Pediatric short bowel syndrome: redefining predictors of success

Abstract

Objective: To determine predictors of survival and of weaning off parenteral nutrition (PN) in pediatric short bowel syndrome (SBS) patients.

Summary background data: Pediatric SBS carries extensive morbidity and high mortality, but factors believed to predict survival or weaning from PN have been based on limited studies. This study reviews outcomes of a large number of SBS infants and identifies predictors of success.

Methods: Multivariate Cox proportional hazards analysis was conducted on 80 pediatric SBS patients. Primary outcome was survival; secondary outcome was ability to wean off PN. Nonsignificant covariates were eliminated. P < 0.05 was considered significant.

Results: Over a mean of 5.1 years of follow-up, survival was 58 of 80 (72.5%) and 51 weaned off PN (63.8%). Cholestasis (conjugated bilirubin > or =2.5 mg/dL) was the strongest predictor of mortality (relative risk [RR] 22.7, P = 0.005). Although absolute small bowel length was only slightly predictive, percentage of normal bowel length (for a given infant's gestational age) was strongly predictive of mortality (if <10% of normal length, RR of death was 5.7, P = 0.003) and of weaning PN (if > or =10% of normal, RR of weaning PN was 11.8, P = 0.001). Presence of the ileocecal valve (ICV) also strongly predicted weaning PN (RR 3.9, P < 0.0005); however, ICV was not predictive of survival.

Conclusions: Cholestasis and age-adjusted small bowel length are the major predictors of mortality in pediatric SBS. Age-adjusted small bowel length and ICV are the major predictors of weaning from PN. These data permit better prediction of outcomes of pediatric SBS, which may help to direct future management of these challenging patients.

FIGURE 1. Survival was assessed using multivariate Cox proportional hazards analysis for 80 neonatal patients with short bowel syndrome (SBS). Cholestasis (conjugated bilirubin ≥2.5 mg/dL) was highly predictive of mortality.

FIGURE 2. Survival of pediatric SBS patients was stratified according to percentage of expected normal bowel length remaining (based on a subject's gestational age). Risk of mortality strongly increased in infants with <10% of the expected small intestinal length for a given gestational age.

FIGURE 3. Two very strong predictors of mortality in pediatric SBS patients are cholestasis (conjugated bilirubin ≥2.5 mg/dL) and length of small intestine <10% of the expected normal length for an infant's gestational age. As a useful approximation, 10% of the expected normal length (in centimeters) is equal to gestational age minus 10. For example, a normal 35-week infant is expected to have ≈250 cm of bowel; thus, 10% of the normal length at 35 weeks is about 25 cm.

FIGURE 4. Percentage of normal bowel length was highly predictive of ability to wean off PN. Infants with at least 10% or more of the expected small intestinal length for a given gestational age were significantly more likely to wean off PN. The time variable (days) is the time from onset of SBS to the time that a patient achieved complete independence from PN (or until last follow-up, if not weaned).

FIGURE 5. Presence of the ICV was a highly significant predictor of the ability to wean off PN. The time variable (days) is the time from onset of SBS to the time that a patient achieved complete independence from PN (or until last follow-up, if not weaned).