AJG-09-1936-Online Supplementary Appendix
1 of 17
Supplementary Data
TECHNICAL APPENDIX: SUMMARY OF SCIENTIFIC EVIDENCE
PRE-ENDOSCOPIC INDICATORS
QI 1: If patients present with suspected non-variceal upper gastrointestinal hemorrhage (NVUGIH), then they should receive the following tests: complete blood count, levels of electrolytes, blood type, and cross-match at the time of initial evaluation. (Grade Ic)
Routine blood tests allow estimation of severity of bleeding and presence of existing comorbidities as well as preparation for blood transfusion, if needed. Although recommended by guidelines,1,2 there are no direct data to support that obtaining these tests impact patient outcomes. Despite this, the Delphi panelbelieved that the benefit associated with obtaining routine blood tests outweighs the risk by a wide margin and that the lack of data reflect the obviousness of this practice and should not be used as a compelling argument against routine measurement of these tests in a patient with NVUGIH. (Class I, Level C).
QI 2: If patients present with suspected NVUGIH, then they should have a documentation of risk stratification using one of the previously validated measures or the individual components of these measures (e.g., Blatchford or pre-endoscopic Rockall score) at the time of initial evaluation. (Grade Ib)
The Blatchfordscore and pre-endoscopic Rockall score use clinical and laboratory data to identify patients who require intervention,3,4 whereas the complete Rockall score also uses endoscopic variables to predict rebleeding and mortality.4 Published guidelines and authoritative reviews recommend use of these prognostic scales for early stratification of patients into low- and high-risk categories for early intervention, rebleeding, and mortality.1,2, 5 The data supporting the validity of these prognostic scores has been recently published elsewhere.5
Emerging data show that application of these prognostic scores in routine clinical care can accurately identify patients with a low need for early intervention.6,7These selected patients can be safely managed as outpatients, thus shortening the duration of hospitalization. For example, Stanley et al prospectively assessed the impact of introducing Blatchford scoring system at two hospitals in the UK.6 Introduction of the Blatchford scoring system classified 15% patients with upper GI hemorrhage as low risk; these were managed as outpatients without adverse events. The proportion of individuals with low risk GI bleeding admitted to hospital also fell from 96% in the pre- to 71% in the post-Blatchford era (p<0.00001).6 Similarly, Soncini et al reported their experience before and after introduction of Rockall score in routine practice.7 The mean hospital stay became shorter (7.1+/-5.0 vs. 6.3+/-4.5 days), and significantly fewer causes of bleeding remained undefined after the introduction of the Rockall score.7
Collectively, these data not only further support the validity of these prognostic scores, but also show that routine use of these scores can indeed change clinical practice and impact outcomes in NVUGIH. Based on these data, the panel believed that the benefit associated with documentation of risk stratification outweighs the risks by a wide margin. (Class I, Level B).
QI 3: If patients with suspected NVUGIH have normal resting vital signs (e.g., pulse < 100; systolic blood pressure > 100 mmHg), then they should have orthostatic vital signs documented in the records at the time of initial evaluation (Grade Ic)
It is essential to categorize patients with suspected NVUGIH based on the severity of bleeding. One risk factor for severe bleeding is the presence of shock (defined as a pulse rate of more than 100 beats/min and systolic blood pressure less than 100 mm Hg).1 Although we did not find any data that prospectively examined the impact of documenting orthostatic vital signs on clinical outcomes in patients with NVUGIH, an exploratory analysis of RUGBE (Registry on Non-variceal Upper GastrointestinalBleeding and Endoscopy) Canadian cohort study found that presence of orthostatic changes in vital signs correlated highly with the severity of bleeding (unpublished data), while a recent systematic review suggested this initial finding predicts subsequent rebleeding post endoscopic hemostasis.10Finally a recent review of studies that identified criteria for early discharge from hospital also suggested the importance of initial hemodynamic instability not quickly resolved with adequate fluid resuscitation.2Thus, the panelists believed that in patients without clinical signs of shock, it may beimportant to check for orthostatic changes in vital signs to ascertain the severity. (Class I, Level C).
