Diabetes is consistently an independent predictor of mortality in population studies of
• In Scotland, diabetes was as an independent predictor of mortality in both genders with CHF: hazard ratio 1.55 (1.41-1.70) and 1.50 (1.38-1.62) for men and women, respectively (Macintyre etal., 2000).
• In Rotterdam, the presence of diabetes conferred a worse prognosis in patients with CHF: hazard ratio 3.19 (1.80-5.65) (Mosterd etal., 2001).
• In the USA, diabetes was an independent predictor of mortality in 170 239 Medicare patients, with hazard ratios of 1.11 (1.06-1.16, P < 0.05) in Black patients and 1.22 (1.24-1.25, P < 0.05) in White patients (Croft etal., 1999).
• In another small study of 495 patients with CHF, diabetes independently predicted mortality: odds ratio 1.71 (1.16-2.51, P = 0.0065) (Kamalesh and Nair, 2005).
Only one population study has considered whether or not diabetes might predict mortality according to aetiology of CHF. In 1246 French patients, diabetes was a risk factor for mortality only in patients with CHF secondary to CHD (hazard ratio 1.54 (1.13-2.09, P=0.006) but not for those with CHF secondary to other aetiologies (hazard ratio 0.65 (0.39-1.07, P = 0.09) (de Groote etal., 2004).
In clinical trials of CHF, patients with diabetes have a consistently higher mortality rate than patients without diabetes (Ryden etal., 2000; Erdmann etal., 2001; Haas etal., 2003; Deedwania etal., 2005). Diabetes is an independent predictor of mortality in patients with CHF (Shindler etal., 1996; Dries etal., 2001; Domanski etal., 2003; Brophy etal., 2004; Gustafsson etal., 2004; Pocock etal., 2006).
In the SOLVD (enalapril versus placebo), BEST (bucindolol versus placebo) and DIG (digoxin versus placebo) studies, diabetes was an independent risk factor for mortality in patients with CHF. In these three trials the increased risk appeared to be confined to patients with CHF due to CHD. In the SOLVD study, the hazard ratio (HR) for those with diabetes was 1.29 (1.1-1.5). The HR was 1.37 (1.21-1.55) and 0.98 (0.76-1.32) for CHF secondary to CHD and non-CHD, respectively (Shindler etal., 1996; Dries etal., 2001). In the BEST study, the HR for those with diabetes and CHF secondary to CHD and non-CHD was 1.33 (1.12-1.58, P=0.001) and 0.98 (0.74-1.30, P = 0.89), respectively (Domanski etal., 2003). In the DIG study, the HR for those with diabetes and CHF secondary to CHD was 1.43 (1.26-1.63) (Brophy etal., 2004). In DIG, no HR is available for CHF of non-CHD aetiology.
That diabetes is a predictor of mortality only in those with CHF due to CHD is not a consistent finding. Both DIAMOND-CHF (dofetilide versus placebo in CHF) and CHARM (candesartan versus placebo in CHF) reported that diabetes was an independent predictor of mortality regardless of the aetiology of CHF (Gustafsson etal., 2004; Pocock etal., 2006). In the CHARM study, patients with CHF and diabetes treated with insulin had an 80% increased risk of death compared to those without diabetes (HR 1.80 (1.56-2.08)). Patients with CHF and diabetes not treated with insulin had a 50% increased risk of death compared to those without diabetes (HR 1.50 (1.34-1.68)) (Pocock etal., 2006).
Subgroup analysis of the Framingham study suggested that diabetes might be a predictor of mortality in women but not in men with CHF (HR 1.70 (1.21-2.38) and 0.99 (0.70-1.40) for women and men, respectively) (Ho etal., 1993). The only clinical trial to report outcome of patients with diabetes by gender (the DIAMOND-CHF study) did not report this apparent mortality difference (Gustafsson etal., 2004). The relative risk for mortality was 1.7 for women (95% CI 1.4-1.9, P < 0.0001) and 1.4 for men (95% CI 1.3-1.6, P< 0.0001).
Was this article helpful?