would you recommend the below to be a good meta-analysis focused topic with good impact and feasibility Outcomes of endovascular versus surgical management of post-pancreatectomy hemorrhage: a systematic review and pooled proportion meta-analysis
Searching PubMed
post-pancreatectomy hemorrhage endovascular surgical management outcomes
Searching PubMed
post-pancreatectomy hemorrhage meta-analysis systematic review
Searching the Web
post-pancreatectomy hemorrhage endovascular vs surgical systematic review meta-analysis 2023 2024 2025
Searching PubMed
post-pancreatectomy hemorrhage management outcomes angioembolization reoperation
Reading Article
Searching PubMed
postpancreatectomy hemorrhage diagnosis management systematic review meta-analysis
Searching the Web
van Oosten 2019 post-pancreatectomy hemorrhage meta-analysis HPB endovascular surgical comparison
I now have enough information for a thorough, well-evidenced evaluation. Here is my full assessment:
Evaluation: Post-Pancreatectomy Hemorrhage (PPH) - Endovascular vs. Surgical Management Meta-Analysis
Short verdict: Conditionally recommended - the topic has genuine clinical importance and a real evidence gap, but you must carefully position it relative to the existing 2019 meta-analysis and make several methodological decisions before proceeding.
1. Clinical Impact Assessment - STRONG
Post-pancreatectomy hemorrhage (PPH) is a rare but high-stakes complication:
- Incidence: 3-16% after pancreatectomy (weighted mean ~5%)
- Overall mortality: ~21% in pooled data (van Oosten et al., 2019, PMID 30962134)
- Mortality gap is wide: 16% with primary endovascular vs. 37% with primary relaparotomy (van Oosten 2019)
- Grade C PPH (severe, life-threatening) carries mortality exceeding 50% in surgical series
- The ISGPS classification (early vs. late, intraluminal vs. extraluminal, Grade A/B/C) is now well-established, giving any new meta-analysis a standardized framework to work within
The clinical question - whether to go endovascular first or straight to OR - is a daily high-stakes decision at HPB centers. The topic is publishable in HPB-tier journals (HPB, Annals of Surgical Oncology, JHBPB, Langenbeck's).
2. Existing Literature Landscape - IMPORTANT CONSTRAINT
The key existing paper you must address:
van Oosten A et al. (2019). "Diagnosis and management of postpancreatectomy hemorrhage: a systematic review and meta-analysis." HPB (Oxford) 21(8):953-961. [PMID 30962134]
This is a direct precedent. It included 14 studies, 467 patients with late PPH, and found:
- Interventional angiography mortality: 16%
- Relaparotomy mortality: 37%
- Endoscopy mortality: data also reported
Critical finding: No meta-analysis specific to endovascular vs. surgical comparison has been published since 2019. A 2026 single-center study from Turkey (Aydogan et al., BMC Surgery, PMID 41782132) reported 60% surgical vs. 25% endovascular mortality, reinforcing the gap - and there have been multiple new single-institution series since 2019 (covered stents, TAE, REBOA, hybrid approaches), none synthesized in a new meta-analysis.
This is your differentiation opportunity, but you must be explicit about how your work updates and expands van Oosten 2019.
3. Feasibility Assessment
Evidence Pool
| Factor | Assessment |
|---|---|
| Dedicated PPH series (2019-2026) | Moderate - ~15-25 new studies likely identifiable |
| Comparative data (endo vs. surgery) | Limited in direct RCTs - all observational/retrospective |
| Standardized outcomes reported | Partially - ISGPS grading not uniformly used pre-2016 |
| Study heterogeneity expected | HIGH - case series, single-center, variable definitions |
What Makes it Feasible
- PPH is well-defined by ISGPS criteria (since 2007) - gives inclusion/exclusion clarity
- Most series report technical success, rebleeding, mortality, and hospital stay - enough for pooled proportions
- A pooled proportion meta-analysis (as you've specified) is the methodologically correct choice here given the absence of RCTs - this is appropriate and matches what the literature can support
- The topic does NOT require direct head-to-head RCT data to be valid - proportion-based comparisons with sensitivity analyses are publishable
What Makes it Challenging
- Selection bias is irreducible: endovascular is generally selected for hemodynamically stable patients; sicker patients go to OR - this confounding is nearly impossible to adjust for in a pooled proportion design
- Study-level data only: most primary studies won't allow subgroup adjustment - you need to acknowledge this as a primary limitation
- Small sample sizes per study: PPH is uncommon; many series have < 30 patients; heterogeneity (I²) will likely be very high
- Pooled proportions are non-comparative by nature: you'll be comparing two separate pools (endovascular studies and surgical studies), not head-to-head matched data - reviewers will push back on causal language
4. Key Refinements to Strengthen the Proposal
A. Narrow the scope strategically:
- Focus on late PPH only (>24h, Grade B/C) - this is where the intervention decision is most clinically contested and where most published series focus
- Specify extraluminal bleeding separately from intraluminal - these are managed very differently
B. Define your primary outcomes clearly upfront:
- Primary: 30-day/in-hospital mortality (pooled proportion per arm)
- Secondary: technical success rate, rebleeding rate, complication rate (hepatic ischemia, abscess after TAE), need for crossover to surgery, hospital LOS
C. Distinguish endovascular subtypes:
- TAE (coil embolization) vs. covered stent placement - these have meaningfully different hepatic complication profiles and rebleeding rates; lumping them reduces clinical utility
D. Add a quality-of-evidence layer:
- Use GRADE or the Newcastle-Ottawa scale for cohort data; this significantly increases publication acceptance at top-tier journals
E. Consider ISGPS Grade C only as a subgroup:
- Grade C is where management choice matters most; Grade A is self-limiting and Grade B often resolves conservatively - a Grade C-focused subgroup analysis would be the highest-impact finding
5. Overall Recommendation
| Dimension | Score | Comment |
|---|---|---|
| Clinical impact | ★★★★☆ | High-stakes decision, real mortality signal |
| Novelty | ★★★☆☆ | 2019 precedent exists; gap since then is ~7 years of new data |
| Feasibility (data volume) | ★★★☆☆ | Enough for pooled proportions; not enough for direct comparison meta-analysis |
| Methodology fit | ★★★★☆ | Pooled proportion design is exactly right for available data |
| Publication potential | ★★★☆☆ | HPB, Pancreatology, Langenbeck's, JHBPB are realistic targets |
Recommendation: Yes, proceed - but reframe the title slightly. Instead of positioning it as head-to-head ("versus"), which implies direct comparison you cannot make rigorously, consider:
"Outcomes of endovascular and surgical management of post-pancreatectomy hemorrhage: a systematic review and meta-analysis of pooled proportions"
The word "versus" implies a controlled comparison; "and" accurately reflects the pooled proportion design and will reduce reviewer pushback while preserving the comparative interpretation in your discussion. Your discussion section can still compare the two arms and make practice recommendations - but the design label should match the methodology.
One additional check: run a PROSPERO search before registering to confirm no group has registered this exact protocol since 2022. The 2019 van Oosten paper leaves a clear 7-year update window, and the 2026 single-center Turkish paper (PMID 41782132) explicitly calls for larger pooled analyses - which is a direct citation you can use to justify your study's rationale.
This is a shared conversation. Sign in to Orris to start your own chat.