QI 4: If patients withsuspected NVUGIH receive a nasogastric tube, then they should have documentation of the findings in the chart, that include the possibilities of fresh red blood, coffee grounds, bilious, or non bilious aspirate. (Grade Ib)
Nasogastric aspirate is useful in predicting high-risk lesions in patients with NVUGIH. An analysis of the Canadian RUGBE showed a high specificity of a bloody nasogastric aspirate for high-risk lesions on a subsequent EGD with a negative predictive value of 78% (95% CI 73%-82%).8A clear nasogastric aspirate reduced the likelihood to having a high-risk lesion to 15%.8 Although use of a nasogastric tube may identify patients who would benefit from earlier endoscopy, the panel disagreed on whether a nasogastric tube is warranted for routine use in all patients who present with suspected NVUGIH. However, based on the RUGBE data, they agreed that in patients who undergo nasogastric aspiration, the nasogastric aspirate findings (bloody, "coffee ground," clear/bile) should be documented. (Class I, Level B)
QI 5-8: Early resuscitation QI (Grade Ic)
If patients present with suspected NVUGIH, then they receive large bore IV lines at the time of initial evaluation.
If patients with suspected NVUGIH exhibit signs of hypovolemia (e.g., pulse >100, systolic blood pressure <100mmHg; or orthostatic changes), then they should receive crystalloids for fluid resuscitation at the time of initial evaluation.
If patients with suspected NVUGIH have hypoxemia (arterial blood test or an oximeter reading <90%) then they should receive supplemental oxygen at the time of initial evaluation
If patients with suspected NVUGIH have hypovolemia (pulse >100, systolic blood pressure <100 mmHg) that is not responsive to initial fluid resuscitation, then they should be admitted to the intensive care or a monitored unit
Despite advances in medical and endoscopic management, mortality in patients with NVUGIH remains high. Inadequate early resuscitation is likely an important contributor to persistently high morality rates in NVUGIH. In a prospective study including 2 cohorts of patients with clinically significant upper GI hemorrhage(defined as hemodynamic instability; ~90% with NVUGIH), early intensive resuscitation decreased morbidity and mortality and this was likely related to the decreased time interval between admission to achievement of hemodynamic stability.9 Although these data provide direct support to the importance of early resuscitation in improving clinical outcomes in upper GI hemorrhage, the study did not provide results stratified by the common clinical circumstances specified in the resuscitation QIs above.
The importance of early resuscitation is recognized by published clinical guidelines in upper GI hemorrhage.1,5,11,12 The guidelines recommend that intravenous access must be achieved in all patients. The guidelines also specify that for patients who are hemodynamically compromised (pulse >100, systolic blood pressure <100 mmHg)1appropriate crystalloids should be infused to achieve a fall in pulse rate and rise in blood pressure, 1,11,12 followed by admission for close monitoring.1,11 Patients with evidence of severe hypovolemia should be admitted to an intensive care setting.11 (Class I, Level C)
QI 9: If patients with suspected NVUGIH have active hemetemesis with mental status changes, then they should receive airway protection (i.e., intubation) before upper endoscopy. (Grade Ic)
In a retrospective study, outcomes were compared for intensive care unit patients with upper gastrointestinal hemorrhage (UGIH) for 1 year during which prophylactic endotracheal intubation was not performed before endoscopy, with outcomes during a subsequent year in which endotracheal intubation was routine for airway protection before or during EGD when there was hematemesis, altered mentation, unstable cardiopulmonary status, or large amounts of blood in the proximal GI tract.13 Although, there was no difference in the length of stay and mortality, patients in the latter year had no episodes of aspiration during EGD (2.0% vs. 0% in the pre vs. post year; p = 0.21), did not require emergent post-EGD endotracheal intubation (6.0% vs. 0%; p < 0.05), and had fewer in-hospital cardiopulmonary arrests (12.9% vs. 5.0%; p < 0.05). Another recent retrospective analysis included 53 patients who underwent elective prophylactic intubation before EGD and similar number of propensity-matched controls (patients without intubation but with similar probability of receiving intubation).14There were fewer cardiac arrests within 12 hours of EGD in patients who received prophylactic intubation (4 patients in controls vs 1 in cases), but a small sample size precluded finding a statistically significant difference. Overall, there was no difference in the incidence of cardiopulmonary complications, ICU length of stay, hospital length of stay, or hospital mortality between patients who received intubation and propensity-matched controls. The small sample size limited any sub-group analyses (i.e., patients with or without altered mental status).
