Frailty Linked to Adverse Events After Upper-GI ESD
Key Takeaways
- Among Japanese adults aged at least 75 selected for first esophageal or gastric ESD, claims-defined frailty was significantly associated with more composite short-term adverse events.
- Perforation/peritonitis and delayed bleeding were each significantly more frequent in the claims-defined frail group.
- Death, thromboembolism, and aspiration pneumonia within 30 days did not differ significantly in the matched comparison.
Investigators retrospectively analyzed Japan’s nationwide DeSC Healthcare Late-Stage Elderly Healthcare System (LSEHS) claims component in their Journal of Gastroenterology study. The analysis covered April 2014 through February 2025 and included 15,184 eligible adults aged at least 75 undergoing first esophageal or gastric ESD. Patients with less than two years of observation, prior esophageal or gastric resection, or no corresponding hospitalization record were excluded. The Hospital Frailty Risk Score (HFRS) used weighted diagnosis codes from the two years before ESD; HFRS ≥5 indicated elevated claims-based risk, not measured physical function. Propensity-score matching yielded 4,970 pairs balanced on age, sex, procedure site, Charlson Comorbidity Index (CCI), facility ESD volume, outpatient visits, and medication categories. Outcome models adjusted for residual CCI and visit-frequency imbalance. The primary composite captured perforation/peritonitis during the index hospitalization, bleeding on days 2–30, and aspiration pneumonia, thromboembolism, or all-cause death within 30 days.
Composite adverse events occurred in 12.3% of frail versus 9.3% of non-frail participants in the matched cohort, an observed difference of 3.0 percentage points. The adjusted odds ratio (aOR) was 1.35 (95% confidence interval [CI] 1.19–1.53).
Perforation/peritonitis occurred in 4.5% versus 2.9% of frail and non-frail participants, respectively, with an aOR of 1.55 (95% CI 1.25–1.92). Delayed bleeding occurred in 7.1% versus 5.9% and was also significantly associated with frailty. The composite association appeared after both gastric and esophageal ESD, without significant interaction by site, although the esophageal estimate was less precise. Inverse-probability-of-treatment weighting (IPTW), a shorter outcome window, and exclusion of perforation/peritonitis from the composite yielded associations in the same direction. Aspiration pneumonia was associated with frailty in the IPTW analysis but not in the matched analysis. Frail participants also had lower overall survival during follow-up; median hospital stay was the same, hospitalization costs were modestly higher, and readmissions did not differ significantly.
Claims-based HFRS does not directly measure physical performance, nutrition, sarcopenia, or cognitive function. Diagnostic coding and frequency of healthcare contact can influence classification. Lesion size and complexity, procedure details, and operator experience were unavailable, leaving potential residual confounding, particularly for perforation/peritonitis; matching on facility volume cannot replace those measures. These observational associations do not establish causation, and neither changing frailty nor intensified monitoring or supportive care was evaluated. In Japan, scheduled hospitalization for ESD may limit how resource-use findings translate to other healthcare systems. U.S. endoscopy guidance addresses ESD for early esophageal and gastric cancers, but these Japanese claims-based associations have not been validated in U.S. care.
The authors concluded that preprocedural claims-defined frailty marked vulnerability to selected short-term events after upper-GI ESD, principally perforation/peritonitis and delayed bleeding, among older adults who underwent the procedure. They suggested that routine claims may contribute to characterizing that vulnerability.
Clinician Questions
Do frailty findings after upper-GI ESD extend to older patients who were not offered the procedure?
In the Japanese cohort, only older adults who underwent their first esophageal or gastric ESD contributed to the associations. People considered too frail for ESD may have been underrepresented, so these associations cannot automatically extend to all older adults with upper-GI neoplasia.
How was new delayed bleeding defined after esophageal or gastric ESD in the frailty comparison?
Investigators counted bleeding on days 2–30 only if it required endoscopic hemostasis or red blood cell transfusion. An outcome-specific diagnosis code in the preceding 90 days identified a pre-existing condition rather than a new complication.
Can lower survival in frail older adults after upper-GI ESD be attributed to the endoscopic procedure?
In this selected Japanese cohort, claims-defined frailty was associated with lower longer-term all-cause survival. Cause-specific mortality was unavailable, so the survival difference cannot be attributed to ESD or upper gastrointestinal neoplasia.