The evidence based dilemma of intraperitoneal drainage for pancreatic resection - a systematic review and meta-analysis

Abstract

Background: Routine placement of intraperitoneal drains has been shown to be ineffective or potentially harmful in various abdominal surgical procedures. Studies assessing risks and benefits of abdominal drains for pancreatic resections have demonstrated inconsistent results. We thus performed a systematic review of the literature and meta-analyzed outcomes of pancreatic resections with and without intraoperative placement of drains.

Methods: A database search according to the PRISMA guidelines was performed for studies on pancreatic resection with and without intraperitoneal drainage. The subgroup 'pancreaticoduodenectomy' was analyzed separately. The quality of studies was assessed using the MINORS and STROBE criteria. Pooled estimates of morbidity, mortality and length of hospital stay were calculated using random effects models.

Results: Only two randomized trials were identified. Their results were contradictory. We thus included six further, retrospective studies in the meta-analysis. However, with I2 = 68% for any kind of complication, the estimate of inter-study heterogeneity was high. While overall morbidity after any kind of pancreatic resection was lower without drains (p = 0.04), there was no significant difference in mortality rates. In contrast, pooled estimates of outcomes after pancreaticoduodenectomy demonstrated no differences in morbidity (p = 0.40) but increased rates of intraabdominal abscesses (p = 0.04) and mortality (p = 0.04) without intraperitoneal drainage.

Conclusion: Although drains are associated with slightly increased morbidity for pancreatic resections, routine omission of drains cannot be advocated, especially after pancreaticoduodenectomy. While selective drainage seems reasonable, further efforts to generate more reliable data are questionable because of the current studies and the presumed small differences in outcomes.

Trial registration: Systematic review registration number CRD42014007497.

Figure 1

Flow chart with the number of screened, assessed, and finally included studies in the meta-analysis.

Figure 2

Funnel plot for the primary outcome “any complication”. Despite the relatively low number of included studies (n = 8), no clear asymmetry was detected that would suggest selection bias. OR, odds ratio, SE(log_OR), standard error of the log OR.

Figure 3

Forrest plots for risk of any complication, minor and major complications for patients after all kinds of pancreatic resection, with and without intraperitoneal drain. The overall effect is shown by the diamond, with level of significance (p-value), 95% confidence intervals (CIs), and diversity between studies (Heterogeneity, I²).

Figure 4

Forrest plots for risk of fistula, abscess, and interventional radiology procedures for patients after all kinds of pancreatic resection, with and without intraperitoneal drain. The overall effect is shown by the diamond, with level of significance (p-value), 95% confidence intervals (CIs), and diversity between studies (Heterogeneity, I²).

Figure 5

Forrest plots for risk of re-operation and mortality, and difference in the length of hospital stay for patients after all kinds of pancreatic resection, with and without intraperitoneal drain. The overall effect is shown by the diamond, with level of significance (p-value), 95% confidence intervals (CIs), and diversity between studies (Heterogeneity, I²).

Figure 6

Forrest plots for risk of any complication, abscess, and mortality for patients after pancreaticoduodenectomy, with and without intraperitoneal drain. The overall effect is shown by the diamond, with level of significance (p-value), 95% confidence intervals (CIs), and diversity between studies (Heterogeneity, I²).