American Journal of Respiratory and Critical Care Medicine

We investigated the endemicity of Pseudomonas aeruginosa in intensive care units (ICUs) through analyses of surveillance cultures (from the rectum, stomach, oropharynx, and trachea; n = 1,089), and clinical cultures (n = 2,393) from 297 consecutive patients. Multiple isolates of P. aeruginosa (n = 353) were genotyped. Variables associated with acquisition of respiratory tract colonization (RTC) were tested in a risk factor analysis. The mean daily prevalence of colonization was 34%. On admission, 22 patients had intestinal colonization and 13 had RTC. Twenty patients acquired colonization in the intestinal and 24 in the respiratory tract. Forty-four different genotypes were found; 38 (86%) were isolated from individual patients only. In all, 37 patients had RTC with a total of 38 genotypes: 13 (34%) were colonized on admission, 9 (24%) acquired RTC with a novel genotype during a stay in the ICU, five (13%) acquired colonization from their intestinal tract and three (8%) were colonized via cross-acquisition. In eight patients (21%), no route could be demonstrated for colonization. Antibiotics providing P. aeruginosa with a selective growth advantage were associated with acquired RTC. Endemicity of colonization with P. aeruginosa is characterized by polyclonality, and seems to be maintained by continuous admittance of colonized patients and selection pressure from antibiotics rather than by cross-acquisition.

Pseudomonas aeruginosa is a pathogen frequently causing ventilator-associated pneumonia (VAP) in intensive care units (ICUs) (1, 2). However, infection rates in ICUs often represent only the tip of an iceberg, whereas the true bacterial load within an ICU is represented by colonization rates. Respiratory tract colonization (RTC) is the most important risk factor for the development of VAP caused by P. aeruginosa (3, 4). RTC with P. aeruginosa is endemic in many ICUs worldwide. Understanding the mechanisms establishing and maintaining endemicity of P. aeruginosa colonization is therefore important. In the pathogenesis of RTC, microorganisms may originate from endogenous or exogenous sources. The intestines are considered the most important endogenous source of microorganisms reaching the respiratory tract via the gastropulmonary route (5, 6), via colonization of the skin, or via transiently colonized hands of health-care workers during tending (7-10). The potential importance of exogenous sources of P. aeruginosa (e.g., contaminated equipment) has been repeatedly demonstrated during outbreaks of infections with this organism (11-16). In addition, colonized patients serve as continuous exogenous sources of microorganisms from which other patients can be colonized via cross-acquisition (17). Knowledge about the relative importance of exogenous and endogenous routes of colonization is essential in order to design targeted strategies for infection prevention. Obviously, preventive strategies will be different when each of the two routes is dominant. The relative importance of both routes of colonization in endemic settings has seldom been determined.

Acquisition of RTC by P. aeruginosa has been associated with the duration of ICU stay, the necessity for and duration of mechanical ventilation (4), and with previous antibiotic use, providing P. aeruginosa with a selective growth advantage (18). Furthermore, the risk for cross-acquisition may be associated with the number of other patients in the ward who are already colonized, thereby increasing colonization pressure. This association has been demonstrated for vancomycin-resistant enterococci (VRE) (19). Additionally, the severity of illness during an ICU stay may predict which patients will be at risk for acquisition of P. aeruginosa.

To study the epidemiology of P. aeruginosa colonization in detail during endemicity, we analyzed rectal, gastric, and respiratory tract colonization with P. aeruginosa prospectively during a 300-d period. We also studied variables influencing acquisition of P. aeruginosa. Similarity of isolates of P. aeruginosa was based on molecular typing techniques.


The study was performed in one of two general ICUs of the University Hospital Maastricht, a 700-bed hospital. The ICU in which the study was done is a nine-bed ward containing a mixed population of adult medical, surgical, neurologic, neurosurgical, and trauma patients, and pediatric patients. The ICU has four boxes (cubicles that partition off a distinct area for a patient's bed and equipment needed to care for the patient) and the potential to accommodate five beds on the open ward (distance between the beds on the ward is ± 3 m). During the study period there were no problems with multiresistant pathogens, and patients placed in boxes were subjected to the same level of infection control measures as were patients on the open ward. There was no predefined selection for admittance of patients to a box or to the open ward, although children were preferably treated in a box. The patient-to-nurse ratio during the day is 1:1, with cross-coverage during breaks, and is 2:1 during the night. There are two large sinks on the ward, one near the entrance and one placed in the center of the ICU. Each box has its own sink. The handwashing regimen on entrance to and exit from the ICU, and in the interval between patient contacts, includes washing with disinfecting soap and/or alcohol.

