High COPD prevalence at high altitude: does household air pollution play a role?

Abstract

Studies comparing chronic obstructive pulmonary disease (COPD) prevalence across altitudes report conflicting results. However, household air pollution (HAP), a major COPD risk factor, was mostly not accounted for in previous analyses and never objectively measured. We aimed to compare the prevalence of COPD and its risk factors between low-resource highlands and lowlands, with a particular focus on objectively measured HAP.We conducted a population-based, observational study in a highland (∼2050 m above sea level) and a lowland (∼750 m above sea level) setting in rural Kyrgyzstan. We performed spirometry in randomly selected households, measured indoor particulate matter with an aerodynamic diameter <2.5 µm (PM_2.5) and administered a questionnaire on other COPD risk factors. Descriptive statistics and multivariable logistic regressions were used for analyses.We included 392 participants: 199 highlanders and 193 lowlanders. COPD was more prevalent among highlanders (36.7% versus 10.4%; p<0.001). Their average PM_2.5 exposure was also higher (290.0 versus 72.0 µg·m^-3; p<0.001). In addition to high PM_2.5 exposure (OR 3.174, 95% CI 1.061-9.493), the altitude setting (OR 3.406, 95% CI 1.483-7.825), pack-years of smoking (OR 1.037, 95% CI 1.005-1.070) and age (OR 1.058, 95% CI 1.037-1.079) also contributed to a higher COPD prevalence among highlanders.COPD prevalence and HAP were highest in the highlands, and were independently associated. Preventive interventions seem warranted in these low-resource, highland settings. With this study being one of the first spirometry-based prevalence studies in Central Asia, generalisability needs to be assessed.

FIGURE 1

Flowchart of participants.

FIGURE 2

Levels of particulate matter with an aerodynamic diameter <2.5 µm (PM_2.5) in the lowlands versus the highlands (measured as time-weighted average concentration per household). The median and interquartile range are indicated. The dashed line is the maximum average 24-h PM_2.5 concentration of 25 µg·m⁻³ as stated in the World Health Organization indoor air quality guideline [33]. Missing values were excluded from the study (figure 1). *: difference is statistically significant (p<0.001).

FIGURE 3

Chronic obstructive pulmonary disease (COPD) prevalence in the lowlands versus the highlands. *: difference is statistically significant (p<0.001).

FIGURE 4

Chronic obstructive pulmonary disease (COPD) prevalence in the lowlands and highlands by sex and household air pollution (HAP) exposure. HAP exposure was categorised into tertiles, with time-weighted average concentrations of particulate matter with an aerodynamic diameter <2.5 μm of ≤72, >72–293 and >293 µg·m⁻³.

FIGURE 5

Potential predictors for chronic obstructive pulmonary disease (COPD): a) univariable analyses and b) multivariable analysis. HAP: household air pollution. COPD (n=93) versus no COPD (n=299). Generalised estimating equation analyses adjusted for a clustering effect within households. Age per year increase. Higher education means the highest level of completed education is above secondary education. HAP exposure was categorised into tertiles, with time-weighted average concentrations of particulate matter with an aerodynamic diameter <2.5 μm of ≤72, >72–293 and >293 μg·m⁻³. The lowest category is the reference category. Data are presented as OR (95% CI). Note that with a COPD prevalence >10%, OR no longer approximates relative risk. *: difference is statistically significant (95% CI of the OR excluding 1; p<0.05).