Search

Background: Many patients with upper gastrointestinal (GI) bleeding have a benign outcome and could receive less intensive and costly care if accurately identified. We sought to determine whether early endoscopy performed shortly after admission in the emergency department could significantly reduce the health care use and costs of caring for patients with nonvariceal upper GI bleeding without adversely affecting the clinical outcome. Methods: All eligible patients with upper GI bleeding and stable vital signs were randomized after admission to undergo endoscopy in 1 to 2 days (control) or early endoscopy in the emergency department. Patients with low-risk findings on early endoscopy were discharged directly from the emergency department. Clinical outcomes and costs were prospectively assessed for 30 days. Results: We randomized 110 consecutive stable patients with nonvariceal upper GI bleeding during the 12-month study period. The baseline demographic features, endoscopic findings, and the clinical outcomes were no different between the two groups. However the findings of the early endoscopy allowed us to immediately discharge 26 of 56 (46%) patients randomized to that group. No patient discharged from the emergency department suffered an adverse outcome. The hospital stay (median of 1 day [interquartile range of 0 to 3 days] vs. 2 days [interquartile range of 2 to 3 days], p = 0.0001) and the cost of care ($2068 [interquartile range of $928 to $3960] versus $3662 [interquartile range of $2473 to $7280], p = 0.00006) were significantly less for the early endoscopy group. Conclusions: Early endoscopy performed shortly after admission in the emergency department safely triaged 46% of patients with nonvariceal upper GI bleeding to outpatient care, which significantly reduced hospital stay and costs. (Gastrointest Endosc 1999;50:755-61.)

The use of diagnostic and therapeutic endoscopic retrograde cholangiopancreatography (ERCP) is increasing in the management of pancreatobiliary diseases in children.Over a 32-month period, we performed 34 ERCP procedures for the treatment of pancreatitis in 22 children at two university hospitals. Demographics and clinical data and ERCP findings were documented. Clinical status was assessed 6 months before the first ERCP and 6 months after the last ERCP, according to general condition, severity and frequency of pain, and health care encounters (emergency department visits, clinic visits, and hospital admissions related to the pancreatitis).Mean age of the patients was 10.7 years (range 1.5 to 17 years). Abdominal pain was the main presenting symptoms with hyperamylasemia and hyperlipasemia. Clinical diagnoses included acute pancreatitis (6), recurrent pancreatitis (5), and chronic pancreatitis (11). The mean follow-up was 16.4 months. Nine patients had sphincter manometry, with abnormal results leading to biliary sphincterotomy in 4. Fifteen patients underwent a total of 23 therapeutic ERCP procedures unrelated to sphincter dysfunction. There were 2 complications of 34 procedures (6%), both being mild pancreatitis after sphincter manometry. There were no deaths. There was a significant reduction in frequency (p0.01) and severity of pain (p0.01) after intervention. Patients without pancreatographic changes of chronic pancreatitis had the most marked clinical improvement (p0.05). In those with ductal changes of chronic pancreatitis, clinical improvement was not predicted by the extent of ductal changes. There was a significant decrease in health care encounters (p0.05) and improvement in general condition (p0.01) after endoscopic therapy, especially in those with a normal pancreatogram.Therapeutic ERCP is safe in pediatric patients with pancreatitis. Significant clinical improvement is achieved in patients with biliary or pancreatic stone disease. Prospective studies with long-term follow-up are needed to determine the impact of endoscopic therapy in patients with chronic pancreatitis and sphincter of Oddi dysfunction.

SHOULD WE RELY ON RAPID UREASE TEST OR ROUTINE STAINING FOR THE ACCURATE DIAGNOSIS OF HELICOBACTER PYLORI RK Hsu MD M Chang MD, A Yu MD, B Yost* MD, B Ruebner* MD, JG Lee MD and JW Leung MD. Division of Gastroenterolgy and Department of Pathology*, University of California, Davis Medical Center, Sacramento CA Introduction Rapid Urease Test and histology are two common endoscopic tests for Hp detection. Discrepancies between these two tests can cause confusion in the actual diagnosis. In praetice, Genta stain is more sensitive than H&E and Giemsa stain but are not used in routine. The aim of this study is to examine the error rate of RUT and histology using routine staining as reference to the consensus results by pathologists with special interest in GI pathology. Method: Standard antral biopsy were obtained from 40 outpatients presented with ulcer and nonulcer dyspepsia using Wilson Cook Maxum biopsy forceps with cup diameter 2.5mm. The patients were not on treatment with proton pump inhibitors. The HP status was detected using Hpfast rapid urease test (RUT) and initial histology by routine H&E or Giemsa stain interpreted by general pathologists. The histology slides were then reviewed by two pathologists with special interest in GI pathology who were blinded to the initial results. Genta stain was used for the final diagnosis in doubtful cases. Using these consensus results as the gold standard, comparison of the error rate of RUT and general staining was made. The Hpfast RUT results were reevaluated independently by l0 different physicians and nurses. Results: (n = 40) Error rate with 95% Confident interval Group # error Rate SE U95 L95 RUT 16 0.4 0.l 0.9544 0.2614 H&E/Giemsa 4 0.1 0.05 0.2386 0.0307 The greatest discrepancies in the interpretation of Hpfast results were observed at the color change that were seen at the range of pH 5.8-6.5. Conclusions: 1. Inter-observational variability in the interpretation of color changes in Hpfast RUT in the marginal cases could be a factor for the discrepancies seen in routine practice. A single trained observer may avoid the problem. 2. When both tests agree with each other, the accuracy rate is 92%. 3. In doubtful cases with routine staining, Genta stain is required to confirm the diagnosis.