Table 2.
Associations between consumption of ultra-processed food and incident type 2 diabetes
| Quartiles of ultra-processed food consumption | |||||||
|---|---|---|---|---|---|---|---|
| First | Second | Third | Fourth | P-trend | Continuousa | P value | |
| Cases/population | 255/17,604 | 247/17,606 | 272/17,605 | 354/17,606 | 1128/70,421 | ||
| Incidence, % | 1.4 | 1.4 | 1.5 | 2.0 | 1.6 | ||
| Model 1b | 1 | 1.11 (0.93, 1.33) | 1.39 (1.17, 1.65) | 2.17 (1.83, 2.58) | < 0.001 | 1.33 (1.26, 1.41) | < 0.001 |
| Model 2c | 1 | 1.08 (0.90, 1.30) | 1.30 (1.07, 1.57) | 1.87 (1.52, 2.30) | < 0.001 | 1.27 (1.18, 1.36) | < 0.001 |
| Model 3d | 1 | 1.08 (0.90, 1.30) | 1.28 (1.06, 1.55) | 1.80 (1.47, 2.20) | < 0.001 | 1.25 (1.16, 1.34) | < 0.001 |
| Model 4e | 1 | 1.04 (0.87, 1.26) | 1.20 (0.99, 1.45) | 1.56 (1.27, 1.92) | < 0.001 | 1.17 (1.09, 1.26) | < 0.001 |
aContinuous model indicates OR (95% CI) for an absolute increment of 10% consumption of ultra-processed food in the total diet
bModel 1: OR (95% CI) derived from multivariate logistic regression models adjusted for age and sex, n = 70,421
cModel 2: OR (95% CI) derived from multivariate logistic regression models adjusted for Model 1 covariates plus Lifelines diet score, total energy intake, and alcohol intake, n = 70,421
dModel 3: OR (95% CI) derived from multivariate logistic regression models adjusted for Model 2 covariates plus smoking status, educational level, non-occupational moderate-to-vigorous physical activity level, and TV watching time, n = 70,418
eModel 4: OR (95% CI) derived from multivariate logistic regression models adjusted for Model 3 covariates plus BMI, n = 70,403