Higher mortality in “normal weight” adults: should the BMI range for normal weight be redefined?
CCBYNCOpen access
Rapid response to:
Research
BMI and all cause mortality: systematic review and non-linear dose-response meta-analysis of 230 cohort studies with 3.74 million deaths among 30.3 million participants
Higher mortality in “normal weight” adults: should the BMI range for normal weight be redefined?
In this systematic review and meta-analysis, Aune and colleagues concluded that their analysis “provides strong evidence that overweight and obesity increases the risk of all-cause mortality” [1]. The conclusion that overweight associated with increased mortality contradicted a previous meta-analysis [2], in which Flegal and colleagues reported that overweight (BMI: 25-29.9) was associated with lower mortality relative to normal weight (BMI:18.5-24.9) [2]. Aune and colleagues explained the seemingly contradictory conclusions as the previous meta-analysis being prone to confounding by smoking and existing illness [1]. However, it appears that their data could not support this explanation.
To illustrate the relation between BMI and mortality from their published data, we generated Figure 1 for never smokers and Figure 2 for healthy never smokers [1]. We included a narrower BMI range in those figures to focus on the comparison between normal weight (BMI: 18.5-24.9) and overweight (BMI: 25-29.9). Among never smokers, the relative risk (RR) was 1.01 for BMI of 25 and 1.07 for BMI of 27.5, with BMI of 23 as the reference. The RR for BMI of 27.5 was equivalent to that for BMI of 20.7, indicating that normal weight adults with BMI of 18.5-20.7 had a higher risk than overweight adults with BMI of 25-27.5. Therefore, the lower risk in the overweight group could not be explained by the confounding effect of smoking. Even in healthy never smokers, those with BMI of 25 still had a lower risk (1.03) than those in the lower normal weight range (BMI: 18.5-20) (Figure 2), suggesting that the lower risk associated with overweight could not be fully explained by the confounding effect of smoking and pre-existing illness.
There was a key methodological difference between two meta-analyses. Flegal and colleagues used the WHO BMI categories with normal weight (18.5-24.9) as the reference while Aune and colleagues used BMI of 23, which was about the nadir of the BMI-mortality curve, as the reference. It would be interesting to know if both meta-analyses would have reached the same conclusion had they used the same reference.
Aune and colleagues further suggested that their results support the WHO recommendation of the normal BMI of 18.5-24.9. On the contrary, as shown in Figures 1 and 2, BMI of 18.5-20 marked a substantially increase in mortality, and the risk decreased steeply with increasing BMI. Therefore, instead of debating about whether overweight is associated with higher or lower mortality relative to normal weight, we should shift our focus to evaluating the appropriateness of the current WHO BMI normal weight range. As large amount of data have been accumulated, researchers are in a good position to reassess and redefine normal BMI values. Based on large Korean data, Yi and colleagues [3] found that optimal BMI ranges associated with minimal risk of death varied with age and sex, and were generally higher than the current normal weight (BMI of 18.5-24.9), suggesting that the WHO normal BMI range is questionable for Korean adults. In another study of three Danish cohorts, Afzal and colleagues found that BMI associated with the lowest mortality increased by 3.3 from 1976-1978 (BMI: 23.7) to 2003-2013 (BMI: 27.0) [4]. Therefore, it is possible that the current normal BMI range based on evidence from several decades ago may be no longer applicable to contemporary adults.
Although there were over 9 million never smokers included in this meta-analysis, the nonlinear curves were established based on a very limited number of data points available in the original articles. Since a large amount of relevant data have already been collected, sharing de-identified original data with individual observations is critical for updating BMI values for normal weight.
References
1. Aune D, Sen A, Prasad M, et al. BMI and all cause mortality: systematic review and non-linear dose-response meta-analysis of 230 cohort studies with 3.74 million deaths among 30.3 million participants. BMJ 2016;353:i2156 doi: 10.1136/bmj.i2156.
2. Flegal KM, Kit BK, Orpana H, Graubard BI. Association of all-cause mortality with overweight and obesity using standard body mass index categories: a systematic review and meta-analysis. JAMA 2013;309(1):71-82 doi: 10.1001/jama.2012.113905.
3. Yi SW, Ohrr H, Shin SA, Yi JJ. Sex-age-specific association of body mass index with all-cause mortality among 12.8 million Korean adults: a prospective cohort study. Int J Epidemiol 2015 doi: 10.1093/ije/dyv138.
4. Afzal S, Tybjaerg-Hansen A, Jensen GB, Nordestgaard BG. Change in Body Mass Index Associated With Lowest Mortality in Denmark, 1976-2013. JAMA 2016;315(18):1989-96 doi: 10.1001/jama.2016.4666.
