The Clinical Value of the Duodenal Biopsy: A Nation-Wide Clinicopathologic Analysis of 20,000 Patients

Purpose: Duodenal biopsies are routinely obtained from patients undergoing EGD for the investigation of iron-deficiency anemia, shown in prospective studies to be associated with celiac sprue in ∼3% of patients. The yield of these biopsies in clinical practice is unknown. Using a nation-wide sample of patients, we assessed the value of duodenal biopsies in relation to the indications for EGD. Methods: We analyzed data from Caris Diagnostics, a specialized gastrointestinal pathology group receiving specimens from endoscopy centers in 40 states; biopsy interpretation is maintained uniform through standardized procedures, diagnostic criteria, and terminology. The database includes demographic, clinical, and endoscopic information, site of origin, and histopathologic report for each biopsy. We extracted data from all patients who underwent EGD with gastric and duodenal biopsies from 4/07 to 3/08. Statistics were performed using SigmaStat 3.5; odd ratios were calculated using Hutchon's online calculator. A p value <0.05 was considered significant. Results: Duodenal biopsies were available from 20,461 unique patients (66% women; median age 52 years, range 0-92). Anemia was the main EGD indication in 3,079 patients; dyspepsia in 8,941; GERD in 5,970; and “rule out sprue” in 511. Irrespective of the indication, > 82% of duodenal biopsies were normal, except when sprue was suspected (71%, OR 1.98, 95% CI 1.63-21). Flattened mucosa consistent with sprue (Marsh 3b-3c) was seen in 0.9% of all patients, in 1.4% of those with anemia (OR 1.63, 95% CI 1.17-2.28); and in 10.8% of patients with suspected sprue (OR 13.9, 95% CI 10.1-19.0; p<0.001). Variable villous atrophy (Marsh 2-3a) was 10 times as common in patients with suspected sprue than in all other patients (7.8% vs. 0.8%; OR 9.76, 95% CI 6.84-13.94). Intraepithelial lymphocytosis (Marsh 1) was almost twice as common in patients with suspected sprue (7.4% vs. 4.1%) than in all others (OR 1.85, 95% CI 1.32-2.60; p<0.01). Marsh 1, 2, and 3a lesions were not more frequent in patients with anemia. Peptic duodenopathy was unrelated to the EGD indication. Conclusion: In this population duodenal biopsies were highly valuable in patients with suspected (or serologically documented) celiac sprue. Patients investigated for anemia had a slight but significantly greater prevalence of Marsh 3b and 3c lesions than patients who had EGD for other indications. This suggests that, perhaps combined with serologic tests, duodenal biopsies may have a role in the evaluation of anemia. In patients with other indications and no endoscopic abnormalities 5% to 10% of duodenal biopsies yielded clinically useful results.