Endoscopy Timing in Older Adults (≥75 Years) with Upper Gastrointestinal Bleeding Undergoing Esophagogastroduodenoscopy: A Retrospective Cohort Study

Introduction

Upper gastrointestinal bleeding (UGIB) is a common and potentially life-threatening condition, particularly in older adults who often have multiple comorbidities and reduced physiological reserve. Prior studies have highlighted the distinct clinical characteristics and outcomes of UGIB in very elderly populations [¹].

Current international guidelines recommend performing esophagogastroduodenoscopy (EGD) within 24 hours of presentation [²,³]. However, variation in endoscopy timing in real-world settings remains common, particularly among older adults. Frailty, comorbidities, and logistical constraints often influence clinical decision-making in this population.

Older adults represent a distinct population with higher comorbidity burden, frailty, and competing risks of mortality, which may modify the relationship between endoscopy timing and outcomes [⁴]. In this population, factors such as hemodynamic stabilization, anticoagulation management, anesthesia risk, and goals-of-care considerations may influence the timing of endoscopy. However, data describing real-world timing patterns specifically among older adults undergoing inpatient EGD remain limited.

The clinical benefit of early endoscopy in older adults remains uncertain, with mixed evidence from randomized trials and observational studies [⁵,⁶,⁷]. Therefore, we aimed to describe the timing of endoscopy and associated short-term clinical outcomes among older adults ($\geq$75 years) undergoing inpatient EGD for UGIB.

Methods

Results

A total of 108 patients aged $\geq$75 years were included in the final analysis. The mean age was 88.1 $\pm$ 6.1 years, with a median age of 90 years (IQR 83 – 93; range 75 – 100). Using the prespecified definition of early endoscopy as $\leq$24 hours from presentation, 35 (32.4%) of 108 patients underwent early EGD, and 73 (67.6%) underwent delayed EGD.

Baseline characteristics and endoscopic findings are summarized in Table 1, and clinical outcomes are presented in Table 2. Patients in the early and delayed EGD groups were broadly similar in age, hemoglobin, BUN, and major comorbidities, including cirrhosis, chronic kidney disease, and peptic ulcer disease. However, patients undergoing early EGD were more likely to present with shock (5 [14.3%] of 35 vs 1 [1.4%] of 73, p = 0.013), suggesting that patients with greater initial severity may have been prioritized for earlier intervention. No statistically significant differences were observed in active bleeding (4 [11.4%] of 35 vs 7 [9.6%] of 73, p = 0.74) or endoscopic therapy (16 [45.7%] of 35 vs 24 [32.9%] of 73, p = 0.21).

The mean door-to-scope time was 18.2 $\pm$ 6.4 hours in the early EGD group and 51.2 $\pm$ 44.7 hours in the delayed EGD group (p < 0.001). Across the entire cohort, the median time to EGD was 29.0 hours (IQR 23.0 – 45.3; range 4.0 – 339.0). Timing categories were distributed as follows: 35 (32.4%) of 108 patients underwent EGD within 24 hours, 52 (48.1%) between 24 and 48 hours, and 21 (19.4%) after 48 hours. Clinical outcomes are summarized in Table 2. Overall, 28-day all-cause mortality occurred in 5 (4.6%) of 108 patients. Mortality was 1 (2.9%) of 35 in the early EGD group and 4 (5.5%) of 73 in the delayed group (p = 1.00), with no clear difference observed between groups. The absolute risk difference for 28-day mortality was -2.6% (95% CI, -10.7% to 9.5%). Similarly, no statistically significant differences were observed in ICU admission (5 [14.3%] of 35 vs 5 [6.8%] of 73, p = 0.29), 30-day readmission (10 [28.6%] of 35 vs 21 [28.8%] of 73, p = 1.00), packed red blood cells (PRBCs) transfusion requirements (mean 1.7 $\pm$ 1.4 vs 1.9 $\pm$ 1.6 units, p = 0.56), or length of stay (median 3 vs 3 days, p = 0.16). The absolute risk differences were 7.4% (95% CI, -4.0% to 23.0%) for ICU admission, -0.2% (95% CI, -16.8% to 18.6%) for 30-day readmission, and 12.8% (95% CI, -6.2% to 31.6%) for endoscopic therapy. Estimates for all outcomes were imprecise due to the low number of events.

Discussion

In this real-world cohort of older adults with UGIB undergoing inpatient EGD, approximately one-third of patients underwent endoscopy within 24 hours, reflecting variation in timing within this scoped cohort. This finding is consistent with prior large-scale audits demonstrating variability in real-world management of UGIB [⁸].

Despite guideline recommendations emphasizing early endoscopy, no clear association was observed between early EGD and improved clinical outcomes, including mortality. These findings align with prior randomized and observational studies demonstrating mixed benefits of early endoscopy [⁵,⁶,⁷]. While early endoscopy facilitates timely diagnosis and therapeutic intervention, its impact on mortality remains inconsistent, particularly in lower-risk populations.

Importantly, our study population comprises an older cohort, in whom outcomes are likely influenced more by baseline comorbidities, frailty, physiologic reserve, and procedural timing. Prior studies have highlighted the distinct clinical characteristics and outcomes of UGIB in very elderly patients [¹]. Frailty, a multidimensional syndrome characterized by decreased physiologic reserve and increased vulnerability to stressors, may represent an important unmeasured confounder influencing both timing and outcomes [⁴]. These factors may have influenced both the timing of endoscopy and the observed outcomes in our study.

Patients undergoing early endoscopy in our cohort were more likely to present with shock, suggesting the presence of selection bias whereby patients perceived to be at higher risk are prioritized for earlier intervention. This phenomenon has been described in prior observational studies and may confound the relationship between timing and outcomes.

The relatively low overall mortality rate observed in this cohort likely reflects advances in supportive care, including early resuscitation, pharmacologic therapy, and modern transfusion strategies [⁹]. However, the low number of events limits statistical power and may obscure clinically meaningful differences between groups.

Consensus guidelines emphasize early endoscopy and risk stratification as key components of UGIB management, yet timing remains variable [¹⁰,²,³]. Delayed presentation, the need for clinical stabilization, resource limitations, and institutional workflow challenges may contribute to this variation, although these factors were not directly measured in the present study.

Our findings should be interpreted in light of several important sources of bias. Selection bias is present because only patients who underwent EGD were included. Survivor bias may occur because patients in the delayed group had to survive long enough to receive endoscopy. In addition, confounding by indication is likely, as patients with more severe presentations may have been prioritized for earlier endoscopy. These limitations restrict causal interpretation of the observed associations.

These findings are best interpreted as descriptive and hypothesis-generating and should not be used alone to guide clinical decision-making.

Conclusion

In older adults with UGIB undergoing inpatient EGD, no clear difference in short-term outcomes was observed between early and delayed endoscopy. These findings are descriptive and hypothesis-generating and should be interpreted cautiously given the modest sample size, low event count, and potential for bias.

Limitations

This study is limited by its retrospective single-center design, modest sample size, and low number of outcome events, which reduced statistical power and resulted in imprecise estimates.

The study included only patients who underwent inpatient EGD and does not represent all UGIB admissions. The analysis does not distinguish between medically appropriate and system-related delays in endoscopy. Code status and goals-of-care limitations were not available and may represent important sources of residual confounding. Outcomes occurring outside the health system may have been under-captured.