Impact of perioperative corticosteroids on postoperative morbidity after pancreaticoduodenectomy: A systematic review, meta-analysis, and trial sequential analysis.
Corticosteroids may reduce the risk of postoperative pancreatic fistula after pancreaticoduodenectomy, but the evidence is not strong enough to recommend their routine use.
Where it sits
this study against the rest of the pasireotide corpusSummary and findings
This study evaluated the impact of perioperative systemic corticosteroids on postoperative morbidity in adults undergoing pancreaticoduodenectomy. The analysis included 454 patients across four randomized controlled trials, comparing corticosteroids with placebo or active comparators. The findings suggest a potential reduction in clinically relevant postoperative pancreatic fistula, but the evidence remains inconclusive.
Abstract
<h4>Background</h4>Postoperative morbidity after pancreaticoduodenectomy remains clinically relevant despite improvements in perioperative care and surgical technique. Perioperative systemic corticosteroids have been proposed to attenuate the inflammatory response after pancreatic surgery, but the certainty of the randomized evidence remains unclear.<h4>Methods</h4>We performed a systematic review, meta-analysis, and trial sequential analysis of randomized controlled trials evaluating perioperative systemic corticosteroids in adults undergoing pancreaticoduodenectomy. Eligible studies compared hydrocortisone, dexamethasone, or analogous systemic corticosteroids with placebo, no corticosteroids, or active pharmacologic comparators; trials with a placebo or no-treatment control constituted the primary analysis, and the trial with an active pharmacologic comparator was included in a sensitivity analysis. The main outcomes were mortality, major morbidity, and clinically relevant postoperative pancreatic fistula. Secondary outcomes included postpancreatectomy hemorrhage, delayed gastric emptying, reoperation, readmission, corticosteroid-related adverse events, and length of hospital stay. Trial sequential analysis was performed under a range of prespecified anticipated effects, and the certainty of the evidence for the critical outcomes was rated using Grading of Recommendations Assessment, Development, and Evaluation.<h4>Results</h4>Three randomized controlled trials with a placebo or no-treatment control (397 patients) constituted the primary analysis; a fourth trial with an active comparator (pasireotide) was added in a sensitivity analysis, for a total of 4 randomized controlled trials and 454 patients. Corticosteroid use was not associated with a significant reduction in mortality (primary analysis: risk difference, -0.00; 95% confidence interval, -0.03 to 0.03; sensitivity analysis: risk difference, -0.01; 95% confidence interval, -0.03 to 0.02). Major morbidity was numerically lower with corticosteroids, but the difference was not statistically significant (risk difference, -0.04; 95% confidence interval, -0.11 to 0.02 in both analyses). Clinically relevant postoperative pancreatic fistula was significantly reduced with corticosteroids (primary analysis: risk difference, -0.09; 95% confidence interval, -0.17 to -0.01; number needed to treat, 11; sensitivity analysis: risk difference, -0.08; 95% confidence interval, -0.16 to 0.00; number needed to treat, 12). Trial sequential analysis showed that the required information size was not reached for any of the critical end points under any prespecified anticipated effect (20%-30% relative risk reduction) or under the observed effect, in either analysis, suggesting that the current evidence remains underpowered. No significant differences were observed for delayed gastric emptying, postpancreatectomy hemorrhage, readmission, or length of stay in the primary analysis, whereas reoperation (risk difference, -0.04; 95% confidence interval, -0.08 to -0.01; number needed to treat, 23) was less frequent in the corticosteroid group. No trial reported an increase in corticosteroid-related adverse events (infections, hyperglycemia) with corticosteroids compared with controls.<h4>Conclusion</h4>Perioperative systemic corticosteroids may reduce the incidence of clinically relevant postoperative pancreatic fistula after pancreaticoduodenectomy, but the certainty of the randomized evidence is low to moderate, and the current data are insufficient to support their routine clinical use.
Background
Postoperative morbidity after pancreaticoduodenectomy is a significant concern despite advancements in surgical techniques. Previous studies suggested that perioperative systemic corticosteroids might mitigate the inflammatory response associated with surgery. This study aims to clarify the effectiveness of corticosteroids in reducing postoperative complications, particularly focusing on mortality and major morbidity.
Methods
This research involved a systematic review, meta-analysis, and trial sequential analysis of randomized controlled trials. The primary analysis included three trials with a placebo or no-treatment control, totaling 397 patients, while a fourth trial with an active comparator (pasireotide) was included in a sensitivity analysis, bringing the total to 454 patients. The main outcomes assessed were mortality, major morbidity, and clinically relevant postoperative pancreatic fistula, with secondary outcomes including various postoperative complications.
Results
The primary analysis indicated that corticosteroid use did not significantly reduce mortality (risk difference, -0.00; 95% confidence interval, -0.03 to 0.03). Although major morbidity was numerically lower with corticosteroids, this difference was not statistically significant (risk difference, -0.04; 95% confidence interval, -0.11 to 0.02). However, there was a significant reduction in clinically relevant postoperative pancreatic fistula (risk difference, -0.09; 95% confidence interval, -0.17 to -0.01; number needed to treat, 11).
Interpretation
The findings suggest a potential benefit of corticosteroids in reducing postoperative pancreatic fistula, aligning with some previous literature on their anti-inflammatory effects. However, the lack of statistically significant findings for mortality and major morbidity raises questions about the clinical relevance of the results. The study's limitations, including its underpowered nature and reliance on a small number of trials, suggest caution in applying these findings to clinical practice.
Key findings
- Corticosteroid use was not associated with a significant reduction in mortality (risk difference, -0.00; 95% confidence interval, -0.03 to 0.03).
- Major morbidity was numerically lower with corticosteroids, but the difference was not statistically significant (risk difference, -0.04; 95% confidence interval, -0.11 to 0.02).
- Clinically relevant postoperative pancreatic fistula was significantly reduced with corticosteroids (risk difference, -0.09; 95% confidence interval, -0.17 to -0.01; number needed to treat, 11).
- Trial sequential analysis showed that the required information size was not reached for any of the critical endpoints.
- No significant differences were observed for delayed gastric emptying, postpancreatectomy hemorrhage, readmission, or length of stay.
- Reoperation was less frequent in the corticosteroid group (risk difference, -0.04; 95% confidence interval, -0.08 to -0.01; number needed to treat, 23).
Limitations
- small n=454 across four trials
- evidence rated low to moderate certainty
- trial sequential analysis indicated underpowered results
- no significant differences in major morbidity or mortality
- single-site studies may limit generalizability