Change in the body mass index distribution for women: analysis of surveys from 37 low- and middle-income countries

Abstract

Background: There are well-documented global increases in mean body mass index (BMI) and prevalence of overweight (BMI ≥ 25.0 kg/m(2)) and obese (BMI ≥ 30.0 kg/m(2)). Previous analyses, however, have failed to report whether this weight gain is shared equally across the population. We examined the change in BMI across all segments of the BMI distribution in a wide range of countries, and assessed whether the BMI distribution is changing between cross-sectional surveys conducted at different time points.

Methods and findings: We used nationally representative surveys of women between 1991-2008, in 37 low- and middle-income countries from the Demographic Health Surveys ([DHS] n = 732,784). There were a total of 96 country-survey cycles, and the number of survey cycles per country varied between two (21/37) and five (1/37). Using multilevel regression models, between countries and within countries over survey cycles, the change in mean BMI was used to predict the standard deviation of BMI, the prevalence of underweight, overweight, and obese. Changes in median BMI were used to predict the 5th and 95th percentile of the BMI distribution. Quantile-quantile plots were used to examine the change in the BMI distribution between surveys conducted at different times within countries. At the population level, increasing mean BMI is related to increasing standard deviation of BMI, with the BMI at the 95th percentile rising at approximately 2.5 times the rate of the 5th percentile. Similarly, there is an approximately 60% excess increase in prevalence of overweight and 40% excess in obese, relative to the decline in prevalence of underweight. Quantile-quantile plots demonstrate a consistent pattern of unequal weight gain across percentiles of the BMI distribution as mean BMI increases, with increased weight gain at high percentiles of the BMI distribution and little change at low percentiles. Major limitations of these results are that repeated population surveys cannot examine weight gain within an individual over time, most of the countries only had data from two surveys and the study sample only contains women in low- and middle-income countries, potentially limiting generalizability of findings.

Conclusions: Mean changes in BMI, or in single parameters such as percent overweight, do not capture the divergence in the degree of weight gain occurring between BMI at low and high percentiles. Population weight gain is occurring disproportionately among groups with already high baseline BMI levels. Studies that characterize population change should examine patterns of change across the entire distribution and not just average trends or single parameters.

Figure 1. Change in specific parameters of the BMI distribution (5th and 95th percentile and SD) or categories of BMI (underweight, overweight, obese) versus change in mean or median BMI.

Change is calculated from baseline survey to most recent survey and each data point is represented by a two-letter country code. Multilevel models across all country-survey cycles are contained in Table 2.

Figure 2. QQ plots of BMI in 37 countries.

Country plots are ordered in terms of increasing mean weight between baseline and final survey, from smallest to largest. x-axis BMI at the baseline survey; y-axis BMI at final survey cycle. The diagonal line y = x is the line of equality between baseline survey and final survey. Δ BMI, change in mean BMI from first survey to last survey. Red text indicates decline in mean BMI; green text indicates increase in mean BMI.

Figure 3. Population distribution of BMI over time in Bangladesh, Bolivia, Ghana, Peru, and Egypt.

BMI histograms (panels on left). Vertical reference lines (dotted) represent established thresholds for BMI: 16.0 (chronic energy deficiency), 18.5 (underweight), 25.0 (overweight), and 30.0 (obese). Shaded regions represent proportion with BMI<16.0 or >25.0 in Bangladesh and proportion with BMI <18.5 or >30.0 in Bolivia, Ghana, Peru, and Egypt. QQ plots (panels on right). x-axis BMI at the baseline survey. y-axis BMI at successive survey cycles. The diagonal line y = x is the line of equality between baseline survey and subsequent survey. Vertical reference lines (dotted) represent the 10th, 50th, and 90th percentile, with the value of BMI at the baseline survey at the top of each line.

Figure 4. QQ plots of different models of population change in BMI over time.

The observed pattern of change (from Figure 2) in green, uniform change model in red, and a proposed model of ideal change in blue. The corresponding table indicates impact of each model of change on (1) mean BMI, (2) dispersion of BMI, and (3) risk level of low and high BMI population segments.