When good haemostatic tools fail: lessons from delayed bleeding after EMR and ESD (1/2)
The prevention of delayed bleeding after advanced colorectal endoscopic resection represents a complex clinical challenge that appears resistant to single-modality prophylactic strategies.
1. The Limits of Mechanical Closure (De Cristofaro et al.)
In the largest Western multicentre study to date, De Cristofaro et al. evaluated the impact of complete prophylactic clip closure after colorectal endoscopic submucosal dissection (ESD) using a robust propensity score–matched design based on validated Limoges Bleeding Score (LBS) risk factors. Despite analysing more than 3,000 ESD cases and 944 matched pairs, complete defect closure did not significantly reduce clinically significant delayed bleeding (CSDB), even among patients receiving antithrombotic therapy or those classified as high risk (LBS 5–8). Importantly, this lack of benefit persisted despite closure being a deep, mechanically intuitive intervention and despite meticulous adjustment for known bleeding predictors.
2. Reinterpreting the PURPLE Trial
These findings have important implications for the interpretation of subsequent prophylactic haemostasis studies. In this context, the negative results of the PURPLE trial, which evaluated prophylactic application of a self-assembling peptide hydrogel (PuraStat) after EMR, should not be viewed as evidence of failure of the agent itself. Rather, when considered alongside De Cristofaro et al., PURPLE reinforces the concept that delayed bleeding is a multifactorial, time-dependent phenomenon that is unlikely to be prevented by a single local intervention, whether mechanical (clip closure) or topical (hydrogel application). The absence of benefit observed when PuraStat was applied as a stand-alone prophylactic measure after EMR is therefore pathophysiologically coherent, rather than contradictory, given that even complete closure of large ESD defects does not confer consistent protection.
3. Caution with Retrospective Data (Barone et al.)
Similarly, the multicentre retrospective case-control study by Barone et al., which reported a higher rate of delayed bleeding associated with self-assembling peptide gel use after EMR of large non-pedunculated lesions not amenable to closure, must be interpreted with caution. This study inherently selected a highly enriched high-risk population in whom the decision to apply the gel likely reflected unmeasured intraprocedural factors such as extent of thermal injury, presence of high-risk exposed vessels, operator concern, or centre-specific haemostatic practice. In such a setting, confounding by indication is highly probable, and the observed association cannot be taken as proof of causality or harm, particularly in light of small event numbers and wide confidence intervals.
4. The Value of Multimodal Strategy (POPS Trial)
In contrast, data from the POPS trial, a large multicentre prospective observational study evaluating PuraStat during high-risk tissue resection (HRTR) across both EMR and ESD, provide important real-world context. In POPS, PuraStat was frequently used as an adjunct to conventional haemostatic techniques, including targeted coagulation and focal clipping, rather than as an isolated prophylactic intervention. Within this multimodal haemostatic strategy, high intraprocedural haemostatic efficacy and a low overall delayed bleeding rate were observed. Although the absence of a control arm precludes causal inference, POPS supports the concept that the clinical value of self-assembling peptide hydrogels lies in their complementary role within layered haemostasis, rather than in replacement of deep vessel control.
More thoughts in the next post.
#3Dmatrix
#Endoscopy
#GIbleeding
#DelayedBleeding
#PURPLEtrial
#POPStrial
#PuraStat
#EMR
#ESD
#MultimodalHaemostasis
#EvidenceBasedMedicine