Hyponatremia in acute spinal cord injury

Objective To define the occurrence rate, time course, and potential etiologic factors of hyponatremia in patients with acute spinal cord injury. Design Analysis of data obtained from a retrospective review of medical records and from a systematized, prospective database pertaining to patients with spinal cord injury. Setting A university hospital with a federally funded regional spinal cord injury center and a dedicated spinal cord injury intensive care unit. Patients Two hundred eighty-two patients admitted between January 1, 1988 and December 31, 1989 with acute (<24-hr duration) spinal cord or vertebral column injury. Interventions None. Measurements and Main Results The mean age of patients was 36.7 ± 17.6 (SD) yrs; 225 (80%) of the patients were male and 57 (20%) were female. Hyponatremia, when it occurred, developed at a mean time of 6.4 ± 6.7 days postadmission, reached its nadir at 8.7 ± 8.8 days, and occurred in 28% of those patients with cervical injuries, 34% with thoracic injuries, and 27% with lumbar injuries (p = NS). Logistic regression analysis demonstrated that the type of spinal cord injury (Frankel class: range is A = complete neurologic lesion to E = no neurologic lesion) was the strongest predictor of hyponatremia. The occurence rate of hyponatremia was as follows: Frankel class-A 62%; Frankel class-B 48%; Frankel class-C 41%; Frankel class-D 23%; Frankel class-E 16% (p < .0001). Conclusions The prevalence of hyponatremia in acute spinal cord injury is much higher than in the general medical or surgical patient population. This abnormality usually occurs within the first week postinjury. The most significant predictor of hyponatremia is the type rather than the level of spinal cord injury. The potential etiological factors are many and these factors are probably interrelated. The pathophysiologic mechanisms that result in hyponatremia must be explored so that this occurrence and its consequences can be prevented. (Crit Care Med 1994; 22:252–258)