Table 2.
Excess cause-specific mortality at ages 35–74 years associated with previously diagnosed or undiagnosed diabetes at recruitment
|
Number of deaths |
Death RR (95% CI) vs no diabetes* |
Attributable mortality† | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| No diabetes (n=110 181) | Undiagnosed diabetes (n=6541) | Previously diagnosed diabetes |
No diabetes | Undiagnosed diabetes | Previously diagnosed diabetes |
|||||
| HbA1c <9% (n=8475) | HbA1c ≥9% (n=8465) | HbA1c <9% | HbA1c ≥9% | |||||||
| All-cause mortality | 3944 | 595 | 1184 | 1960 | 1·0 | 2·1 (1·9–2·2) | 3·0 (2·8–3·3) | 5·2 (4·9–5·5) | 35% | |
| Any vascular, renal, or infectious | 1709 | 379 | 822 | 1491 | 1·0 | 3·0 (2·7–3·4) | 4·9 (4·5–5·3) | 9·1 (8·5–9·8) | 51% | |
| Renal | 224 | 115 | 301 | 679 | 1·0 | 8·0 (6·4–10·1) | 15·8 (13·2–18·9) | 33·6 (28·8–39·3) | 79% | |
| Cardiac | 697 | 127 | 244 | 362 | 1·0 | 2·4 (2·0–2·9) | 3·3 (2·8–3·8) | 5·3 (4·7–6·0) | 37% | |
| Infectious | 434 | 86 | 166 | 266 | 1·0 | 2·8 (2·2–3·5) | 4·0 (3·3–4·8) | 6·6 (5·6–7·7) | 43% | |
| Cerebrovascular | 268 | 37 | 79 | 132 | 1·0 | 1·7 (1·2–2·4) | 2·6 (2·0–3·3) | 4·7 (3·8–5·8) | 32% | |
| Other vascular | 86 | 14 | 32 | 52 | 1·0 | 2·1 (1·2–3·8) | 3·5 (2·3–5·3) | 6·3 (4·4–9·0) | 40% | |
| Acute diabetic crises‡ | 48 | 28 | 86 | 156 | .. | .. | .. | .. | 100% | |
| Neoplastic | 1051 | 85 | 102 | 121 | 1·0 | 1·1 (0·9–1·4) | 1·0 (0·8–1·2) | 1·2 (1·0–1·5) | 2% | |
| Cirrhotic | 436 | 42 | 58 | 55 | 1·0 | 1·2 (0·9–1·7) | 1·4 (1·0–1·8) | 1·3 (1·0–1·8) | 6% | |
| Chronic obstructive pulmonary disease | 136 | 12 | 27 | 20 | 1·0 | 0·9 (0·5–1·7) | 1·5 (1·0–2·3) | 1·2 (0·8–2·0) | 6% | |
| Ill-defined, other, or external | 564 | 49 | 89 | 117 | 1·0 | 1·2 (0·9–1·7) | 1·7 (1·4–2·2) | 2·3 (1·9–2·8) | 14% | |
RR=rate ratio.
Death RR estimates for those with versus without diabetes at recruitment were adjusted for age, sex, district, educational level, smoking status, and anthropometric measures.
For each of previously diagnosed diabetes with HbA1c ≥9%, previously diagnosed diabetes with HbA1c <9%, and undiagnosed diabetes, the number of deaths attributable to the excess risk associated with that level of diabetes was calculated as number of deaths × (RR–1)/RR, where RR is the cause-specific death RR for that group relative to those without diabetes. These three numbers were then summed and presented as a percentage of all such deaths. For example, for renal death, the calculation was 100 × ([679 × 32·6/33·6] + [301 × 14·8/15·8] + [115 × 7·0/8·0])/1319=79%.
Death RR estimates are not shown for deaths attributed to acute diabetic crises as all such deaths were due to diabetes, irrespective of whether diabetes was diagnosed before recruitment.