Given the scarcity of data and based on clinical experience, the panel believed that airway intubation is important for patients with active hemetemesis and altered mentation, but not for the more commonly encountered circumstances, such as hemetemesis without altered mental status or blood in the nasogastric lavage. (Class I, Level C)
ENDOSCOPIC INDICATORS
QI 1: If patients suspected present with NVUGIH and do not have contraindications to EGD, then they should receive an EGD within 24 hours of presentation (Grade Ib)
The performance of an early EGD (within 24 hours of presentation) allows for safe and prompt discharge of patients classified as low risk, improves patient outcomes for patients classified as high risk, and reduces resource utilization for patients classified as either low or high risk.
Several observational studies15-20 and a systematic review21 support the use of early endoscopic stratification for all risk groups, as defined within the first 24 hours following presentation. RCT data have shown resultant decreases in transfusion requirements and length of hospital stay in high-risk patients with a bloody nasogastric tube aspirate but not in those with clear or “coffee grounds” aspirates.22 Studies in all patient-risk groups have demonstrated statistically significant reductions in length of hospital stay17,21-26, as well as significant cost reductions in low risk patients.25,26 Importantly, studies in low-risk patients have shown no major complications in those triaged to outpatient care with early endoscopy.15,16,18,19,21,25-27
Administrative data have found that early endoscopy is associated with a decreased need for surgery in elderly patients without an improvement in mortality.28 Moreover, NVUGIH patients admitted on weekends had higher rates of adjusted in-hospital mortality,29,30surgical interventions,30 while they had a longer mean time to endoscopy and were less likely to undergo early endoscopy within 1 day of hospitalization,29 suggesting a relationship between early endoscopy and mortality. However, in one study, weekend admission remained an independent predictor of increased mortality even after adjusting for the timing of endoscopy.30 Early EGD is associated with significantly shorter hospital stays29,30 and lower hospitalization charges.30 Although not fully reported as of yet, a recent small retrospective analysis of 395 patients treated at a US Veteran’s Administration hospital suggested that the performance of early endoscopy may be associated with a significant 85% reduction in mortality.31
The panelists recognized that there may be justifiable reasons for not undergoing early EGD. For example, EGD may need to be delayed or deferred in selected high-risk patients, such as those with active acute coronary syndrome, suspected perforation, or those in terminal malignancy. The QI explicitly accounted for this and excluded patients who may be too sick to undergo endoscopy. (Class I, Level B).
QI 2: If patients have an ulcer related bleeding on EGD, then the stigmata of bleeding should be documented in the procedure note for the index endoscopy using a standardized taxonomy (e.g., Forrest classification, NIH Consensus Conference taxonomy) (Grade Ia)
Endoscopic predictors of increased risk of rebleeding and mortality include active bleeding, non-bleeding visible vessel, or adherent clot. In a meta-analysis of 30 randomized controlled trials evaluating hemostatic endoscopic treatment in patients with acute NVUGIH, endoscopic treatment decreased rates of further bleeding, surgery, and mortality in patients with high-risk endoscopic features, such as active bleeding or non-bleeding visible vessels.32 These findings were corroborated by a second meta-analysis of 25 trials.33These data show that the identification of the bleeding stigmata directly impacts the selection of treatment modality, risk stratification, and subsequent management in patients with NVUGIH. (Class I, Level A).
QI 3: If patients with NVUGIH undergo an EGD, then a large (single or double) channel endoscope should be used. (Grade IIc)
Experts recommend use of large-channel therapeutic endoscopes in patients undergoing EGD for upper GI bleeding.34 A double channel endoscope (therapeutic endoscope) is particularly useful in patients with more severe bleeding mainly for the purpose of lavage, better visualization, and ability to use the 10-Fr heater probe, although there are no clear data onits advantages. Thus, although the benefit of using large-channel endoscope may be greater than the risk, additional data are needed to confirm this advantage. (Class II, Level C).