Patient Population

All patients admitted to the ICU between August 20, 1994 and June 14, 1995 were included in the study. Demographic characteristics of each patient and the patient's diagnosis on admission, underlying diseases, vital signs, medications, and laboratory values were recorded prospectively. The Acute Physiology and Chronic Health Evaluation (APACHE) II scoring system was used to assess the severity of illness of adult patients (20). The Therapeutic Intensity Scoring System (TISS-28) was used to estimate the severity of illness and intensity of patient care, with scoring done daily by the nursing staff (21).


Gastric, oropharyngeal, and rectal swab samples were collected twice weekly at defined time points (Monday and Thursday, 8:00 a.m.) from all patients treated in the ICU. Tracheal aspirates were obtained from intubated patients only. Additionally, surveillance cultures were also obtained on admission from patients needing mechanical ventilation. Further specimens were collected and analyzed as required. Patients were studied until discharge from the ICU or death. No rectal swabs were taken from children, but stool was analyzed when available.

Microbiologic Analysis and Genotyping

Semiquantitative and/or quantitative microbiologic analyses of culture samples were performed according to standard microbiologic methods (22, 23). Specimens were inoculated immediately after sampling, 7 d per week, on Pseudomonas Agar (Oxoid 559, Basingstoke, UK) enriched with Pseudomonas C-F-C Supplement (Oxoid SR103E). From each patient and each body site or clinical specimen positive for P. aeruginosa, three colonies were randomly selected for analyses, and these were stored separately at a temperature of −70° C until further processing. Antibiotic susceptibility was determined by means of a microbroth dilution method according to National Commission for Clinical Laboratory Standards (NCCLS) guidelines (40). P. aeruginosa ATCC 27853 (American Type Culture Collection, Rockville, MD) and Escherichia coli ATCC 25922 and ATCC 29212 were used as reference strains. The criteria used for susceptibility and resistance were those specified in the NCCLS guidelines. Typing was performed by random amplification of polymorphic DNA (RAPD) as described by Renders and colleagues (24), with minor modifications.

Electrophoresis was performed in 0.5× Tris-borate-ethylenediamine tetraacetic acid buffer on a Protean II electrophoresis system (Bio-Rad, Inc.). Banding patterns were visualized by staining the electrophoresis gel with ethidium bromide. Strains of P. aeruginosa were differentiated according to published criteria (24), without consideration of single-band differences. Typing was performed by macrorestriction of the genome with SpeI (Boehringer Mannheim, Germany) and separation of the DNA fragments through pulsed-field gel electrophoresis (PFGE) on a Bio-Rad CHEF DR-II system (25). Strains were differentiated according to the criteria of Tenover and coworkers (26).

Definitions of Colonization

Colonization was defined as the isolation of P. aeruginosa from two consecutive cultures in the absence of infection. Colonization on admission was established by a positive culture within 24 h. Colonization was designated as acquired if there was no colonization on admission and there were at least two positive cultures with P. aeruginosa from consecutive samples. The initial site of colonization was defined as the site at which colonization was first identified. Sequences of colonization were determined by the chronologic comparison of acquisition of colonization at different body sites.

RTC was defined as being of endogenous origin when, in an individual patient, P. aeruginosa strains with similar genotypes were isolated from the intestinal tract before RTC was demonstrated. In contrast, acquired RTC was judged to be due to cross-colonization (i.e., exogenous colonization) when genotypically identical isolates were identified that had previously been cultured from another patient who was in the ICU at the same time. Primary RTC was defined as acquired colonization with a genotype of P. aeruginosa that had not been isolated previously from the intestinal tract of the same patient or from any other body site of another patient in the ICU.

Definition of VAP and Other Infections

VAP was considered to be ICU-acquired if a clinical condition fulfilling the criteria for pneumonia developed after the patient had been in the ICU for at least 3 d. In cases of clinically suspected pneumonia, bronchoscopy with bronchoalveolar lavage (BAL) and the protected-specimen brush (PSB) technique was performed as previously described (27). The diagnosis of VAP was established with either a positive quantitative culture of samples obtained by BAL (cutoff point: 104 cfu/ml) or the PSB technique (cutoff point: 103 cfu/ml) or with a positive blood culture unrelated to another source of infection; a new, progressive, or persistent infiltrate on chest radiography; and when at least three of the following four criteria were met: (1) rectal temperature > 38.0° C or < 35.5° C; (2) blood leukocytosis (> 10.103/mm3) and/or a decrease in the white blood cell count or blood leukopenia (< 3.103/mm3); (3) > 10 leukocytes per high-power field in a Gram-stained smear of tracheal aspirate; and (4) a positive culture from a tracheal aspirate. All other infections were diagnosed using the criteria defined by the Centers for Disease Control and Prevention (28).

