TABLE 4.
Illustrative example based on 1binary logistic regression examining the added sugar-disease relation based on self-reported added sugar intake compared with added sugar intake estimated by the δ¹³C isotope-based equation (n = 106, Southwest Virginian adults)
| Quartiles of added sugar intake | |||||
|---|---|---|---|---|---|
| Q1 (ref) | Q2 | Q3 | Q4 | P trend | |
| Self-reported | |||||
| n (case) | 29 (8) | 28 (10) | 31 (15) | 28 (16) | — |
| Median (range) added sugar intake, g/d | 36.5 (13.5–45.6) | 56.0 (45.9–67.9) | 79.3 (68.3–98.5) | 156.5 (104.8–256.4) | — |
| OR (95% CI) | 1.00 | 1.46 (0.47, 4.48) | 2.46 (0.84, 7.22) | 3.50 (1.16, 10.58) | 0.016 |
| δ¹³C isotope-based | |||||
| n (case) | 29 (12) | 29 (13) | 29 (11) | 29 (13) | — |
| Median (range) added sugar intake, g/d | 55.7 (44.8–60.8) | 64.3 (61.6–68.8) | 73.7 (69.2–77.3) | 86.5 (77.3–108.3) | — |
| OR (95% CI) | 1.00 | 1.15 (0.41, 3.26) | 0.87 (0.30, 2.48) | 1.15 (0.41, 3.26) | 0.933 |
1
Calculated from a mock dataset generated based on the reverse-engineered data from Hedrick et al. (49).