RESUMO
OBJECTIVE: To evaluate (99m)Tc-HMPAO leukocyte scintigraphy as an investigation for inflammatory bowel disease (IBD). STUDY DESIGN: Scintigraphy was performed in 95 children undergoing investigation for IBD in a tertiary Gastroenterology Department. Diagnosis was based on conventional investigations including small bowel barium contrast radiology (BCR), upper gastrointestinal endoscopy (UGIE), colonoscopy, and endoscopic biopsy (the "gold standards"). IBD was confirmed in 73 (57 Crohn's disease; 10 ulcerative colitis; 6 indeterminate colitis) and excluded in 22 (controls). Scintigraphy was (1) evaluated as a screening test, (2) compared with individual conventional tests, (3) assessed for each gut segment. RESULTS: Screening test: sensitivity 0.75 (95% CI, 0.63-0.85), specificity 0.82 (95% CI, 0.59-0.94), PPV 0.93, NPV 0.5. Comparison with BCR: sensitivity 0.87 (95% CI, 0.72-0.96), specificity 0.57 (95% CI, 0.39-0.73), PPV 0.69, NPV 0.2. Comparison with UGIE: specificity 0.9 (95% CI, 0.79-0.96), NPV 0.13 (sensitivity and PPV unavailable). Comparison with colonoscopy: sensitivity 0.57 (95% CI, 0.41-0.73), specificity 0.71 (95% CI, 0.54-0.85), PPV 0.71, NPV 0.42. Comparison with biopsies paralleled that with endoscopy. False negatives were especially common (NPV< or =0.2) in the proximal gut. CONCLUSIONS: (99m)Tc-HMPAO leukocyte scintigraphy should not be relied on as a screening test for IBD because false negative results are common. This method is especially unreliable at detecting disease in the proximal gut.