Design and Statistics of Analysis of Acquisition of Colonization

We created a statistical model of the relative influences of colonization pressure, antibiotic use, APACHE II score on admission to the ICU, and mean TISS-28 of daily scores on acquisition of RTC by P. aeruginosa. All patients without RTC by P. aeruginosa at the time of admission were included in this analysis. A similar model has been used to study the epidemiology of VRE (19). For each study day, the point prevalence of P. aeruginosa in the ICU was calculated as follows: (number of patients colonized with P. aeruginosa on that day divided by the number of patients treated in the ICU on that day). Subsequently, for each patient without RTC by P. aeruginosa on admission, we calculated the average point prevalence of P. aeruginosa for all days in the ICU until acquisition of RTC by P. aeruginosa or until discharge (if the patient did not acquire RTC). This calculated number reflects the colonization pressure with P. aeruginosa for the period of the patient's ICU stay during which the patient was not colonized.

The percentage of days on which noncolonized patients received antibiotics not affecting P. aeruginosa, and thus providing these bacteria with a selective growth advantage (“antibiotic pressure”) was calculated until acquisition of P. aeruginosa or discharge from the ICU, as were mean daily TISS-28 scores.

The effects of colonization pressure, antibiotic pressure, APACHE II score on admission, and mean of daily TISS-28 scores on acquisition of RTC by P. aeruginosa were analyzed in a Cox proportional hazards model (SPSS for Windows, release 6.1.2; SPSS, Inc., Chicago, IL), in which the number of days until acquisition of P. aeruginosa was the dependent variable. Each variable was analyzed separately in a Cox regression model, and the best-fitting model was constructed through the simultaneous inclusion of the significantly related variables. Hazard ratios with corresponding 95% confidence intervals (CIs) were computed. A hazard ratio of a variable with two categories represents the ratio of the two category-dependent probabilities for occurrence of the event toward which the variable contributes. In our study, the integration of the hazard function over a certain period of time (e.g., day by day) gave the probability of acquiring RTC by P. aeruginosa in that period. The significant covariates were obtained by means of a forward stepwise selection, using the p values of the likelihood ratio test. Goodness-of-fit was assessed from the three plots of partial residuals against time, with one plot for each covariate. In the plots the points were randomly distributed. For other statistical analyses, the t test, Mann–Whitney U-test, and chi-square test were used when appropriate.

Study Population

A total of 297 patients (219 adults [74%] and 78 children [26%]) were admitted during the 300-d study period (Table 1). Fifty-one percent of all patients had been hospitalized be-fore ICU admission, and 51% received mechanical ventilation. Adults were ventilated for a longer average period than were children (p < 0.05), although the proportions of patients receiving ventilation for at least 3 d were similar. The mean duration of ICU stay was comparable for adults and children, at 8 ± 13 d (mean ± SD) and 9 ± 22 d, respectively (p = 0.6). The daily occupancy rate of the ICU was 79 ± 14%.


Adults (n = 219)Children (n = 78)
Age, yr (mean ± SD)58 ± 183 ± 4
APACHE II score (mean ± SD)19 ± 8 (n = 124)ND
Hospitalization before ICU admission
 Number (%)116 (53)36 (47)
 Duration (mean ± SD) (median)11 ± 16 (4)12 ± 25 (4)
Mechanical ventilation
 Number (%)115 (52)36 (47)
 Duration (mean ± SD)10 ± 145 ± 4
 Duration ⩾ 3 d (%)75 (34)23 (29)
Medical specialty
Duration of stay in ICU (mean ± SD) 8 ± 139 ± 22
 In ICU (%)40 (18)6 (8)
 In hospital (%)19 (9)0

Definition of abbreviations: APACHE = Acute Physiology and Chronic Health Evaluation; ICU = intensive care unit; ND = not determined.


In all, 1,089 surveillance cultures and 2,393 clinical cultures were analyzed (Table 2). Cultures were not available from 72 adults and 33 children. The mean durations of stay in the ICU for these patients were 1.9 ± 0.7 d (range: 1 to 5 d) and 2.4 ± 1.2 d (range: 1 to 7 days) for adults and children, respectively. These patients accounted for 140 of 1,723 patient days (8%) for adults and 78 of 681 patient days (11%) for children.


