was developed and validated. This effort is because most public hospitals from countries with limited financial resources for health spending are eager for simple and affordable methods for assessing body fat of their patients. Particularly, severe obese patients make up a population at risk of complications associated with excess body fat and whose BF% estimation is constantly limited due to the unavailability of methods which happen to contain

21 20 the size and weight of patients with this profile in clinical practice. Therefore, we tested whether BAI could be adopted as a tool with a good performance to estimate BF% in severely Brazilian obese. As the Reviewer 2 corrected point out, another two studies has addressed this question in other severely non- Brazilian obese 1 and in Brazilians obeses 2. The first one mainly differs from our study by the ethnic background of the population assessed, and the performance of BAI has not been consistent in populations with characteristics different from those used for its development and validation, which may include ethnicity/race. The second one did not focuses on obese but overweight patients (mean BMI = 26.9) and our study focuses only on obese patients (BMI 30 kg/m 2, mean BMI = 47.45), by considering this population the most susceptible for complications associated with excessive body fat and lacking in simple methods that enable the analysis of their BF%. Therefore, our study was the first to assess one severely obese patient with a Brazilian background. Based on Reviewer 2 comments we realize the consistent need to highlight the characteristics of our population as the main differential our study. For this purpose we included some information at the introduction (page 5, lines here, highlighted in red), as follows: ( ) both of the databases that were used for the development and validation of the BAI included groups of subjects with varying degrees of BF% (BMIs average: ~ 30 kg/m², ranging from normal to obese) 8. It is worth noting that more severely obese patients are at higher complications risk associated with excess body fat. Their large size and weight limits de performance of simple (such as skin fold) and more sophisticated (such as DXA) methods for the clinical BF% estimation. Therefore, we tested whether BAI could have a good performance to estimate BF% in severely obese Brazilian patients. References for this answer: 1. Geliebter A, Atalayer D, Flancbaum L, Gibson CD. Comparison of body adiposity index (BAI) and BMI with estimations of % body fat in clinically severe obese women. Obesity (Silver Spring) 2013; 21:

22 21 2. Cerqueira M, Amorim P, Magalhães F, Castro E, Franco F, Franceschini S et al. Validity of body adiposity index in predicting body fat in a sample of Brazilian women. Obesity (Silver Spring) 2013; 21: E696 E699. Comment 9: Abstract: BAI estimates percent body fat. I would change line 28 to reflect that. Also, throughout the manuscript, I urge you to be very specific and use BF% when you mean that rather than just BF. When I see BF, I start thinking about kg rather than %. Another example is line 45. Answer: Thank you for this keen observation that was promptly attended throughout our manuscript text. Comment 10: Discussion: lines I don't think it's appropriate to say that BAI was found to be LESS ACCURATE than your standard, since it, too, is an estimation, and perhaps ADP would vary from DXA in your study population. Perhaps you could say that the values are not in agreement with each other or something like this? Answer: Thank you for this keen observation that was promptly attended (page 9, lines here highlighted in red), as follows: For estimating BF% in severely obese Brazilian patients, BAI was found to provide differing values from those estimated by ADP, especially for patients with lower BF% values.