|
|
| |||||||||||||||||||||||||||||||||||||||||
Background Critically ill patients who require mechanical ventilation are at increased risk for gastrointestinal bleeding from stress ulcers. There are conflicting data on the effect of histamine H2-receptor antagonists and the cytoprotective agent sucralfate on rates of gastrointestinal bleeding, ventilator-associated pneumonia, and mortality.
Methods In a multicenter, randomized, blinded, placebo-controlled trial, we compared sucralfate with the H2-receptor antagonist ranitidine for the prevention of upper gastrointestinal bleeding in 1200 patients who required mechanical ventilation. Patients received either nasogastric sucralfate suspension (1 g every six hours) and an intravenous placebo or intravenous ranitidine (50 mg every eight hours) and a nasogastric placebo.
Results The patients in the two groups had similar base-line characteristics. Clinically important gastrointestinal bleeding developed in 10 of 596 (1.7 percent) of the patients receiving ranitidine, as compared with 23 of 604 (3.8 percent) of those receiving sucralfate (relative risk, 0.44; 95 percent confidence interval, 0.21 to 0.92; P = 0.02). In the ranitidine group, 114 of 596 patients (19.1 percent) had ventilator-associated pneumonia, as compared with 98 of 604 (16.2 percent) in the sucralfate group (relative risk, 1.18; 95 percent confidence interval, 0.92 to 1.51; P = 0.19). There was no significant difference between the groups in mortality in the intensive care unit (ICU) (23.5 percent in the ranitidine group and 22.8 percent in the sucralfate group) or the duration of the stay in the ICU (median, nine days in both groups).
Conclusions Among critically ill patients requiring mechanical ventilation, those receiving ranitidine had a significantly lower rate of clinically important gastrointestinal bleeding than those treated with sucralfate. There were no significant differences in the rates of ventilator-associated pneumonia, the duration of the stay in the ICU, or mortality.
Randomized trials of prophylaxis against stress ulcers, as compared with no prophylaxis, indicate that histamine H2-receptor antagonists and antacids prevent clinically important gastrointestinal bleeding.12 Observational studies have suggested, however, that a higher gastric pH is associated with gastric microbial growth,13 tracheobronchial colonization,14 and nosocomial pneumonia.15 In randomized trials, the cytoprotective agent sucralfate, which does not alter the gastric pH, has been associated with a trend toward a lower incidence of ventilator-associated pneumonia, both as compared with histamine H2-receptor antagonists and antacids16 and as compared with H2-receptor antagonists alone.17,18,19,20,21,22,23,24,25,26,27,28,29,30
The need to evaluate patients at highest risk for both bleeding and pneumonia31 and the need to blind care givers and research personnel to the treatment assignments, in order to minimize inflated treatment effects,32 prompted us to evaluate the rates of clinically important gastrointestinal bleeding, ventilator-associated pneumonia, and mortality in a randomized trial of 1200 patients who required mechanical ventilation and who were assigned to receive either ranitidine or sucralfate.
Methods
Enrollment
From October 1992 to May 1996, we screened consecutive patients admitted to 16 participating intensive care units (ICUs) to identify adults who were projected to require mechanical ventilation for at least 48 hours. Criteria for exclusion were a diagnosis of gastrointestinal bleeding or pneumonia on admission, gastrectomy, a prognosis considered to be hopeless, previous randomization in this or another trial, or receipt of two or more previous doses of open-label prophylactic therapy. The protocol was approved by the institutional review boards of all the participating centers, and the patients or their proxies gave informed consent.
Randomization
Patients were randomly assigned to study groups in blocks of six, with stratification according to center, by means of a computer-generated random-number table prepared at the McMaster University Methods Center and managed by the ICU study pharmacist at each site who administered the coded drugs. All care givers and other research personnel were unaware of the randomization schedule and the block size.
