Table 4.
Sensitivity analysis of CVD mortality risk associated with late non-fasting plasma glucose (dichotomous) in 4042 participants, a stratified by HbA1c.
| HbA1c | Model 1 | Model 2 | ||
|---|---|---|---|---|
| HR (95% CI) | P value | HR (95% CI) | P value | |
| Plasma glucose ≥ 105 vs < 105 mg/dL | ||||
| < 5.7% | 0.91 (0.65–1.27) | 0.574 | 0.93 (0.66–1.30) | 0.651 |
| 5.7%-6.4% | 1.30 (0.92–1.84) | 0.141 | 1.22 (0.86–1.74) | 0.267 |
| ≥ 6.5% | 1.56 (0.96–2.55) | 0.076 | 1.27 (0.74–2.18) | 0.394 |
| Plasma glucose ≥ 110 vs < 110 mg/dL | ||||
| < 5.7% | 1.01 (0.65–1.57) | 0.969 | 1.03 (0.66–1.59) | 0.913 |
| 5.7%-6.4% | 1.80 (1.15–2.81) | 0.011 | 1.67 (1.06–2.63) | 0.028 |
| ≥ 6.5% | 1.38 (0.88–2.16) | 0.161 | 1.06 (0.63–1.79) | 0.824 |
| Plasma glucose ≥ 115 vs < 115 mg/dL | ||||
| < 5.7% | 2.09 (1.19–3.68) | 0.011 | 2.17 (1.22–3.83) | 0.008 |
| 5.7%-6.4% | 1.94 (1.09–3.46) | 0.025 | 1.87 (1.05–3.33) | 0.034 |
| ≥ 6.5% | 1.46 (0.95–2.24) | 0.084 | 1.13 (0.68–1.89) | 0.640 |
CI, confidence interval; CVD, cardiovascular disease; HbA1c, hemoglobin A1c; HR, hazard ratio.
aThe late non-fasting sub-cohort included 5564 participants, among which 1522 died of non-CVD causes. This sensitivity analysis was conducted on the remaining 4042 participants after exclusion of these 1522 participants.
Model 1: adjusted for age, sex, ethnicity, obesity, education, poverty-income ratio, survey period, physical activity, alcohol consumption, smoking status, self-reported hypertension, and self-reported hypercholesterolemia.
Model 2: adjusted for all the factors in Model 1 plus natural log-transformed HbA1c.