Acute diabetes insipidus in severe head injury: a prospective study.

BACKGROUND

The incidence and risk factors for acute diabetes insipidus after severe head injury and the effect of this complication on outcomes have not been evaluated in any large prospective studies.

METHODS

We conducted a prospective study of all patients admitted to the surgical ICU of a Level I trauma center with severe head injury (head Abbreviated Injury Score [AIS] >or= 3). The following potential risk factors with p < 0.2 on bivariate analysis were included in a stepwise logistic regression to identify independent risk factors for diabetes insipidus and its association with mortality: age, mechanism of injury (blunt or penetrating), blood pressure, Glasgow Coma Scale, Injury Severity Score, head and other body area AIS, skull fracture, cerebral edema and shift, intracranial hemorrhage, and pneumocephaly.

RESULTS

There were 436 patients (blunt injuries, 392; penetrating injuries, 44); 387 patients had isolated head injury. Diabetes insipidus occurred in 15.4% of all patients (blunt, 12.5%; penetrating, 40.9%; p < 0.0001) and in 14.7% of patients with isolated head injury (blunt, 11.8%; penetrating, 39.5%; p < 0.0001). The presence of major extracranial injuries did not influence the incidence of diabetes insipidus. Independent risk factors for diabetes insipidus in isolated head injury were Glasgow Coma Scale3. Diabetes insipidus was an independent risk factor for death (adjusted odds ratio, 3.96; 95% CI [1.65, 9.72]; adjusted p value = 0.002).

CONCLUSIONS

The incidence of acute diabetes insipidus in severe head injury is high, especially in penetrating injuries. Independent risk factors for diabetes insipidus include a Glasgow Coma Scale3. Acute diabetes insipidus was associated with significantly increased mortality.