Blinding
The patients, research nurses, and all ICU care givers were unaware of the treatment assignments for the duration of the study. Therefore, clinicians did not monitor gastric pH. The radiologists, outcome adjudicators, all investigators, and the study statistician were also blinded until all events had been adjudicated and the analyses completed.
Drug Preparation, Dispensing, and Administration
The bags of ranitidine (Zantac, Glaxo Wellcome; prepared by Baxter) and ranitidine placebo were identical in appearance. The sucralfate (Sulcrate, Hoechst Marion Roussel) and sucralfate placebo slurries were identical in color, taste, and consistency. From coded ranitidine bags and sucralfate bottles stored in each ICU pharmacy, study pharmacists dispensed either active ranitidine (50 mg every eight hours) and sucralfate placebo or active sucralfate (1 g every six hours) and ranitidine placebo.
Ranitidine was administered in intravenous bolus form, with the dose adjusted for renal failure as follows: standard dose, 50 mg every 8 hours; dose for patients with an estimated creatinine clearance rate of 25 to 50 ml per minute, 50 mg every 12 hours; dose for patients with an estimated creatinine clearance rate below 25 ml per minute, 50 mg every 24 hours; and dose for patients dependent on dialysis, 50 mg every 12 hours. Sucralfate suspension was given through a nasogastric tube or orally. We followed all patients until they died or were discharged from the ICU.
Demographic Characteristics of the Patients
We documented each patient's age, sex, admitting diagnosis, location before admission to the ICU, score on the Acute Physiology and Chronic Health Evaluation (APACHE II) scale (range of scores, 0 to 71, with higher scores indicating more severe illness),33 and multiple-organ-dysfunction (MOD) score (range, 0 to 24, with higher scores indicating more severe organ dysfunction).34
Gastrointestinal Bleeding
We monitored patients for signs of overt gastrointestinal hemorrhage, including hematemesis, nasogastric aspirate containing blood or coffee-grounds material, melena, or hematochezia. The decision to perform endoscopy was left to the discretion of the intensive care specialist. Two physicians from the study's bleeding-adjudication committee examined all relevant clinical and diagnostic documents related to possible bleeding episodes in duplicate and independently, using previously defined criteria.35 We defined clinically important bleeding as overt bleeding plus one of the following four features, in the absence of other causes: a spontaneous drop of 20 mm Hg or more in the systolic or diastolic blood pressure within 24 hours after upper gastrointestinal bleeding; an increase in the pulse rate of 20 beats per minute and a decrease in the systolic blood pressure of 10 mm Hg on the patient's assuming an upright position; a decrease in the hemoglobin concentration of at least 2 g per deciliter in 24 hours and the transfusion of 2 units of packed red cells within 24 hours after bleeding; or failure of the hemoglobin concentration (in grams per deciliter) to increase after transfusion by at least the number of units transfused minus 2 (i.e., if 4 units of packed cells were transfused, the bleeding would be considered clinically important if the hemoglobin concentration did not rise by at least 2 g per deciliter).
Ventilator-Associated Pneumonia
Attending intensivists at each center used a modified version of the criteria of the Centers for Disease Control and Prevention (CDC)36 to identify patients in whom pneumonia was suspected on clinical grounds. These criteria were a new radiographic infiltrate that had persisted for at least 48 hours (as interpreted by designated study radiologists blinded to the patients' treatment assignments) plus at least two of the following: a temperature above 38.5°C or below 35.0°C, a leukocyte count of more than 10,000 per cubic millimeter or less than 3000 per cubic millimeter, purulent sputum, or isolation of pathogenic bacteria from an endotracheal aspirate.