No. of PatientsNo. of CulturesNo. of PatientsNo. of CulturesTotal No. of Cultures
 Rectal swabs 9125511 16271
 Gastric aspirates 8429518 50345
 Oropharyngeal swabs 9538131 92473
Clinical cultures
 Tracheal aspirates10046334105568
 Respiratory tract samples 32178 3  4182
 Blood 7747812 32510
 Urine11851621 36552
 Intravenous lines 8626914 25294
 Others 7424117 46287

Prevalence of Colonization

The mean daily prevalence of colonization with P. aeruginosa was 34 ± 16%, with mean prevalences of intestinal colonization alone of 12 ± 9%, and of RTC with or without intestinal colonization of 22 ± 13% (Figure 1). Patients from whom no cultures were available were assumed not to be colonized with P. aeruginosa.

Colonization on Admission

No children were colonized with P. aeruginosa on admission (Table 3). Among all adult patients with available cultures, 22 (24%) were colonized in the intestinal tract, and 10 (9%) and 7 (7%), respectively, were colonized in the oropharynx and trachea. Because four of these patients were colonized at both sites, a total of 13 patients (13%) had RTC by P. aeruginosa on admission.


Colonization on admission
 Rectum22/91 (24%)022/102 (22%)
 Stomach4/84 (5%)04/102 (4%)
 Oropharynx10/95 (9%)010/126 (8%)
 Trachea7/100 (7%)0/347/134 (5%)
Acquired colonization
 Rectum19/91 (21%)1/11 (9%)20/102 (20%)
 Stomach14/84 (17%)2/18 (11%)16/102 (16%)
 Oropharynx21/95 (22%)2/31 (6%)23/126 (18%)
 Trachea18/100 (18%)2/34 (6%)20/134 (15%)
Number of body sites colonized
Positive clinical cultures24024

Acquired Colonization

Acquired intestinal colonization was demonstrated in 20 patients (19 adults and one child) (Table 3). Acquired RTC was demonstrated in 24 patients, of whom 23 acquired oropharyngeal and 20 acquired tracheal colonization. Again, adults outnumbered children in the acquisition of colonization. When colonized, patients were usually colonized at multiple body sites simultaneously. Only 16 (32%) of 50 patients with P. aeruginosa colonization were colonized at a single body site, which was the intestinal tract in 13 of the 16 patients.

Genotypic Analysis

P. aeruginosa was isolated from 51 patients (17%). Of these patients, 50 were colonized and from 1 patient P. aeruginosa was isolated from a wound culture without further colonization. In all, 353 isolates from 44 patients were genotyped using RAPD. Representative isolates from different RAPD types were further typed with PFGE, which resulted in 44 distinct genotypes. Thirty-eight genotypes (86%) were isolated from individual patients only, four genotypes colonized two patients each, one genotype colonized three patients, and one genotype was isolated from five different patients.

Routes of Colonization

Acquired RTC occurred in 24 patients, and genotyping of isolates could be performed for each patient. In the cases of 23 patients, all P. aeruginosa isolates tested belonged to a single genotype per patient, and one patient acquired RTC with two genotypes. Therefore, routes of acquisition could be analyzed for 25 genotypes. On the basis of chronologic patterns and similarity of genotypes, an endogenous route of colonization was demonstrated for five (20%) of these genotypes (in five patients). In all cases, the intestine was the initial source pf P. aeruginosa preceding RTC. The mean duration between demonstration of intestinal colonization and RTC was 12 ± 11 d (range: 3 to 31 d). An exogenous source for RTC, with genotypes previously colonizing another ICU patient, was demonstrated for only three genotypes (12%) (three patients), with RTC occurring 3, 4, and 25 d, respectively, after admission to the ICU. Of the remaining 17 genotypes, primary RTC occurred in nine cases (nine patients), after a mean of 10 ± 6 d (range: 2 to 18 d). For the other eight genotypes (eight patients), no route could be reliably demonstrated because intestinal colonization and RTC were found simultaneously, probably because cultures were obtained simultaneously on a twice weekly basis. However, among patients with both intestinal colonization and RTC (either present on admission or acquired) that were not demonstrated simultaneously (n = 9), intestinal colonization preceded RTC in six patients. Therefore, it is likely that some episodes of simultaneous colonization did in fact occur through endogenous routes, and that the true percentage of episodes of endogenous colonization was higher than 20%. If all cases in which intestinal colonization and RTC were demonstrated simultaneously were in fact from an endogenous source, the total proportion of acquired RTC via the endogenous route would have been 52% (n = 13). In summary, 37 patients had RTC by P. aeruginosa; 13 (34%) were colonized on admission, nine (24%) had primary RTC, five (13%) acquired colonization from an endogenous source, eight (21%) acquired colonization simultaneously in the intestinal and respiratory tracts, and three (8%) had RTC via cross-acquisition (Figure 2).

