Area Editoriale
Non-variceal upper-GI bleeding is still a significant problem in the United States despite major advances in endoscopy, pharmacotherapy, and surgery. It remains one of the most common diagnoses that account for admission to the intensive care unit, and the leading reason for urgent consultation with a gastroenterologist. It is now clear that early endoscopy is the most accurate method of determining the cause of bleeding and that endoscopic therapy significantly reduces transfusion requirement, need for urgent surgery, hospital stay, and (probably) mortality from non-variceal upper-GI bleeding. In addition, the findings at endoscopy are a powerful prognostic indicator of ultimate outcome; for example, patients with an ulcer with a clean base have a negligible risk of recurrent bleeding and other adverse outcomes. Given these benefits of endoscopy, it seems intuitively obvious that patients with non-variceal upper-GI bleeding should undergo endoscopy as soon as possible for diagnosis and therapy, and to establish prognosis. There are alternatives, of course, but transfusion and treatment with a proton pump inhibitor or exploratory laparotomy can hardly be justified in most cases.What then is the problem? Is there a problem with the management of non-variceal upper-GI bleeding? The problem I believe is the flow of the patient. Although this is mainly a GI problem that requires GI endoscopy for resolution, the patient first is evaluated by emergency medicine physicians and then cared for by intensivists or hospitalists, so that the gastroenterologist is not brought into the loop for many hours or even days after the onset of bleeding. This delay tends to be magnified in academic institutions, because a prerequisite for a consultation with a gastroenterologist is that the patient first be evaluated by the emergency department resident and then presented to the attending physician, admitted, evaluated by the medical resident team, presented to the ward attending, and then a consultation is requested, all before the patient is seen by the gastroenterologist. Even then, the patient is usually first seen by the resident, then the fellow, and then the attending, before a decision is made, let alone an endoscopy performed. Although some delay is inevitable in every system, delay, nevertheless, increases health care costs and workload, and is inconvenience for the patient. One logical solution to this ingrained system problem of flow is to perform endoscopy early so that an accurate diagnosis is made, the problem is treated, and the patient is triaged to an appropriate level of care in the most efficient, evidence-based manner possible. Bjorkman et al.1 report the results of such an intervention in the current issue of the journal.These investigators from two university medical centers and a Veterans Affairs hospital randomized 93 hemodynamically stable patients with upper-GI bleeding to early endoscopy (within 6 hours of admission) or routine endoscopy (as late as 48 hours after admission). As expected, the two groups were similar with mean Rockall scores of less than 2, (age <80; absence of comorbidity and hypotension). Ulcer was the cause of the bleeding in 61% and 70% of the elective and the urgent endoscopy patients, respectively. Bjorkman et al.1 recommended that 19 of 47 patients (40%) without high-risk lesions after urgent endoscopy be discharged; in actuality, the emergency department physicians discharged only 4 such patients. Consequently, there were no significant differences in final resource utilization for the two groups, because both were hospitalized for about 3 days, regardless of the results of the endoscopy. We are not told why the emergency department physicians did not follow through with the recommendation of the gastroenterologists, but Bjorkman et al.1 speculate that a part of the reason could have been that the study was an effectiveness study in which “endoscopy results are used as part of the decision-marking process by the attending physician, instead of the sole determinant of a therapeutic approach.” The study did not address what was done during the hospitalization for patients with low-risk endoscopic lesions, all of whom were hemodynamically stable, without serious comorbid conditions, able to follow discharge instructions, and relatively young (mean age of 57). Intravenous infusion of a proton pump inhibitor is one possibility; if so, however, this was inappropriate because this form of therapy is only effective in patients with high-risk lesions. Blood transfusion could have been another explanation for admission, but the urgent-endoscopy group had a higher than Hb level at presentation, with only 40% of the total group requiring transfusion. The 19 patients with low-risk lesions on endoscopy did not require endoscopic therapy or surgery, by definition. It seems unlikely that the emergency department physicians did not understand the significance of a highly favorably endoscopy result, inasmuch as the recommendation was that these patients be discharged. Perhaps it was easier and faster for the emergency department to admit instead of discharge the patient and take a chance that the recommendation of the gastroenterologist was wrong. I don't know the reason. However, it is only necessary to look to the inappropriate use of H2-receptor antagonists to understand that physicians do not necessarily practice evidence-based medicine in caring for patients with GI bleeding.2 Bjorkman et al.1 intentionally did not attempt to influence the decision of the emergency department physician to mimic the “typical clinical setting where primary care providers or emergency room physicians are the decision makers and gastroenterologists are consultants who only provide recommendations for care.”We examined a similar strategy of early endoscopy.3 Forty-six percent of the patients in the early endoscopy group in our study met similar criteria for discharge after endoscopy and were discharged, which significantly reduced the cost and the use of resources for patients randomized to early endoscopy. There are several important differences between our study and that of Bjorkman et al.1 that could explain the variation in discharge rates.Our study was carried out in a single center, whereas, the current study was multicenter, which likely increased the variation in physician attitudes, beliefs, and practice styles. Although this may have decreased compliance with an evidence-based approach, experience from a single center is not necessarily generalizable, and, thus, the results of Bjorkman et al.1 may be more realistic. They performed endoscopy within 6 hours, either in the emergency department or the endoscopy unit, whereas we performed endoscopy in the emergency department within 1 to 2 hours of presentation. It is possible that even a small delay in performing endoscopy could reduce the efficiency of a busy emergency department and thereby increase dissatisfaction with the process and, consequently, non-compliance with the recommendations of the gastroenterologists. Finally, the study of Bjorkman et al.1 was conducted by gastroenterologists, whereas our study combined the efforts of gastroenterologists and emergency medicine physicians; as such, the latter group enthusiastically supported the prompt discharge of patients from the emergency department.The results of Bjorkman et al.1 seemingly suggest that it is unrealistic to expect widespread support in clinical practice for the early discharge of patients with non-variceal upper-GI bleeding who are at low risk for adverse outcomes based on endoscopic and other findings. However, the emergency department physicians were unaware of the study hypothesis and, thus, presumably, the rationale and the evidence supporting early discharge. It could, therefore, be concluded that the results of Bjorkman et al.1 are probably the worst-case scenario and that the true clinical results probably lay somewhere between 25% and 100% compliance with early discharge.So what can be concluded from these two studies: (1) early endoscopy identifies slightly less than half of the stable patients with non-variceal upper-GI bleeding who can be discharged safely from the emergency department, (2) early discharge reduces health care costs and resource utilization, and (3) emergency department physicians may or may not accept the recommendation of the gastroenterologist to discharge the patient.As for my question, does it matter if they don't listen to us. It matters. Clearly, it is possible to improve the care of patients with non-variceal upper-GI bleeding. But doing so will take much more than a one-time recommendation. I believe it to be our responsibility to educate our non-gastroenterologist colleagues as to the growing body of evidence supporting early discharge. When this is accomplished, and it must be accomplished, the management of patients with non-variceal upper-GI bleeding is certain to improve.
Leggi l’articolo "