Case Presentation: A 76 year-old woman with coronary artery disease, hypertension, and chronic kidney disease presented to the ER with a one-day history of profound generalized weakness and acute severe anemia. She had known diverticulosis and was taking aspirin 81 mg twice daily status-post a left total hip arthroplasty three weeks prior. Review of systems was negative for abdominal pain, hematemesis, melena and hematochezia. In the ER, she was hypotensive with left thigh edema. Labs (baseline values) showed hemoglobin 4.3 (11.7), MCV 100 (89.8), BUN 81 (45), creatinine 1.99 (1.43) and positive fecal occult blood test. Head CT was negative. Abdomen and pelvis CT showed chronic diverticulosis. The working diagnosis was diverticular bleeding.The admitting Medicine team suspected left thigh hematoma as the primary bleeding source. However, hip imaging showed a postoperative seroma and incidental fracture (Timeline, Fig 1). The consulting Gastroenterology team obtained additional history of darker stools post-surgery and proposed a working diagnosis of upper GI bleed. Urgent EGD on hospital day 1 (HD1) showed a single oozing superficial duodenal ulcer and diminutive nonbleeding erosions in the gastric antrum. On HD2, she became hemodynamically unstable (hemoglobin 6.0, down from 8.2 post-transfusion, Fig 2) after passing ≥ 8 black and then bright red stools. Following ICU transfer, a repeat EGD was negative. Colonoscopy was advised on HD3 and performed on HD5, showing diffuse moderate diverticulosis with stigmata of recent bleeding. The final diagnosis was severe gastrointestinal hemorrhage likely secondary to diverticular source, preceded by a subacute duodenal bleed.
Discussion: This case was included in a large retrospective study (UPSIDE) investigating diagnostic errors in the hospital. The missed opportunity to make a timely diagnosis of diverticular bleeding constitutes a diagnostic error. A protracted stay and interval ICU course imply temporary patient harm.Several contextual factors likely contributed to a delayed diagnosis. First, initial exam findings were likely overvalued since ER providers’ concern for lower GI bleed (LGIB) was dismissed in favor of a low-yield evaluation to exclude left thigh hematoma. Second, failure to pursue urgent colonoscopy despite the patient’s known diverticulosis and life-threatening anemia suggests diagnostic “anchoring” on a bleeding source in the upper GI tract due to suboptimal weighing of all data. Third, delayed discovery of darker stools post-surgery may have impeded the diagnostic process since melena may occur in right-sided LGIB. Fourth, subacute EGD findings confirmed the presumed (sole) bleeding source, but were likely overvalued, swaying the focus away from alternative or additional diagnoses. Of note, interval hematochezia was potentially pathognomonic for LGIB, making a repeat EGD (and delaying colonoscopy) controversial. Next, omission of LGIB from the differential diagnoses of the inpatient teams was not reassessed until HD3, suggesting gaps in active discourse or inadequate shared decision-making among providers. Lastly, once advised, colonoscopy was delayed by another 2 days, implying barriers to non-urgent weekend procedures.
Conclusions: Flawed clinical reasoning and suboptimal collaboration among consulting teams may have significantly hindered the diagnostic process during this encounter. In hindsight, same-day upper and lower endoscopy may have expedited a final diagnosis of severe LGIB with subacute duodenal bleeding.
