5

Double-blind study of selective decontamination of the digestive tract in intensive care

Selective decontamination of the digestive tract (SDD), by means of non-absorbable antibiotics, to prevent infection in intensive-care units (ICUs) remains controversial; there is evidence that the regimen reduces the incidence of secondary infection, but no convincing reduction in morbidity or mortality has been shown and the costs and effect on microbial resistance patterns need further study. In a double-blind, placebo-controlled trial, we have tried to find out whether SDD should be used routinely in all ICU patients at high risk of secondary infection. All patients admitted to the ICU who were thought likely to stay in the unit for at least 5 days and to need intubation for longer than 48 h were enrolled and randomly allocated to groups receiving placebo or SDD (amphotericin, colistin, and tobramycin applied to the oropharynx and enterally); all patients received intravenous cefotaxime for 72 h. Of 322 patients randomised, 83 were withdrawn (80 ICU stay or duration of intubation too short, 3 protocol violations). 239 medical, trauma, and surgical patients completed the trial period (114 SDD, 125 placebo). There were no differences between SDD and placebo groups in incidence of infection (30 [26%] vs 43 [34%] patients; p=0·22), duration of ICU stay (mean 16·2 [14·3] vs 16·8 [12·3] days), hospital stay (29·9 [SD 25·0] vs 31·9 [22·2] days), or mortality (21 [18%] vs 21 [17%]). SDD increased the costs of intensive care. substantially Mechanisms other than bacterial colonisation of the gut may bring about substantial numbers of secondary infections in ICUs. Routine use of SDD in multidisciplinary ICUs cannot be recommended. Introduction Nosocomial infection contributes substantially to morbidity and mortality in the intensive-care unit (ICU).l The overall incidence of infection varies among units, depending on the environment, nursing care, and types of patients admitted.2 The most common site of infection is the lower respiratory tract and the commonest causative organisms are multiresistant gram-negative bacteria. 1,3-5 Most of these infections are thought to be endogenous and secondary to the aspiration of oropharyngeal secretions that have become colonised by resistant organisms from the hospital environment.3,6 Selective decontamination of the digestive tract (SDD), by which the anaerobic bacteria are preserved and potentially pathogenic aerobic gram-negative bacteria are eliminated from the gastrointestinal tract with nonabsorbable enterally administered antibiotics, has been studied widely;7 however, because of design flaws, including

of historical control groups and small numbers of patients for analysis, the results are inconclusive. Nevertheless, SDD has been adopted by many units; although the procedure reduces secondary infections, especially of the respiratory tract, there is no convincing evidence of a reduction in morbidity, mortality, or the overall cost of intensive care. Our study of SDD was a randomised, double-blind, placebo-controlled trial in patients who had been intubated for more than 48 h and had been treated in a multidisciplinary ICU for more than 5 days; we aimed to find out whether SDD should be used routinely in patients at high risk of secondary infection. the

use

Patients and methods The trial included all

patients admitted to the ICU who were require endotracheal intubation for longer than 48 h and to stay in the ICU for at least 5 days; this assessment was made by the attending physicians using guidelines from previous analysis of admissions to our unit, which showed that patients with asthma, drug overdoses, and those admitted electively after operations were unlikely to meet the enrolment criteria. Patients not enrolled but who subsequently proved suitable were included in a separate analysis to seek selection bias. Only patients with a history of hypersensitivity to the study drugs or of anaphylaxis to penicillin were excluded. Consent was obtained from the patient, a close relative, or the responsible physician (who was not taking part in the study). The study was approved by the ethics committee of the University of Cape Town. On admission, patients were randomly assigned by means of computer-generated numbers to groups receiving placebo or the SDD regimen. The regimen consisted of 2 ml methylcellulose gel containing amphotericin 2%, tobramycin 2%, and colistin 2% applied to the oral mucosa every 6 h with a gloved finger (or indistinguishable placebo), and 10 ml of a solution of amphotericin 500 mg, tobramycin 80 mg, and colistin 100 mg by mouth or by nasogastric tube (or 10 ml saline). SDD was continued for 48 h after

