One-Year Incidence of Hyperosmolar States and Prognosis in a Geriatric Acute Care Unit

Isabelle Bourdel‐Marchasson, Sebastien Proux, Patrick Dehail, François Muller, S. Richard‐Harston, T. Traissac, M. Rainfray

Gerontology · 2004 · 32 citations · 25 references

Concepts

Abstract

BACKGROUND: Hyperosmolar syndromes are associated with high mortality rates, yet little is known about their incidence and their prognosis. OBJECTIVE: To determine the 1-year incidence of hyperosmolar states and the prognostic factors for in-hospital and 1-year mortality. METHOD: A 6-month prospective cohort study was conducted in a 40-bed acute care geriatric unit and included all patients who developed plasma osmolarity of 320 mosm/l or greater. Age, sex and known cognitive impairment as possible risk factors of hyperosmolarity were assessed. In-hospital and 1-year mortality were calculated and risk factors for death among baseline patient characteristics were sought. RESULTS: 48 (11) of the 436 inpatients in the study were identified as hyperosmolar. Diabetic hyperosmolarity was found in 8 patients. Cognitive impairment was a risk factor for hyperosmolarity (relative risk 2.39, 95% confidence interval 2.18-3.33, p < 0.001), but not age or sex. Infections were accompanied by hyperosmolarity in 30 (62.5). Thirty-five patients (72.9) were bed- or chair-ridden. In-hospital mortality was higher in hyperosmolar patients (35.4) than in the others (16.7%, p = 0.003). Causes of death were infection in 5 (29.4), terminal cachexia in 5, thrombosis in 3, gastric bleeding in 1, renal failure in 2 and heart failure in 1. Functional dependency for mobility was a risk factor for in-hospital mortality but not the degree of hyperosmolarity. One-year mortality was 68.7%. Functional dependency and pressure ulcers were independent predictors of 1-year mortality (p = 0.005 and p = 0.044, respectively). CONCLUSION: Hyperosmolar states occurred in cognitively impaired and dependent patients and resulted in high mortality rates at short and at mid-term. Mortality was related to functional dependency rather than to hyperosmolarity.

References

25