Background: Studies on the association between mortality and delirium in older hospital inpatients have produced conflicting results. This insconsistency might be explained by case-mix differences in terms of clinical or underlying patho-physiological processes. For example, both albumin and C-reactive protein (CRP) have been reported as predictors of in-hospital mortality and interleukin-6 of longer-term mortality.
Methods: We used data from a longitudinal study of delirium to investigate the delirium-mortality relationship. A cohort of 164 patients, 70+ years were assessed within 3 days of acute hospital admission and hence twice weekly until hospital discharge, for the presence and severity of delirium and a range of clinical and laboratory measures, including initial albumin (n = 149), CRP (n = 76) and cytokine (n = 60) levels. In-hospital and 6-months mortality were determined from clinical records and telephone contact.
Results: During hospitalisation 14 (8.5%) patients died, 6 with delirium: mortality was not associated with delirium. At 6 months, 119 of 150 (77.3%) discharged patients were still alive, 21 (14.0%) dead, and 13 (8.7%) uncontactable. In bivariate analysis, 6-months mortality was associated with older age (P = 0.013), lower albumin (P = 0.001), higher CRP (P = 0.014) and higher interleukin-6 levels (P = 0.007), but not with presence or severity of in-hospital delirium. After controlling for other variables significant predictors (P < 0.05) for six-month mortality were initial MMSE, albumin, interferon-lambda and interleukin-6.
Conclusions: The lack of demonstrable association between delirium and mortality may reflect inadequate statistical power in this study due to low numbers. These findings, however, highlight specific patho-physiological factors which may be important in the prognosis after delirium.