Patients in whom ventilator-associated pneumonia was suspected on clinical grounds underwent bronchoalveolar lavage or protected brush-catheter sampling by the study bronchoscopist.37,38 Two members of the pneumonia-adjudication committee examined all relevant clinical and diagnostic documents related to possible cases of pneumonia in duplicate and independently, classifying patients according to several different methods. One method used was the modified CDC criteria described above36; another was the Clinical Pulmonary Infection Score devised by Pugin et al. (range, 0 to 12, with pneumonia defined by a score of 7 or higher).39 In addition, adjudicators used the criteria of the Memphis Ventilator-Associated Pneumonia Consensus Conference for definite ventilator-associated pneumonia (if there was radiographic evidence of abscess and a positive needle aspirate, or if there was histologic proof of pneumonia at biopsy or autopsy) and probable ventilator-associated pneumonia (if bronchoalveolar lavage or protected brush-catheter sampling yielded positive quantitative or semiquantitative cultures, if there was a positive blood culture of an organism found within 48 hours of isolation in the sputum, if there was a positive pleural-fluid culture of an organism found within 48 hours of isolation in the sputum, or if histologic examination showed formation of an abscess or consolidation with polymorphonuclear-cell infiltration).40 To make a final decision as to whether each patient had ventilator-associated pneumonia, adjudicators made a summary judgment based on all available information; disagreement was resolved through discussion. We determined a priori that the adjudication committee's consensus rate of pneumonia would be used for the primary analysis and that other definitions would be used in the secondary analyses.
Statistical Analysis
Since ventilator-associated pneumonia has been considered an important potential adverse effect of gastric pH–altering prophylaxis against stress ulcers and because previous trials did not examine pneumonia as rigorously as gastrointestinal bleeding, we designed our trial to have the statistical power to detect a difference in the rates of pneumonia. On the basis of data published through 1991, when our study was designed, we anticipated a 25 percent incidence of pneumonia and identified a 25 percent reduction in the risk of pneumonia associated with sucralfate as being plausible and clinically important. This led to the calculation of a sample size of 1200 patients as necessary to give the study 75 percent power to detect such a difference, assuming a two-sided significance test at the 0.05 level. We analyzed all patients in the groups to which they were randomly assigned, according to the intention-to-treat principle. We compared means using Student's t-test and compared proportions using the chi-square test, with two-tailed P values.41 We used Fisher's exact test when the number of data points was small. For the length of the stay in the ICU, we report data as medians and interquartile ranges.
We conducted analyses both with and without adjustment for the following variables, treated categorically: age (<65 or >65 years), sex, location before admission to the ICU (emergency room, operating room, hospital ward, or other ICU), medical or surgical (elective or emergency) status, APACHE II score (<25 or >25), MOD score (<5 or >5), and medical center (of which there were 16).42
Results
Enrollment
Of 7986 patients admitted to ICUs, 6786 were excluded for the following reasons: mechanical ventilation was expected to be needed for less than 48 hours (3754 patients), the prognosis was considered hopeless (481), the patient had pneumonia (988) or gastrointestinal bleeding (432) at admission, gastrectomy had been performed (41), the patient had previously undergone randomization in this or another trial (565), two or more previous doses of open-label prophylaxis had been received (230), the patient had undergone mechanical ventilation in another ICU (38), the patient or proxy was unable or unwilling to give informed consent (184), an administrative error occurred (44), the physician declined to have the patient enrolled (14), or another reason (15). Excluded patients were a mean (±SD) of 60.8±15.3 years of age and had a mean APACHE II score of 21.4±7.9.
We randomly assigned 1200 patients to the sucralfate or ranitidine group. Demographic and base-line physiologic characteristics were similar in the two groups (Table 1).
|
No patient received active drug instead of the assigned placebo, or vice versa. Of the scheduled doses of ranitidine and sucralfate, 94.2 percent and 91.7 percent, respectively, were administered. Among patients who missed doses, the mean number of doses missed was 2.3 (median, 3; interquartile range, 2 to 3) for ranitidine and 2.9 (median, 4; interquartile range, 1 to 4) for sucralfate.
Fourteen patients in the ranitidine group (2.3 percent) and 16 in the sucralfate group (2.6 percent) received an additional drug as prophylaxis against stress ulcers outside of the study protocol (P = 0.74). Most patients received enteral nutrition (70.3 percent and 71.8 percent, respectively; P =0.55). Enteral feeding was started a median of three days (interquartile range, two to four) after admission to the ICU.
Gastrointestinal Bleeding
In the ranitidine group, 10 of 596 patients (1.7 percent) had clinically important gastrointestinal bleeding, as compared with 23 of 604 (3.8 percent) in the sucralfate group (relative risk, 0.44; 95 percent confidence interval, 0.21 to 0.92; P = 0.02); the absolute reduction in the risk of bleeding was 2.1 percent (95 percent confidence interval, 0.29 to 3.97). The results of the analysis in which we adjusted for other factors were similar (relative risk, 0.45; 95 percent confidence interval, 0.22 to 0.92; P = 0.03).
Seventeen patients underwent upper gastrointestinal endoscopy; additional diagnostic tests included laparotomy (in four), angiography (three), sigmoidoscopy (three), and red-cell scanning (one); two patients were examined at autopsy. One patient in the sucralfate group had bleeding from a gastroduodenal anastomosis, and another in the same group had bleeding from an aortic graft–enteric fistula. The source of bleeding was unclear for 8 patients in the ranitidine group and for 11 patients in the sucralfate group. Table 2 lists the 14 patients who had bleeding from gastric, esophageal, or duodenal erosions or discrete ulcers along with their study group, indication of bleeding, and diagnostic tests or procedures (Table 2). Post hoc subgroup analysis of patients with documented esophageal, gastric, or duodenal ulcers or erosions generated results consistent with our main findings (relative risk of bleeding in the ranitidine group as compared with the sucralfate group, 0.41; 95 percent confidence interval, 0.13 to 1.29; P = 0.22).
|
Among 596 patients receiving ranitidine, 114 (19.1 percent) had ventilator-associated pneumonia, as compared with 98 of 604 (16.2 percent) in the sucralfate group (relative risk, 1.18; 95 percent confidence interval, 0.92 to 1.51; P = 0.19); the absolute difference in risk was 2.9 percent (95 percent confidence interval, -1.4 to 7.2). Results of the adjusted analysis were similar (relative risk, 1.14; 95 percent confidence interval, 0.91 to 1.44; P = 0.26).
Table 3 presents the rates of pneumonia when different diagnostic criteria were used. As the definition of pneumonia became more strict, the overall rate decreased, although the difference between the two groups persisted. Table 4 shows the organisms isolated from endotracheal aspirates from patients with pneumonia. There was no significant difference in the types of isolates between the groups.
|
|
All patients were included in the analysis. The median length of the ICU stay was 9 days (interquartile range, 5 to 15) in the ranitidine group and 9 days (interquartile range, 5 to 17) in the sucralfate group (P = 0.27). The duration of intubation was 7 days (interquartile range, 4 to 13) in the ranitidine group and 8 days (interquartile range, 4 to 15) in the sucralfate group (P = 0.25).
Mortality in the ICU
Mortality in the ICU was similar among patients receiving ranitidine (140 of 596 [23.5 percent]) and those receiving sucralfate (138 of 604 [22.8 percent]; relative risk, 1.03; 95 percent confidence interval, 0.84 to 1.26; P = 0.79).
Discussion
We found that patients receiving ranitidine had a significantly lower risk of gastrointestinal bleeding than patients receiving sucralfate (relative risk, 0.44; 95 percent confidence interval, 0.21 to 0.92). This finding appears to contrast with the results of a recent meta-analysis, which suggested that the drugs' effect on bleeding was equivalent (relative risk, 0.95;95 percent confidence interval, 0.17 to 5.36)12; however, the confidence limits around these two estimates overlap widely, and the difference could be attributable to chance. We found no significant difference in the incidence of pneumonia between patients receiving an H2-receptor antagonist and those receiving sucralfate (relative risk, 1.18; 95 percent confidence interval, 0.92 to 1.51), a finding that is consistent with pooled data from previous randomized trials (relative risk, 1.19; 95 percent confidence interval, 0.98 to 1.44).12
The strengths of our trial include the measures taken to conceal the patients' treatment assignments; the blinding of care givers, research personnel, and analysts; the high rates of compliance; the adjudication of outcomes according to rigorous criteria; and the examination of the relation between prophylaxis and the incidence of ventilator-associated pneumonia defined according to a variety of criteria, given the absence of a well-accepted reference standard. Our study was limited by uncertainty about the causes of gastrointestinal bleeding; even in patients whose condition was stable enough for them to undergo endoscopy or other diagnostic tests, a site of bleeding could not always be identified. The source of bleeding remained unknown for 19 of 33 patients. Moreover, although our sample was large, the confidence intervals around key estimates were wide.
Previous randomized trials indicate that H2-receptor antagonists substantially lower the incidence of overt and clinically important bleeding as compared with no prophylaxis.12 Our findings show that rates of bleeding are lower among patients given H2-receptor antagonists than among those given sucralfate. The incidence of bleeding in the sucralfate group in this trial (3.8 percent) approximates the 3.7 percent incidence observed among untreated patients receiving mechanical ventilation in our natural-history study.11 Previous trials comparing sucralfate with no prophylaxis have found a significantly lower rate of overt bleeding with sucralfate23,29,43,44; however, in the single trial from which data on clinically important bleeding events could easily be extracted,29 1 of 30 patients in the control group and 1 of 24 patients in the sucralfate group had clinically important bleeding. Given all this evidence, it is plausible that sucralfate has no effect on clinically important bleeding.
Although we found no significant difference in the rates of pneumonia between the two groups, the relative risk suggests a trend toward a lower rate of pneumonia among patients receiving sucralfate. It is possible that sucralfate appears to have a small protective effect against pneumonia because stress-ulcer prophylactic medications that increase the gastric pH themselves increase the incidence of pneumonia. This contention is supported by direct comparisons of trials of H2-receptor antagonists with no prophylaxis, which show a trend toward higher rates of pneumonia among the patients receiving H2-receptor antagonists (odds ratio, 1.25; 95 percent confidence interval, 0.78 to 2.00).12 Furthermore, the relative effects of sucralfate and no prophylaxis against stress ulcers on the incidence of pneumonia is unclear. Among the 226 patients enrolled in two randomized trials, there was a trend toward a higher incidence of pneumonia among those receiving sucralfate than among those given no prophylaxis (odds ratio, 2.11; 95 percent confidence interval, 0.82 to 5.44).23,44
Previous cost-effectiveness studies and practice guidelines have appropriately considered not only the base-line risk of bleeding and the effectiveness of prophylaxis against stress ulcers, but also the clinical and economic consequences of different preventive approaches.45,46,47 The direct and indirect costs of medicines for prophylaxis vary among health care organizations — a fact that highlights the need to appraise economic evaluations critically to determine their applicability to specific settings. The incremental number of days in the ICU attributable to clinically important upper gastrointestinal bleeding is estimated to be seven, and the attributable risk of mortality is about 12.5 percent.48 The trend toward a lower rate of nosocomial pneumonia among patients receiving sucralfate rather than H2-receptor antagonists12 has been considered in some policy documents about the prophylactic treatment of stress ulcers. The development of ventilator-associated pneumonia may lead to an additional 13 days in the ICU,49 although the effect of this type of pneumonia on mortality is less certain.49,50
The patients at highest risk for clinically important bleeding are those who require mechanical ventilation for more than 48 hours and those with coagulopathy.8,9,10,11 The estimated number of critically ill patients undergoing ventilation who would need to receive prophylaxis with ranitidine rather than sucralfate to prevent one clinically important upper gastrointestinal hemorrhage is approximately 48.
Supported by the Medical Research Council of Canada and Hoechst Marion Roussel. Drugs were supplied by Glaxo Wellcome, Baxter, and Hoechst Marion Roussel. Dr. Cook is a Career Scientist of the Ontario Ministry of Health.
We are indebted to the members of the Canadian Critical Care Trials Group for their help, particularly the research nurses who participated in this study, and to Drs. Thomas Todd, Thomas Noseworthy, Timothy Winton, and Michael Tryba for their support and advice.
Source Information
From McMaster University, Hamilton, Ont. (D.C., G.G., H.F., L.G., A.M.); the University of Toronto, Toronto (J.M.); the University of Western Ontario, London (D.L., F.R., A.K.); Dalhousie University, Halifax, N.S. (R.H.); Memorial University, St. John's, Newf. (S.P.); and Queen's University, Kingston, Ont. (G.W.) — all in Canada. Other authors were Martin Tweeddale, M.D., University of British Columbia, Vancouver; Joe Pagliarello, M.D., University of Ottawa, Ottawa, Ont.; and Richard Johnston, M.D., University of Alberta, Edmonton.
Address reprint requests to Dr. Cook at the Department of Medicine, St. Joseph's Hospital, 50 Charlton Ave., Hamilton, ON L8N 4A6, Canada.
References
In addition to the authors, the study investigators and institutions were as follows: B. Plumstead and L. Frighetto, Vancouver General Hospital, Vancouver, B.C.; T. Noseworthy, E. Konopad, S. Bishop, M.L. Derko, and K. Horon, Royal Alexandra Hospital, Edmonton, Alta.; D. Stollery, M. Goers, and R. Jarman, Grey Nun's Hospital, Edmonton, Alta.; D. Roberts, T. Ostrusniuk, and J. Studney, Winnipeg Health Sciences Center, Winnipeg, Man.; T. Winton, D. Foster, D. Baptiste, M. Steinberg, and M. Lee, Toronto Hospital, General Division, Toronto; G. Darling, M. Culham, and V. Bobiwash, Wellesley Hospital, Toronto; M. Lefcoe (study radiologist), D. McCormack (study bronchoscopist), L. McCarthy, and C. Gawlik, London Health Sciences Center (Victoria Campus), London, Ont.; S. Langdon, M. Johnson, and C. Charters, London Health Sciences Center (University Campus), London, Ont.; M.K. Scott and S. Jansen, St. Joseph's Health Center, London, Ont.; E. McDonald, T. Dinh, and S. Toner, St. Joseph's Hospital, Hamilton, Ont.; S. Salama, Henderson Hospital, Hamilton, Ont.; A. Taite and P. Newman, Kingston General Hospital, Kingston, Ont.; M. Loewen, Ottawa Civic Hospital, Ottawa, Ont.; D. Gibbons, G. Leaman, and G. Gibbons, Health Sciences Center, St. John's, Newf.; P. Roy, M. Coffin, and H. Lummis, Victoria General Hospital, Halifax, N.S.; Methods Center — P. Austin and S. Troyan (project coordinators), L. Buckingham (data-base manager), and S. Duchesne (data entry); Adjudication Committees — D. Heyland, A. Freitag, K. Gough, M. Meade, A. Sarabia, M. Turner, H. Devitt, B. Guslits, M. Heule, R. Jaeschke, and J. Lang.
| |||||||||||||||||||||||||||||||||||||||||
Related Letters:
Prevention of Gastrointestinal Bleeding during Mechanical Ventilation
Mitchell I., Bihari D., Wyncoll D. L.A., Marik P., Varon J., Offenstadt G., Guidet B., Maury E., Cook D., Guyatt G., Leasa D., The Canadian Critical Care Trials Group
Extract |
Full Text
N Engl J Med 1998;
339:266-268, Jul 23, 1998.
Correspondence
This article has been cited by other articles:
HOME | SUBSCRIBE | SEARCH | CURRENT ISSUE | PAST ISSUES | COLLECTIONS | PRIVACY | HELP | beta.nejm.org Comments and questions? Please contact us. The New England Journal of Medicine is owned, published, and copyrighted © 2008 Massachusetts Medical Society. All rights reserved. |
Previous