Intestinal colonization was demonstrated in 42 patients. Acquired intestinal colonization was demonstrated in 20 patients after a mean of 12 ± 10 d (range: 3 to 37 d), and isolates were genotyped from 19 patients, yielding 20 different genotypes. In three patients, acquired intestinal colonization was preceded by RTC, in eight patients it was demonstrated simultaneously with RTC, and in nine patients it appeared de novo.

Gastric colonization never preceded RTC. In contrast, acquired gastric colonization was in 11 of 16 patients (69%) preceded by RTC with identical genotypes.

Infections with P. aeruginosa

Twelve patients developed an infection or clinically suspected infection in which P. aeruginosa was involved. All 12 patients were previously colonized with P. aeruginosa. Five patients developed an episode of VAP, and four patients had clinically suspected VAP that was not established by quantitative cultures from samples obtained via bronchoscopic techniques. Three other patients had non-respiratory tract infections with P. aeruginosa: one had intravenous line sepsis, one had bacteremia with an unknown focus, and one had an abdominal abscess. As a result, eight (16%) of 51 colonized patients developed an infection, which means that for each identified infection five to six other patients were colonized.

Four of nine patients with respiratory tract infections (three with VAP) were already colonized on admission with the causative genotypes of P. aeruginosa, and one patient (with VAP) had primary RTC. Two patients (one with VAP) acquired RTC endogenously, and intestinal colonization and RTC were demonstrated simultaneously in the remaining two patients. In summary, three patients with confirmed VAP, one patient with clinically suspected VAP, and one patient with a nonrespiratory tract infection were already colonized with the causative genotype of P. aeruginosa on admission.

Risk Factors for Acquisition of RTC

Only patients with available cultures were included in the risk factor analysis. Patients who acquired RTC by P. aeruginosa had a longer ICU stay than did patients who did not acquire colonization (Table 4). Acquisition of RTC was associated with a greater use of antibiotics that do not affect P. aeruginosa; the percentages of antibiotic days were 67 ± 41% for patients who did and 37 ± 42% for those who did not acquire RTC (p = 0.03, Mann–Whitney U test). Amoxicillin–clavulanic acid was by far the most frequently used antibiotic, and was administered to 45% (n = 78) of all patients with available cultures (Table 5). Other frequently prescribed antibiotics that do not affect P. aeruginosa were penicillin G (n = 11), flucloxacillin (n = 10), erythromycin (n = 9), and metronidazole (n = 8). Of antibiotics affecting P. aeruginosa, gentamicin (n = 58) and piperacillin (n = 19) were most frequently used. All isolates were susceptible to gentamicin. In addition to antibiotic use, the mean intensity of patient care was higher for patients who acquired RTC than for those who did not; TISS-28 scores were 33 ± 9 and 25 ± 9, respectively (p < 0.001, t test). APACHE II scores on the day of admission and colonization pressure were not associated with acquisition of RTC by P. aeruginosa.


VariablesAcquisition of Colonization
Yes (n = 24)No (n = 157)p Value
Duration of ICU stay in days, mean ± SD  Range24 ± 22 (5–108) 9 ± 17 (1–187)< 0.001t test
Duration until acquisition10 ± 8
APACHE II scores20 ± 819 ± 8NS
(n = 22)(n = 83)
Colonization pressure, %34 ± 1431 ± 12NS
(n = 24)(n = 156)
Antibiotic pressure, %67 ± 3937 ± 410.002
(n = 22)(n = 152)MWU
TISS-28 scores34 ± 1025 ± 9< 0.001
(n = 22)(n = 127) t test

Definition of abbreviations: APACHE = Acute Physiology and Chronic Health Evaluation; ICU = intensive care unit; MWU = Mann-Whitney U test; NS = not significant; TISS = Therapeutic Intensity Scoring System.


