Table 4.
Multivariable linear regression models of predictors of plasma ln(DDE) levels [percent change in DDT levels (95% CI)] among South African women 20–30 years of age, 2010–2011, by exposure group.a
| Predictor | Unsprayed (n = 175) | non-DDT IRS (n = 106) | DDT IRS (n = 100) | |||
|---|---|---|---|---|---|---|
| Percent change (95% CI)a | Adj. R2 | Percent change (95% CI)a | Adj. R2 | Percent change (95% CI)a | Adj. R2 | |
| BMI (kg/m2) | ||||||
| < 21.6 | NS | NS | Referent | |||
| 21.6–24.7 | NS | NS | –49 (–68, –19) | |||
| 24.8–28.3 | NS | NS | –39 (–62, –1) | |||
| ≥ 28.4 | NS | NS | –39 (–63, –1) | 0.04 | ||
| Parity | ||||||
| Nulliparous | NS | Referent | Referent | |||
| 1 | NS | –60 (–76, –32) | –33 (–60, 13) | |||
| > 1 | NS | –51 (–72, –16) | 0.06 | –66 (–81, –41) | 0.12 | |
| Livestock ownership | ||||||
| No | NS | NS | Referent | |||
| Yes | NS | NS | 87 (31, 167) | 0.09 | ||
| Water source | ||||||
| Public tap | Referent | NS | NS | |||
| Piped to yard/home | –61 (–74, –40) | 0.09 | NS | NS | ||
| Butter consumption | ||||||
| < 1 time/month | Referent | NS | NS | |||
| ≥ 1 time/month and < 1 time/day | 85 (13, 202) | NS | NS | |||
| ≥ 1 time/day | 93 (18, 215) | 0.03 | NS | NS | ||
| Milk consumption | ||||||
| < 1 time/month | NS | Referent | NS | |||
| 1–4 times/month | NS | –16 (–47, 35) | NS | |||
| > 4 times/month | NS | –45 (–66, –11) | 0.03 | NS | ||
| Chicken consumption | ||||||
| ≤ 1 time/week | NS | Referent | NS | |||
| 2 times/week | NS | –16 (–52, 47) | NS | |||
| 3 times/week | NS | –49 (–70, –12) | NS | |||
| > 3 times/week | NS | –46 (–69, –6) | 0.05 | NS | ||
| Abbreviations: NA, not applicable; NS, not selected. All models are adjusted for total lipids. aCalculated using the following formula: [exp(β) – 1] × 10. | ||||||