QI 4: If patients have an ulcer related bleeding on EGD, then whether hemostasis was achieved or not should be documented in the procedure note for the index EGD (Grade Ic)
Documentation of hemostasis (or failure thereof) results in a change in subsequent management strategy (use of high dose proton pump inhibitor [PPI], length of hospital stay, interventional radiology/surgery consultation).5,12 Therefore, the panel believed that documentation of hemostasis should be a recommended component of care in the patients undergoing endoscopic hemostasis. (Class I, Level C)
QI 5: If patients with ulcer related bleeding have active spurting, oozing, or visible vessel and normal INR (<1.5), then they should receive endoscopic hemostasis using any of the following modalities: hemoclip, thermal devices, combination epinephrine and contact thermal therapy, or combination epinephrine and hemoclip. (GradeIa)
Several recent meta-analyses have better quantified the efficacy of endoscopic therapies in patients with NVUGIH.35-41 Although monotherapy with epinephrine injection is more effective than medical therapy in patients with high-risk stigmata, it is inferior to other monotherapies or to combination therapythat uses 2 or more methods.35-41Numerous meta-analyses indicate that adding a second procedure, such as a second injectate (for example, alcohol, thrombin, or fibrin glue), thermal contact, or clips, is superior to epinephrine injection alone.35-37,39,41
Epinephrineplus a second method for treating high-risk stigmata significantly reduced rebleeding (OR, 0.51 [CI, 0.39 to 0.66]), surgery (OR, 0.63 [CI, 0.45 to 0.89]), and mortality compared with epinephrine monotherapy (OR, 0.50 [CI, 0.30 to 0.82]).40 Monotherapy with thermal devices, sclerosants, clips, thrombin, or fibrin glue provides more effective endoscopic hemostasis than epinephrine alone35 or pharmacotherapy alone.39 Clips were superior to injection monotherapy in 435,37,38,41 of 5 meta-analyses.35,37-39,41 Clips with injection were superior to injection alone but not to clips alone.38,41 Combination therapy (injection plus second injectate, thermal, or clips) was superior to injection therapy alone, but not to clips or thermal therapy alone.37,41Although the data are insufficient to show superiority or equivalence of the recommended treatments, they are strongest for the use of thermal devices, clips, or combination treatments. (Class I, Level A)
QI 6: If patients with ulcer related bleeding have active spurting, oozing, or visible vessel and INR 1.5 to 2.0, then they should receive endoscopic hemostasis using any of the following modalities: hemoclip or combination epinephrine and hemoclip. (Grade Ic)
There is a paucity of prospective data on the effect of different endoscopic hemostasis modalities in patients with coagulopathy. A cohort study in patients who underwent endoscopic treatment found no differences in rebleeding, surgery, mortality, or complication rates between patients receiving warfarin whose INRs were corrected to 1.5 to 2.5 by using fresh frozen plasma, and a control group who did not receive anticoagulants.42 This study suggests that endoscopic treatment with injection or heater probe may be safely performed in patients with an INR less than 2.5.42 In a recently reported large study, endoscopic therapy achieved initial success in 94.7% of patients with INRs between 1.3 and 2.7 by using a variety of hemostatic techniques including injection therapy, heater probe, and hemoclips. However, the rebleeding rate in this series was 23%,and the results were not stratified by the therapeutic modality used for hemostasis.43 Mechanical hemostasis (e.g., hemoclips) may provide therapeutic advantages in patients who must resume anticoagulation after endoscopy, although this has not been rigorously studied.
Although the data are insufficient to confirm superiority or equivalence of different endoscopic treatments in patients with INR 1.5-2.0, the panel members believed that mechanical devices have important advantages and are thus preferable compared to other techniques in these vulnerable patients. Of note, the median rating for the use combination treatment (thermal therapy and epinephrine) was 7 (Appendix Table). However, panelists disagreed on the appropriateness of this practice, with 3 of the 9 rating this QI as “uncertain” (4-6 range). (Class I, Level C)
QI 7: If patients with ulcer related bleeding have clots, then targeted irrigation of the clot should be performed and this should be documented in the procedure note (Grade IC)
Rigorous washing of a clot in an ulcer bed has successfully exposed the underlying stigmata in 26% to 43% of cases,44,45 and the revealed stigmata were high risk in 70% of those cases.45 Theendoscopic findings present after clot removal should be appropriately managed.
The risk for rebleeding with clots that remain adherent after washing without endoscopic therapy (with or without PPI therapy) has been reported to be as low as 0% to 8%,45,46 (94, 95) but also as high as 25% to 35%44,47-49 in clinically high-risk patients. The disparity of these data led to a disagreement among the panelists as to the optimal management of clots that remain adherent after washing. (Class I, Level C)
QI 8: If patients with ulcer related bleeding have a clean-based ulcer or a flat pigmented spot in the ulcer bed, then they should not receive endoscopic hemostasis. (GradeIa)