[Clinical assessment of peritoneal drainages for necrotizing enterocolitis. A Bayesian approach]

To assess effectiveness of peritoneal drainages for necrotizing enterocolitis.

Retrospective cohort study (years 2000 to 2006). Laparotomy or patient's death were considered as failure.

sex, gestational age, weight at delivery, Apgar score at minutes 1 and 5, modified Bell score, radiology and ventilatory status.

25 patients were diagnosed with necrotizing enterocolitis and treated with peritoneal drainages. Sample's Bell score was: 13(52%) Ia, 6 (24%) IIa, 5 (20%) IIb, and 1 (4%) IIIa. Mean gestational age was 31.8 (+/- 4.2) weeks, and mean weight 1,564 (+/- 810) g. Patients classified as Bell I presented statistically significat differences compared with Bell II-III as for radiology (unspecific), delivery weight (lower) and ventilatory status (higher mechanical ventilation rates). For the 12 patients with Bell scores II-III, peritoneal drains were enough for 5 cases (41.7%) and failed in 7 (58.3%), who were operated on. Multivariate analysis (logistic regression) was not able to show any conection with collected variables. However, a bayesian analysis using data from similar studies showed that the probability for drainage success rate to be higher than 50% is 99%.

In our centre, 52% of peritoneal drainages were used in patinets with low clinical suspect for necrotizing enetrocolitis, maybe in relation with their lower body weight and need for ventilatory support. In patients affected with necrotizing enterocolitis, drainages were effective in 41.7%. Although limited for its retrospective nature, our study suggests that peritoneal drainages can be curative in, at least, 50% of patients with necrotizing enterocolitis without pneumoperitoneum and clinical signs of peritonitis.