thought likely

to

extubation. Cefotaxime 1 g every 8 h was given intravenously for 72 h to SDD and placebo patients. Patients already receiving an antibiotic for a pre-existing infection on admission to the ICU continued on the antibiotic if the organism was known to be sensitive or if the patient seemed to be responding; otherwise, if appropriate, cefotaxime was substituted. We did not use anticytokine therapy. No routine prophylaxis against gastrointestinal bleeding was given; ranitidine was used only in patients with pre-existing ulcer disease or active gastrointestinal haemorrhage and in those at high risk (from burns, head injury, or liver failure). Sucralfate was not used because of concern that it might bind to the active enteral agents.8 Patients were fully examined on entry, an admission diagnosis was defmed, and the severity of illness was assessed on the APACHE II,9 organ failure and injury severity" scoring systems. Factors likely to affect the development of secondary infection were noted.1,12 In all intubated patients nasogastric tubes were placed for ADDRESSES: Respiratory Intensive Care Unit, Departments of Medicine and Anaesthesia (J. M. J. Hammond, MB, P. D. Potgieter, MB), and Department of Medical Microbiology (G. L Saunders, MB, Prof A. A. Forder, MB), University of Cape Town and Groote Schuur Hospital, Cape Town, South Africa. Correspondence to Dr Janet M. J. Hammond, Respiratory ICU, Groote Schuur Hospital, Observatory 7925, Cape Town, South Africa.

6

TABLE III-EFFECT OF SDD ON MORBIDITY AND MORTALITY

TABLE I-BASELINE CHARACTERISTICS I

I

I

I

we did not require positive catheter tip culture or blood culture or the presence of local inflammation for diagnosis. If a nosocomial infection was diagnosed, appropriate cultures were taken for isolation, identification, and sensitivity testing of the causative organisms. All central-venous catheters were cultured on removal. Culture results were available to the clinicians. Aerobic cultures of oropharyngeal swabs, gastric aspirates, faecally soiled rectal swabs, urine, tracheal aspirates, wound swabs, or other secretions were done on admission and then twice a week throughout the ICU stay; final specimens were taken 3 days after discharge from the ICU. To avoid carryover of topical antibiotics from the SDD regimen, specimens were plated out, then placed in 10 ml broth overnight to dilute any antibiotics present. Growth density was graded 1-5 by a semiquantitative method.13 Enterococci, Acinetobacter spp, Pseudomonas spp, Klebsiella spp, Staphylococcus aureus, S epidermidis, and yeasts were identified and antibiotic sensitivity patterns tested by standard disc-diffusion methods. Enterobacteriaceae were identified to species. The clinical cause of death was characterised as unrelated, possibly related, or definitely related to nosocomial infection; when possible these relations were confirmed at necropsy. The cost of SDD and the additional costs of investigation and treatment of any episodes of nosocomial infection were recorded. Statistical analysis was done by the chi-square test, Fisher’s exact test, Student’s t test for continuous variables, and 2 x2 tables, where appropriate.

catheter;

*By Knaus’ criteria.’0

the duration of ventilation. Routine investigations included daily chest radiography during ventilation, laboratory evaluation of haematology, and serum and urine chemistry; other studies were done as necessary. Microbiological studies were done twice a week or more often if indicated. Daily monitoring included looking for clinical evidence of secondary infection, and careful note was kept of complications developing in ICU. Follow-up was continued until discharge from hospital. Nosocomial infection was diagnosed if signs developed more than 48 h after admission. Septicaemia was defined as clinical observation of all the following with no obvious site of infection: a rise or fall in temperature, a rise in white-blood-cell count or a left shift in morphology, evidence of organ dysfunction (lung, kidney, liver, or central nervous system), and a positive blood culture. Nosocomial pneumonia was defined as a new infiltrate on the chest radiograph more than 48 h after admission to the ICU, purulent bronchial secretions with many leucocytes, temperature above 38°C, white-blood-cell count above 1010/1, increasing, or showing a left shift, substantial numbers of organisms shown by gram-staining of tracheal aspirate, pure growth cultured from tracheal aspirate, and deterioration in gas exchange of 2 kPa or more. Bronchial infection was defined by the presence of all the features of pneumonia except the radiographic changes. Urinary-tract infection in patients who had been catheterised for at least 48 h before collection of the specimen was defmed by the presence of local or systemic signs of infection with culture of bacteria or yeasts and the presence of more than 108 white cells per litre urine. Vascular-catheter-related sepsis was defined as temperature above 38°C, white-blood-cell count above 101°/1 or left shift, with relief of signs on removal of the TABLE t!—)NVAS!VE PROCEDURES USED IN ICU