Acquisition of Respiratory Tract Colonization withPseudomonas aeruginosa
Yes(n = 22)No(n =152)
Antibiotics not affecting P. aeruginosa
 Amoxicillin–clavulanic acid18(82%)60 (39%)
 Penicillin G 1 (5%)10 (7%)
 Amoxicillin 1 (5%) 6 (4%)
 Flucloxacillin 2(10%) 8 (5%)
 Cefuroxime 0 7 (5%)
 Metronidazole 1 (5%) 7 (5%)
 Vancomycin 0 5 (3%)
 Erythromycin 2(10%) 7 (5%)
 Cotrimoxazole 0 4 (3%)
 Rifampicin 1 (5%) 0
 Clindamycin 0 1 (1%)
 Chloramphenicol 0 2 (1%)
Antibiotics affecting P. aeruginosa
 Gentamicin15 (68%)43(28%)
 Piperacillin 4(18%)15(10%)
 Ceftriaxone 0 5 (3%)
 Ceftazidime 0 2 (1%)
 Ticarcillin–clavulanic acid 0 2 (1%)
 Ciprofloxacin 1 (5%) 4 (3%)
 Imipenem 0 1 (1%)

As mentioned earlier, adults acquired colonization more frequently than did children, at 23 of 146 (16%) versus two of 46 (4%) (p < 0.05). Although children were more frequently treated in one of the four ICU boxes (73% versus 27%, p < 0.00001), treatment in a box was not on an overall basis associated with a lower acquisition rate. Acquisition occurred in 10 of 129 patients (8%) treated and in 14 of 174 patients (8%) not treated in a box. Furthermore, both the percentages of days with antibiotic use and colonization pressures were comparable for children and adults. However, the intensity of patient care seemed to have been lower for children; TISS-28 scores were 21 ± 9 for children and 25 ± 9 for adults (p = 0.001, t test).

In the Cox regression, the percentages of antibiotic days with drugs not affecting P. aeruginosa were significantly related to acquisition of RTC by P. aeruginosa, and statistical significance was approached for the mean of daily TISS scores. The hazard ratios were 1.0159 (95% CI = 1.0025 to 1.0296, p = 0.02) for percentages of antibiotic days and 1.0599 (95% CI = 0.9982 to 1.1254, p = 0.06) for the mean of daily TISS scores.

Our study provides a detailed analysis of colonization and infection with P. aeruginosa in an ICU in which colonization is endemic. Under the circumstances tested, the epidemiology of colonization is characterized by polyclonality, with most patients being colonized with unique genotypes of P. aeruginosa. RTC was frequently present on admission, and when it was acquired in the ICU, the intestinal tract seemed to be more important as a source than were cross-acquisition and acquisition from the gastric reservoir. Therefore, increasing compliance with infection-control measures, as well as attempts to interrupt intestinal or gastric colonization, are unlikely to decrease the endemic level of P. aeruginosa colonization. Antibiotics providing P. aeruginosa with a selective growth advantage were the most important risk factors for acquisition of RTC.

The combination of extensive surveillance and genotyping of 353 isolates identified four important aspects relevant to endemicity of colonization with P. aeruginosa.

First, introduction of new strains of P. aeruginosa into the ICU seems to be a driving force for maintaining endemicity; in our study, 35% of all patients with RTC were already colonized on admission. In addition, in 24% of all patients with RTC, isolated genotypes of P. aeruginosa appeared de novo in the respiratory tract. This suggests that RTC may have been present on admission, with a bacterial load below the detection limit of our methods, thereby strengthening the role of introduction of new strains in RTC. However, multiple exogenous sources are not completely excluded.

Second, for both intestinal colonization and RTC with P. aeruginosa, cross-acquisition seemed to be relatively unimportant. On the basis of a comparison of genotypes, only three events of cross-acquisition could be demonstrated, despite a mean prevalence rate of colonization of 35% and the abundance of surveillance cultures and genotype analyses. The absence of an important role of cross-acquisition was supported by the finding that colonization pressure in the ICU was not associated with acquisition of P. aeruginosa. Several other studies also failed to demonstrate an important role of cross-acquisition in the epidemiology of colonization by P. aeruginosa (3, 9, 29-31). Our findings should not be interpreted as downplaying the potential relevance of cross-acquisition, which has been demonstrated repeatedly, mostly because of lapses in compliance with simple infection control measures such as inappropriate handwashing (7, 29). The paramount importance of high hygiene standards in the ICU remains undisputed. However, as demonstrated in our study, high colonization rates are not always equivalent to cross-acquisition.

Third, surveillance cultures and molecular genotyping are indispensable when studying the ICU epidemiology of bacterial colonization. Our data would have been completely different and wrong if only strains from sites of infection and no rectal cultures had been analyzed. In addition, valuable information would have been absent if no genotyping had been done. In our ICU, antibiotic susceptibilities have no discriminatory power, and methods of phenotyping (such as outermembrane protein analysis and serotyping) are less specific than is genotyping (32).

Additionally, the present study confirms the relevance of antibiotic pressure as a variable influencing the epidemiology of colonization and infection by nosocomial pathogens. Antibiotics providing P. aeruginosa with a selective growth advantage were the most important risk factor for acquired RTC. Interestingly, gentamicin use was also associated with acquired RTC in univariate analysis. Since all isolates of P. aeruginosa in our study were susceptible to gentamicin, this strongly suggests that systemic use of gentamicin hardly influences colonization. An antibiotic-induced selective growth advantage increases the prevalence of colonization in the ICU (“colonization pressure”), and may therefore even interfere with the efficacy of infection-control measures. The protective effects of hygiene measures depend on compliance by medical and nursing staff. However, a certain level of compliance may be sufficient when the mean prevalence of colonization is 35%, but may be insufficient when colonization levels double (33). Unfortunately, we had no information about compliance with infection-control measures in our ICU during the study period. However, preliminary data from observational studies at a later period (with similar prevalence rates) showed compliance rates with handwashing of approximately 40% (data not shown).

Despite the fact that 3,482 cultures were analyzed and that 353 isolates were genotyped, our study has two potential insufficiencies that deserve comment. Cultures were not obtained from 105 patients. Since these patients were in the ICU for very short periods and accounted for less than 10% of the total number of patient days studied, it seems unlikely that the absence of cultures in these cases influenced our findings. In addition, surveillance cultures from the inanimate environment were not included in the study. However, the polyclonal character of the endemicity of colonization by P. aeruginosa in our study does not support an important role of exogenous sources. Moreover, environmental contamination was not an important source of P. aeruginosa in two recent studies (3, 31).

How can our findings be used to develop strategies for preventing infections caused by P. aeruginosa? Infection-control measures remain the cornerstone of infection prevention in the ICU. On the basis of the results of the present study, prevention of colonization in the intestines or the stomach would have little effect on RTC and infection rates with P. aeruginosa. Fully effective intestinal decontamination could have prevented from five to 13 episodes of RTC. Since gastric colonization never preceded RTC, decontamination at this site would not have had any effect on RTC. This confirms the results of studies in which intestinal and gastric decontamination were achieved with topical nonabsorbable antibiotics (34, 35). In fact, analyses of sequences of colonization of bacteria causing VAP have repeatedly shown that the stomach is not an important reservoir for P. aeruginosa (36). A regimen that prevents colonization of the upper respiratory tract would be much more effective in preventing VAP. Such an approach has been tested twice, and in both studies oropharyngeal decontamination with nonabsorbable antibiotics significantly reduced the incidence of VAP (37, 38).

Although our study identified antibiotics as an important risk factor for acquisition of RTC by P. aeruginosa, it is questionable whether their reduced use offers an approach for preventing infection. Eradication of the normal human bacterial flora facilitates overgrowth with P. aeruginosa. Amoxicillin– clavulanic acid accounted for the vast majority of antibiotic provision of a selective growth advantage for P. aeruginosa. Therefore, eradication of the anaerobic flora probably occurred with this drug. It has been suggested that this anaerobic flora provides colonization resistance to potentially pathogenic bacteria such as P. aeruginosa (39). However, whether a change of first-choice therapy from amoxicillin–clavulanic acid to a second-generation cephalosporin, for example, will decrease rates of RTC by P. aeruginosa is unknown.

The authors thank the ICU medical staff and nurses who helped make this study possible, and Mr. Anton W. Ambergen for statistical advice.

Supported by Grant 28-2125-1 from the Praevention Foundation, The Netherlands.

