Table 5.
Estimates for effect modification by age (dichotomized at the median value of 40 years) for the percentage difference in fractional exhaled nitric oxide in relation to measures of exposure to household air pollution (per 25% increase in 24-h average measured pollution, or by stove type) among traditional and Justa stove users, rural Honduras 1.
| n | Adjusted Percent Difference 1 | 95% CI | p-Value for Interaction | |
|---|---|---|---|---|
| 24-h average kitchen PM2.5 (µg/m3) 2 | 0.7 | |||
| Age < 40 | 65 | <0.1 | (−3.3, 3.4) | |
| Age ≥ 40 | 31 | 1.0 | (−3.0, 5.1) | |
| 24-h average personal PM2.5 (µg/m3) 2 | 0.2 | |||
| Age < 40 | 66 | −0.9 | (−5.7, 4.3) | |
| Age ≥ 40 | 31 | 5.2 | (−3.3, 14.3) | |
| 24-h average kitchen Black Carbon (µg/m3) 2 | 0.8 | |||
| Age < 40 | 65 | −0.3 | (−2.6, 2.1) | |
| Age ≥ 40 | 31 | 0.2 | (−2.7, 3.2) | |
| 24-h average personal Black Carbon (µg/m3) 2 | 0.7 | |||
| Age < 40 | 65 | −0.5 | (−3.1, 2.3) | |
| Age ≥ 40 | 31 | 0.5 | (−3.3, 4.7) | |
| Stove Type 3 (traditional compared to Justa [ref]) | ||||
| Traditional | ||||
| Age < 40 | 72 | −12.2 | (−31.9, 13.1) | 0.3 |
| Age ≥ 40 | 67 | 11.1 | (−22., 59.2) |
Cl: Confidence interval; PM2.5: fine particulate matter. 1 All models adjusted for age, height, waist-to-hip ratio, body mass index, dietary-diversity score, years of education (<6 or ≥6 years), and number of assets (<2 or ≥2) (Assets include cars, bikes, motorbikes, televisions, radios, refrigerators, sewing machines, electricity). 2 Exhaled nitric oxide and measured pollution are both log transformed. Beta coefficients were entered into the formula ((1.25^β) − 1) and multiplied by 100. We can interpret the estimate of the continuous pollution exposures as a percent increase in exhaled nitric oxide for each 25% increase in exposure. Example: There is a 0.2% higher FeNO level with a 25% higher kitchen PM2.5 concentration among women less than 40 years old. 3 Exhaled nitric oxide was log-transformed. Categorical variable beta coefficients were entered into the formula (e^β − 1)*100). The estimates for the categorical measures of exposure can be interpreted as the percent difference in FeNO when comparing a specific stove type to the reference (traditional stove).