Competing interests:
No competing interests
18 May 2016
Zhiqiang Wang
Associate Professor
Bin Dong, Yang Peng, Tania Pan, Meina Liu
School of Medicine, University of Queensland
Room 817 Health Sciences Building, RBWH, Herston QLD, Australia
Rapid Response:
Higher mortality in “normal weight” adults: should the BMI range for normal weight be redefined?
In this systematic review and meta-analysis, Aune and colleagues concluded that their analysis “provides strong evidence that overweight and obesity increases the risk of all-cause mortality” [1]. The conclusion that overweight associated with increased mortality contradicted a previous meta-analysis [2], in which Flegal and colleagues reported that overweight (BMI: 25-29.9) was associated with lower mortality relative to normal weight (BMI:18.5-24.9) [2]. Aune and colleagues explained the seemingly contradictory conclusions as the previous meta-analysis being prone to confounding by smoking and existing illness [1]. However, it appears that their data could not support this explanation.
To illustrate the relation between BMI and mortality from their published data, we generated Figure 1 for never smokers and Figure 2 for healthy never smokers [1]. We included a narrower BMI range in those figures to focus on the comparison between normal weight (BMI: 18.5-24.9) and overweight (BMI: 25-29.9). Among never smokers, the relative risk (RR) was 1.01 for BMI of 25 and 1.07 for BMI of 27.5, with BMI of 23 as the reference. The RR for BMI of 27.5 was equivalent to that for BMI of 20.7, indicating that normal weight adults with BMI of 18.5-20.7 had a higher risk than overweight adults with BMI of 25-27.5. Therefore, the lower risk in the overweight group could not be explained by the confounding effect of smoking. Even in healthy never smokers, those with BMI of 25 still had a lower risk (1.03) than those in the lower normal weight range (BMI: 18.5-20) (Figure 2), suggesting that the lower risk associated with overweight could not be fully explained by the confounding effect of smoking and pre-existing illness.
There was a key methodological difference between two meta-analyses. Flegal and colleagues used the WHO BMI categories with normal weight (18.5-24.9) as the reference while Aune and colleagues used BMI of 23, which was about the nadir of the BMI-mortality curve, as the reference. It would be interesting to know if both meta-analyses would have reached the same conclusion had they used the same reference.
Aune and colleagues further suggested that their results support the WHO recommendation of the normal BMI of 18.5-24.9. On the contrary, as shown in Figures 1 and 2, BMI of 18.5-20 marked a substantially increase in mortality, and the risk decreased steeply with increasing BMI. Therefore, instead of debating about whether overweight is associated with higher or lower mortality relative to normal weight, we should shift our focus to evaluating the appropriateness of the current WHO BMI normal weight range. As large amount of data have been accumulated, researchers are in a good position to reassess and redefine normal BMI values. Based on large Korean data, Yi and colleagues [3] found that optimal BMI ranges associated with minimal risk of death varied with age and sex, and were generally higher than the current normal weight (BMI of 18.5-24.9), suggesting that the WHO normal BMI range is questionable for Korean adults. In another study of three Danish cohorts, Afzal and colleagues found that BMI associated with the lowest mortality increased by 3.3 from 1976-1978 (BMI: 23.7) to 2003-2013 (BMI: 27.0) [4]. Therefore, it is possible that the current normal BMI range based on evidence from several decades ago may be no longer applicable to contemporary adults.
Although there were over 9 million never smokers included in this meta-analysis, the nonlinear curves were established based on a very limited number of data points available in the original articles. Since a large amount of relevant data have already been collected, sharing de-identified original data with individual observations is critical for updating BMI values for normal weight.
References
1. Aune D, Sen A, Prasad M, et al. BMI and all cause mortality: systematic review and non-linear dose-response meta-analysis of 230 cohort studies with 3.74 million deaths among 30.3 million participants. BMJ 2016;353:i2156 doi: 10.1136/bmj.i2156.
2. Flegal KM, Kit BK, Orpana H, Graubard BI. Association of all-cause mortality with overweight and obesity using standard body mass index categories: a systematic review and meta-analysis. JAMA 2013;309(1):71-82 doi: 10.1001/jama.2012.113905.
3. Yi SW, Ohrr H, Shin SA, Yi JJ. Sex-age-specific association of body mass index with all-cause mortality among 12.8 million Korean adults: a prospective cohort study. Int J Epidemiol 2015 doi: 10.1093/ije/dyv138.
4. Afzal S, Tybjaerg-Hansen A, Jensen GB, Nordestgaard BG. Change in Body Mass Index Associated With Lowest Mortality in Denmark, 1976-2013. JAMA 2016;315(18):1989-96 doi: 10.1001/jama.2016.4666.
Competing interests: No competing interests