1. Emori T. G., Gaynes R. P.An overview of nosocomial infections, including the role of the microbiology laboratory. Clin. Microbiol. Rev.61993428442
2. Bergmans D. C. J. J., Bonten M. J. M., Gaillard C. A., van Tiel F. H., van der Geest S., de Leeuw P. W., Stobberingh E. E.Indications for antibiotic use in ICU patients: a one-year prospective surveillance. J. Antimicrob. Chemother.391997527535
3. Talon D., Mulin B., Rouget C., Bailly P., Thouverez M., Viel J.-F.Risks and routes for ventilator-associated pneumonia with Pseudomonas aeruginosa. Am. J. Respir. Crit. Care Med.1571998978984
4. Bonten M. J. M., Bergmans D. C. J. J., Ambergen A. W., de Leeuw P. W., van der Geest S., Stobberingh E. E., Gaillard C. A.Risk factors for pneumonia, and colonization of respiratory tract and stomach in mechanically ventilated ICU patients. Am. J. Respir. Crit. Care Med154199613391346
5. Heyland D., Mandell L. A.Gastric colonization by gram-negative bacilli and nosocomial pneumonia in the intensive care unit patient: evidence for causation. Chest1011992187193
6. Inglis T. J. J., Sherratt M. J., Sproat L. J., Gibson J. S., Hawkey P. M.Gastroduodenal dysfunction and bacterial colonization of the ventilated lung. Lancet3411993911913
7. Widmer A. F., Wenzel R. P., Trilla A., Bale M. J., Jones R. N., Doebbeling B. N.Outbreak of Pseudomonas aeruginosa infections in a surgical intensive care unit: probable transmission via hands of a health care worker. Clin. Infect. Dis.161993372376
8. Boyce J. M., Opal S. M., Potter-Bynoe G., Medeiros A. A.Spread of methicillin-resistant Staphylococcus aureus in a hospital after exposure to a health care worker with chronic sinusitis. Clin. Infect. Dis.171993496504
9. Olson B., Weinstein R. A., Nathan C., Chamberlin W., Kabins S. A.Epidemiology of endemic Pseudomonas aeruginosa: why infection control efforts have failed. J. Infect. Dis.1501984808816
10. Noone M. R., Pitt T. L., Bedder M., Hewlett A. M., Rogers K. B.Pseudomonas aeruginosa colonisation in an intensive therapy unit: role of cross infection and host factors. Br. Med. J.2861983341344
11. Craven D. E., Goularte T. A., Make B. J.Contaminated condensate in mechanical ventilator circuits: a risk factor for nosocomial pneumonia? Am. Rev. Respir. Dis.1291984625628
12. Grundmann H., Kropec A., Hartung D., Berner R., Daschner F.Pseudomonas aeruginosa in a neonatal intensive care unit: reservoirs and ecology of the nosocomial pathogen. J. Infect. Dis.1681993943947
13. File T. M., Tan J. S., Thomson R. B., Stephens C., Thompson P.An outbreak of Pseudomonas aeruginosa ventilator-associated respiratory infections due to contaminated food coloring dye—further evidence of the significance of gastric colonization preceding nosocomial pneumonia. Infect. Control Hosp. Epidemiol.161995417418
14. Whitby J. L., Rampling A.Pseudomonas aeruginosa contamination in domestic and hospital environment. Lancet119721517
15. Favero M. S., Carson L. A., Bond W. W., Petersen N. J.Pseudomonas aeruginosa: growth in distilled water from hospitals. Science1731971836838
16. Tredget E. E., Shankowsky H. A., Joffe M. A., Inkson T. I., Volpel K., Paranchych W., Kibsey P. C., MacGregor J. D., Alton, Burke J. F.Epidemiology of infections with Pseudomonas aeruginosa in burn patients: the role of hydrotherapy. J. Infect. Dis.151992941949
17. Maki D. G.Control of colonization and transmission of pathogenic bacteria in the hospital. Ann. Intern. Med.891978777780
18. Olson B., Weinstein R. A., Nathan C., Chamberlin W., Kabins S.Occult aminoglycoside resistance in Pseudomonas aeruginosa: epidemiology and implications for therapy and control. J. Infect. Dis.1521985769774
19. Bonten M. J. M., Slaughter S., Ambergen A. W., Hayden M. K., van Voorhis J., Nathan C., Weinstein R. A.The role of “colonization pressure” in the spread of vancomycin-resistant enterococci: an important infection control variable. Arch. Intern. Med.158199811271132
20. Knaus W. A., Draper E. A., Wagner D. P., Zimmerman J. E.APACHE II: a severity of disease classification system. Crit. Care Med.131985818829
21. Miranda D. R., de Rijk A., Schaufeli W.Simplified therapeutic intervention scoring system: the TISS-28 items—results from a multicenter study. Crit. Care Med.2419966473
22. Stokes, E. J., and G. L. Ridgway. 1980. Clinical Bacteriology, 5th ed. Edward Arnold, London.
23. Isenberg, H. D. 1992. Clinical Microbiology Procedures Handbook, 1st ed. American Society for Microbiology, Washington, DC.
24. Renders N., Romling U., Verbrugh H., van Belkum A.Comparative typing of Pseudomonas aeruginosa by random amplification of polymorphic DNA or pulsed-field gel electrophoresis of DNA macrorestriction fragments. J. Clin. Microbiol.34199631903195
25. Grothues D., Koopmann U., Hardt H., Tummler B..Genome fingerprinting of Pseudomonas aeruginosa indicates colonization of cystic fibrosis siblings with closely related strains. J. Clin. Microbiol.26198819731977
26. Tenover F. C., Arbeit R. D., