1393
Pilot Trial of Selective Decontamination for Prevention of Bacterial Infection in an Intensive Care Unit John Flaherty, Catherine Nathan, Sherwin A. Kabins, and Robert A. Weinstein
From the Section of Infectious Diseases, Department of Medicine, Michael Reese Hospital and Medical Center, and the University of Chicago, Pritzker School of Medicine, Chicago, Illinois
Infections in intensive care unit (ICU) patients are a major cause of morbidity, mortality, and increased hospital costs. Experience in our cardiac surgery intensive care unit (CSICU) has suggested that because most patients are colonized with potential pathogens before admission [1, 2], traditional infection control measures aimed at limiting cross-infections and controlling environmental sources of infection often are ineffective [3]. A better understanding of the pathogenesis of ICU infections and the role of endogenous flora has led to additional control strategies. For example, use of topical antibiotic prophylaxis for "selective decontamination" of the oral and gastrointestinal reservoir in ICU patients [4-7] aims to eliminate endogenous and nosocomial aerobic gram-negative bacilli, prevent overgrowth of yeast, and reduce risk of pulmonary and extrapulmonary infections. Use of sucralfate, a cytoprotective antiulcer medicine that has little effect on gastric pH, can avoid the antacid- or histamine, blocker-related gastric overgrowth by gramnegative bacteria that occurs with a pH >4 [8]. This approach
Received 12 March 1990; revised 2 July 1990. Presented in part: 28th Interscience Conference on Antimicrobial Agents and Chemotherapy, October 1988, Los Angeles. Written informed consent was obtained from all participants. Financial support: Michael Reese Medical Research Institute Council. Reprints or correspondence: Dr. Robert A. Weinstein, Department ofMedicine, Michael Reese Hospital and Medical Center, Lake Shore Drive at 31st Street, Chicago, IL 60616. The Journal of Infectious Diseases 1990;162:1393-1397 © 1990 by The University of Chicago. All rights reserved. 0022-1899/90/6206-0029$01.00
has been proposed to help lessen risk of tracheal colonization and gram-negative bacillary pneumonia in intubated patients [9, 10]. Although sucralfate has been investigated only in pneumonia prevention, theoretically it could reduce the risk of infection at other sites. Because conventional methods for infection control have not been sufficiently effective in intubated patients, we compared the effects of selective decontamination and sucralfate on patient colonization, pneumonia, and other infections. Our pilot trial was conducted among well-characterized CSICU patients, a relatively homogenous, high-risk group of patients who require relatively brief periods of intubation [1, 2]. We felt this study might help in the design of future controlled tests of selective decontamination in more complex heterogeneous ICU patients who often require prolonged periods of intubation.
Patients and Methods Studypopulation and regimens. During an 8-month period, adult cardiac surgery patients were assigned to receive either sucralfate or selective decontamination. Sucralfate (1 g in 10 ml of H20) was given orally or by nasogastric tube every 6 h. The gastric pH was monitored but no antacids or histamine- blockers were given. Patients assigned to selective decontamination had 1 g of sticky paste (Orabase; Colgate-Hoyt, Canton, MA) containing 2 % polymyxin E, 2 % gentamicin, and 100,000 units of nystatin applied to the buccal mucosa four times daily and a lO-ml suspension of polymyxin E (100 mg), gentamicin (80 mg), and nystatin (2 x 106 units) administered by nasogastric tube or orally four times daily. The selective decontamination patients received routine stress ulcer prophylaxis with antacids or histamines blockers; gastric pH was monitored ev-
Downloaded from http://jid.oxfordjournals.org/ at University of Sussex on August 25, 2015
Selective decontamination of the oropharynx and gastrointestinal tract with nonabsorbable antimicrobials and sucralfate, a stress ulcer prophylactic that maintains the 'normal gastric acid bacterial barrier, were compared for prevention of pneumonia in a cardiac surgery intensive care unit. Over 8 months, 51 patients received selective decontamination and 56 received sucralfate. The selective decontamination regimen included polymyxin, gentamicin, and nystatin given as an oral paste and as a solution; patients also received standard antacid or histamin~ blocker stress ulcer prophylaxis. Patients in the selective decontamination group had significantly less colonization of the oropharynx and stomach by gram-negative bacilli (12% vs. 55%, P< .001), significantly fewer infections due to gram-negative bacilli (6% vs. 20%, P = .02), and fewer infections overall (12% vs. 27%, P = .04). There was one episode of pneumonia in the selective decontamination group and five in the sucralfate group. Mortality and length of stay did not differ between the groups, but those receiving selective decontamination had less than one-third as many days of systemic antibiotic therapy with no increase in colonization or infection with resistant gram-negative bacilli. Thus, selective decontamination appeared to reduce both extrapulmonary and pulmonary infections.
1394
Concise Communications
Results Study population. Of 129 adults undergoing open heart surgery, 107 participated in the study. Of those not participating, 10 refused, 5 were missed, 3 were emergency cases from whom consent could not be obtained, 2 died immediately after operation before starting therapy, and 2 were excluded because of underlying renal failure. Of the 107 participants, 51 received nonabsorbable antibiotics (selective decontamination) and 56 received sucralfate. The two groups were demographically similar (table 1). Colonization. Despite nearly identical colonization rates by gram-negative bacilli on admission to the CSICU (table 2), patients in the selective decontamination group had a significantly lower frequency of colonization while on the study regimen (12 %) than did the sucralfate patients (55 %; P < .001). Initial gastric colonization rates were about half those of oropharyngeal rates; both sites responded similarly in proportion and direction. Heavy or moderate colonization was present in 2 % of patients receiving selective decontamination compared with 27 % of those receiving sucralfate (P
14 days postoperatively) occurred in four patients treated with sucralfate and in one treated by selective decontamination. Only two of eight wound infections involved deep sternal sites, both in the sucralfate group. One superficial sternal wound infection occurred in a patient receiving selective decontamination. The remaining five wound infections in-
Downloaded from http://jid.oxfordjournals.org/ at University of Sussex on August 25, 2015
ery 4-6 h with ColorpHast pH indicator strips (EM Science, Cherry Hill, NJ) to maintain a gastric pH of ~5. The study regimens were started immediately after surgery and continued until transfer from the CSICU. For this pilot study, assignment wasalternated every 2 weeks since complete randomization, placing potentially heavily colonized patients near decontaminated patients, could potentially bias colonization and infection rates through cross-infection. All study patients received routine cefazolin prophylaxis (2 g intravenously) at induction of anesthesia, initiation of cardiopulmonary bypass, and every 4 h intraoperatively. Cefazolin was continued postoperatively (1 g intravenously every 6 h for 48 h) in patients with an intraaortic balloon pump or ventricular-assist device. Patients requiring reoperation were given a single preoperative dose of vancomycin. Cultures. Surveillance swabs of the throat and rectum and aspirates of gastric fluid were obtained for culture on admission to the CSICU. Gastric aspirate cultures were repeated daily for 3 days and then twice weekly until the nasogastric tube was removed. Throat and rectal cultures were repeated in 48-72 h and weekly until discharge from the hospital. All specimens were streaked onto MacConkey agar and selective media [1]. Additional cultures and therapy were ordered by the primary care team as clinically indicated. Demographics, definitions, and statistics. We recorded each patient's demographic and clinical data and prospectively monitored patients for evidence of infection throughout their hospital stay and as outpatients. Pneumonia was defined by the new onset of purulent sputum production (>25 leukocytes/low-power field) and a new persistent infiltrateseen on chest radiographs by a blinded observer. Urinary tract infection was defined as >105 cfu/ml of urine. Wound infection was defined as the presence of pus or cellulitis. Frequency comparisons were analyzed by X2 or Fisher's exact tests as appropriate. For analysis of continuous data, Student's t test was used.
1ID 1990;162 (December)
Concise Communications
110 1990;162 (December)
Table 2. Colonization and infection rates of patients treated with sucralfate or selective decontamination to prevent pneumonia. Study regimen
Patient status Oropharyngeal or gastric colonization by gram-negative bacilli On admission, no. (%) During study, no. (%) p
22 (39) 31 (55) .06
19 (37) 6 (12) .003
NS
Pilot Trial of Selective Decontamination for Prevention of Bacterial Infection in an Intensive Care Unit John Flaherty, Catherine Nathan, Sherwin A. Kabins, and Robert A. Weinstein
From the Section of Infectious Diseases, Department of Medicine, Michael Reese Hospital and Medical Center, and the University of Chicago, Pritzker School of Medicine, Chicago, Illinois
Infections in intensive care unit (ICU) patients are a major cause of morbidity, mortality, and increased hospital costs. Experience in our cardiac surgery intensive care unit (CSICU) has suggested that because most patients are colonized with potential pathogens before admission [1, 2], traditional infection control measures aimed at limiting cross-infections and controlling environmental sources of infection often are ineffective [3]. A better understanding of the pathogenesis of ICU infections and the role of endogenous flora has led to additional control strategies. For example, use of topical antibiotic prophylaxis for "selective decontamination" of the oral and gastrointestinal reservoir in ICU patients [4-7] aims to eliminate endogenous and nosocomial aerobic gram-negative bacilli, prevent overgrowth of yeast, and reduce risk of pulmonary and extrapulmonary infections. Use of sucralfate, a cytoprotective antiulcer medicine that has little effect on gastric pH, can avoid the antacid- or histamine, blocker-related gastric overgrowth by gramnegative bacteria that occurs with a pH >4 [8]. This approach
Received 12 March 1990; revised 2 July 1990. Presented in part: 28th Interscience Conference on Antimicrobial Agents and Chemotherapy, October 1988, Los Angeles. Written informed consent was obtained from all participants. Financial support: Michael Reese Medical Research Institute Council. Reprints or correspondence: Dr. Robert A. Weinstein, Department ofMedicine, Michael Reese Hospital and Medical Center, Lake Shore Drive at 31st Street, Chicago, IL 60616. The Journal of Infectious Diseases 1990;162:1393-1397 © 1990 by The University of Chicago. All rights reserved. 0022-1899/90/6206-0029$01.00
has been proposed to help lessen risk of tracheal colonization and gram-negative bacillary pneumonia in intubated patients [9, 10]. Although sucralfate has been investigated only in pneumonia prevention, theoretically it could reduce the risk of infection at other sites. Because conventional methods for infection control have not been sufficiently effective in intubated patients, we compared the effects of selective decontamination and sucralfate on patient colonization, pneumonia, and other infections. Our pilot trial was conducted among well-characterized CSICU patients, a relatively homogenous, high-risk group of patients who require relatively brief periods of intubation [1, 2]. We felt this study might help in the design of future controlled tests of selective decontamination in more complex heterogeneous ICU patients who often require prolonged periods of intubation.
Patients and Methods Studypopulation and regimens. During an 8-month period, adult cardiac surgery patients were assigned to receive either sucralfate or selective decontamination. Sucralfate (1 g in 10 ml of H20) was given orally or by nasogastric tube every 6 h. The gastric pH was monitored but no antacids or histamine- blockers were given. Patients assigned to selective decontamination had 1 g of sticky paste (Orabase; Colgate-Hoyt, Canton, MA) containing 2 % polymyxin E, 2 % gentamicin, and 100,000 units of nystatin applied to the buccal mucosa four times daily and a lO-ml suspension of polymyxin E (100 mg), gentamicin (80 mg), and nystatin (2 x 106 units) administered by nasogastric tube or orally four times daily. The selective decontamination patients received routine stress ulcer prophylaxis with antacids or histamines blockers; gastric pH was monitored ev-
Downloaded from http://jid.oxfordjournals.org/ at University of Sussex on August 25, 2015
Selective decontamination of the oropharynx and gastrointestinal tract with nonabsorbable antimicrobials and sucralfate, a stress ulcer prophylactic that maintains the 'normal gastric acid bacterial barrier, were compared for prevention of pneumonia in a cardiac surgery intensive care unit. Over 8 months, 51 patients received selective decontamination and 56 received sucralfate. The selective decontamination regimen included polymyxin, gentamicin, and nystatin given as an oral paste and as a solution; patients also received standard antacid or histamin~ blocker stress ulcer prophylaxis. Patients in the selective decontamination group had significantly less colonization of the oropharynx and stomach by gram-negative bacilli (12% vs. 55%, P< .001), significantly fewer infections due to gram-negative bacilli (6% vs. 20%, P = .02), and fewer infections overall (12% vs. 27%, P = .04). There was one episode of pneumonia in the selective decontamination group and five in the sucralfate group. Mortality and length of stay did not differ between the groups, but those receiving selective decontamination had less than one-third as many days of systemic antibiotic therapy with no increase in colonization or infection with resistant gram-negative bacilli. Thus, selective decontamination appeared to reduce both extrapulmonary and pulmonary infections.
1394
Concise Communications
Results Study population. Of 129 adults undergoing open heart surgery, 107 participated in the study. Of those not participating, 10 refused, 5 were missed, 3 were emergency cases from whom consent could not be obtained, 2 died immediately after operation before starting therapy, and 2 were excluded because of underlying renal failure. Of the 107 participants, 51 received nonabsorbable antibiotics (selective decontamination) and 56 received sucralfate. The two groups were demographically similar (table 1). Colonization. Despite nearly identical colonization rates by gram-negative bacilli on admission to the CSICU (table 2), patients in the selective decontamination group had a significantly lower frequency of colonization while on the study regimen (12 %) than did the sucralfate patients (55 %; P < .001). Initial gastric colonization rates were about half those of oropharyngeal rates; both sites responded similarly in proportion and direction. Heavy or moderate colonization was present in 2 % of patients receiving selective decontamination compared with 27 % of those receiving sucralfate (P
14 days postoperatively) occurred in four patients treated with sucralfate and in one treated by selective decontamination. Only two of eight wound infections involved deep sternal sites, both in the sucralfate group. One superficial sternal wound infection occurred in a patient receiving selective decontamination. The remaining five wound infections in-
Downloaded from http://jid.oxfordjournals.org/ at University of Sussex on August 25, 2015
ery 4-6 h with ColorpHast pH indicator strips (EM Science, Cherry Hill, NJ) to maintain a gastric pH of ~5. The study regimens were started immediately after surgery and continued until transfer from the CSICU. For this pilot study, assignment wasalternated every 2 weeks since complete randomization, placing potentially heavily colonized patients near decontaminated patients, could potentially bias colonization and infection rates through cross-infection. All study patients received routine cefazolin prophylaxis (2 g intravenously) at induction of anesthesia, initiation of cardiopulmonary bypass, and every 4 h intraoperatively. Cefazolin was continued postoperatively (1 g intravenously every 6 h for 48 h) in patients with an intraaortic balloon pump or ventricular-assist device. Patients requiring reoperation were given a single preoperative dose of vancomycin. Cultures. Surveillance swabs of the throat and rectum and aspirates of gastric fluid were obtained for culture on admission to the CSICU. Gastric aspirate cultures were repeated daily for 3 days and then twice weekly until the nasogastric tube was removed. Throat and rectal cultures were repeated in 48-72 h and weekly until discharge from the hospital. All specimens were streaked onto MacConkey agar and selective media [1]. Additional cultures and therapy were ordered by the primary care team as clinically indicated. Demographics, definitions, and statistics. We recorded each patient's demographic and clinical data and prospectively monitored patients for evidence of infection throughout their hospital stay and as outpatients. Pneumonia was defined by the new onset of purulent sputum production (>25 leukocytes/low-power field) and a new persistent infiltrateseen on chest radiographs by a blinded observer. Urinary tract infection was defined as >105 cfu/ml of urine. Wound infection was defined as the presence of pus or cellulitis. Frequency comparisons were analyzed by X2 or Fisher's exact tests as appropriate. For analysis of continuous data, Student's t test was used.
1ID 1990;162 (December)
Concise Communications
110 1990;162 (December)
Table 2. Colonization and infection rates of patients treated with sucralfate or selective decontamination to prevent pneumonia. Study regimen
Patient status Oropharyngeal or gastric colonization by gram-negative bacilli On admission, no. (%) During study, no. (%) p
22 (39) 31 (55) .06
19 (37) 6 (12) .003
NS
