The high morbidity of colostomy closure after trauma: further support for the primary repair of colon injuries.

BACKGROUND: We examined the recent experience of a large urban trauma center to identify overall morbidity and factors predictive of outcome in patients undergoing colostomy closure after trauma.
METHODS: We did a retrospective analysis of 40 patients who underwent colostomy closure after trauma at our institution between January 1992 and August 1996.
RESULTS: The mechanism of injury was a gunshot wound in 30 patients (75%), a motor vehicle accident in 6 (15%), a stab wound in 3 (7.5%), and a rectal foreign body in 1 (2.5%). Loop colostomies were performed in 28 patients (70%) and end colostomies were performed in 12 patients (30%). Mean time until colostomy closure was 8 months (range, 0.5 to 28 months). Five patients underwent same admission colostomy closure (SACC). Contrast enemas were performed in 36 patients and found to be abnormal in 2 (6%) patients who were found during planning for SACC to have leaks from rectal trauma at 12 and 19 days after injury. Sixteen complications occurred in 12 patients (30%). Intraoperative complications occurred in two patients (5%) who sustained small and large bowel enterotomies. There were 4 major complications (1 fecal fistula, 1 anastomotic stricture, and 2 small bowel obstructions) in 3 patients (7.5%) and 10 minor complications (25%), 7 prolonged ileus and 3 superficial wound infections. Morbidity was significantly higher for patients whose initial injury involved the colon (11 of 20; 55%) as compared with those whose injury involved the rectum (2 of 16; 12.5%). The demographic, injury, and operative characteristics in the 12 patients with complications and the 28 patients without complications were compared to identify predictors of morbidity. The presence of a colon injury (RR = 7.70; p = 0.009) was a statistically significant predictor of morbidity after colostomy closure. The presence of an initial rectal injury, in contrast, was a predictor of low morbidity after closure (RR = 0.22; p = 0.024). No statistically significant differences were found with respect to age, gender, mode of injury, colostomy type, type of repair, need for laparotomy, or right- versus left-sided colostomy. Clinical trends were noted in five groups in whom the relative risk was greater than 2.0: age older than 30 versus less than 30 years (RR = 2.71; p = 0.079), end versus loop colostomy (RR = 2.33; p = 0.130), operative time greater than 2 versus less than 2 hours RR = 2.80; p = 0.141), estimated blood loss greater than 150 versus less than 150 cc (RR = 2.77; p = 0.079), and right- versus left-sided colostomy (RR = 2.00; p = 0.211). Patients with complications had significantly longer mean operative times (3.84 versus 2.46 hours; p = 0.02), higher mean blood loss (468 versus 142 cc; p = 0.006), and longer mean time until closure (11.3 versus 6.33 months; p = 0.02).
CONCLUSIONS: Colostomy closure after trauma remains associated with significant morbidity. The patients in whom a colon injury was the indication for initial colostomy experienced high morbidity (55%) after subsequent closure. Patients who had a colostomy for rectal injury had a low morbidity after closure (6.25%). Intraoperative difficulties (longer operative times, higher blood loss) and long delays until colostomy closure increase complication rates. Timely closure may improve outcome after operation for bowel continuity restoration. Morbidity associated with colostomy closure should be considered additional evidence for performing primary repair of colonic injuries. Because the morbidity of colostomy closure after rectal injuries is low, proximal colostomy for extraperitoneal rectal injuries should remain the treatment of choice.