Association between Personal Exposure to Household Air Pollution and Glycated Hemoglobin among Women in Rural Areas of Guatemala, India, Peru, and Rwanda: Household Air Pollution Intervention Network Trial

Abstract

Background: Household air pollution from biomass cookstoves is a major concern in low- and middle-income countries because it may be linked with increasing rates of metabolic disorders such as diabetes. We assessed cross-sectional associations between household air pollution concentrations and glycated hemoglobin (HbA1c) levels.

Methods: We analyzed data from 346 women 40 to $<80$ years of age who cooked with biomass fuel and were enrolled in the Household Air Pollution Intervention Network (HAPIN) Trial in Guatemala, India, Peru, and Rwanda. We explored associations of 24-h average personal exposure to fine particulate matter [PM $\le 2.5\mu m$ in aerodynamic diameter ( ${\mathrm{PM}}_{2.5}$ )], black carbon (BC), and carbon monoxide (CO) with HbA1c through individual pollutant linear models adjusted for potential confounders. We examined the effect modification of age, body mass index (BMI), and research site on the associations.

Results: We did not observe evidence of associations between HbA1c (percentage points) and 1-unit increases in log-transformed ${\mathrm{PM}}_{2.5}$ [ $-0.07$ ; 95% confidence interval (CI): $-0.18$ , 0.05], BC (0.01; 95% CI: $-0.15$ , 0.13), or CO (0.07; 95% CI: $-0.24$ , 0.10). Effect modification of the BC associations with HbA1c was observed for BMI and research site. An association in the hypothesized direction was observed among women with high BMI ( $\ge 25{\mathrm{kg}/m}^2$ ): 0.13 (95% CI: $-0.06$ , 0.31) compared with low BMI ( $<25{\mathrm{kg}/m}^2$ ): $-0.17$ (95% CI: $-0.38$ , 0.04; $p_{\text{interaction}}=0.04$ ). In the Guatemala research site, there was an association in the hypothesized direction with HbA1c and log-transformed BC (0.36; 95% CI: 0.03, 0.70) that was countered by an association in the opposite direction as that hypothesized for the India site ( $-0.21$ ; 95% CI: $-0.45$ , 0.02) and associations consistent with the null association in the Peru and Rwanda sites ( $p_{\text{interaction}}=0.05$ ). No other evidence of effect modification was observed.

Conclusions: Evidence suggests a need for further research to better understand household air pollution's influence on HbA1c, with particular attention on potential effect modifiers. https://doi.org/10.1289/JHP1053.

Figure 1.

Box-and-whisker plot of HbA1c levels (%) among women at baseline by HAPIN research site. The central line in each box represents the median, the boxes represent the interquartile range (IQR), and the whiskers extend to 1.5 times the IQR. Outliers beyond the whiskers are shown as individual points. The red dashed line represents the standard cutoff for diabetes diagnosis at 6.5%, and the gray shaded region is the prediabetes range of 5.7%–6.4%. Summary statistics are displayed above each box, including counts and mean HbA1c values. Numeric values can be found in Tables 1 and 3. Note: HAPIN, Household Air Pollution Intervention Network; HbA1c, glycated hemoglobin; IRC, international research center.