Results

January, 1989, and December, 1990, 670 ICU. 322 met criteria for study patients on admission; however, 83 were subsequently entry withdrawn, 3 for protocol violations (1 patient refused to continue with the enteral medication,1 had a suspected drug allergy though he was later found to have received placebo, and 1 patient was given sucralfate), and 80 because their stay Between

were admitted to the

in the ICU or duration of intubation was too short. 28 of those excluded, none of whom had infection, died within 5 days of admission. There were also 36 patients who were not initially included, but who were subsequently found to have satisfied the study criteria; their data were analysed

separately. The various groups were similar in terms of baseline characteristics (table i). The SDD and placebo groups did not differ in distributions of medical diagnoses, which included asthma (5), chronic obstructive lung disease (8), adult respiratory distress syndrome (13), pneumonia (36), cardiac disease (4), neurological disease (31), and

*In addition, 1 patient had continuous positive airways pressure IPPV=mtermittent positive pressure ventilation, CPPV=continuous pressure ventilation; ETT = endotracheal tube.

positive

septicaemia (19), or surgical categories--general (20), cardiothoracic (3), obstetric/gynaecological (7). There were a few patients with other diagnoses, including renal and hepatic failure, collagen vascular diseases, burns, leukaemia, and hypothermia. The prevalence of pre-existing disease and the incidence of compromising factors during the ICU stay were similar in

7

TABLE IV-CAUSES OF DEATH

the SDD and placebo

groups-underlying carcinoma, 5 vs 2; previous immunosuppressant (cyclophosphamide and high-dose prednisone) treatment, 10 each; alcohol abuse (>40 g/day), 20 vs 35, diabetes mellitus, 8 vs 10; history of peptic ulcer disease, 6 vs 7; shock on admission, 16 vs 11; admission to hospital more than 48 h before transfer to ICU, 42 vs 47; obesity,11 vs 19; and malnutrition,11 vs 10. Other factors predisposir 19 to infection while in the ICU, such as reduced consciousness (34 vs 48), confmement in bed for at least 3 days (81 vs 95), bulbar dysfunction (9 vs 5), use of total parenteral nutrition (32 vs 44), and nasogastric drainage for more than 5 days (41 vs 36) did not differ significantly between the groups. There were no differences in use of ventilatory and other therapy that might have predisposed to the development of infection (table 11). Hepatic failure (12 vs 14), gastrointestinal haemorrhage (11 vs 14), failed wean (10 vs 14), renal failure (7 vs 7), shock (systolic blood pressure < 70 mm Hg; 8 vs 2), haematological failure (4 vs 3), diarrhoea (19 vs 20), and coma (7 vs 3) occurred with similar frequency during the ICU stay in the SDD and

placebo groups. There were no differences between the two groups in duration of ICU stay, hospital stay, infection rate, incidence of complications, or outcome (table ill). Most deaths were due to multiple organ system failure (table IV). Analysis of various subgroups did not confirm previous reports of special benefit from SDD for some; we found only that the incidence of infection was lowered by SDD in the midrange APACHE II group (table v). All trauma patients admitted to our ICU had chest trauma that required mean

TABLE V-ANALYSIS OF SUBGROUPS WITH PREVIOUSLY REPORTED14 HIGH RISK OF INFECTION

*p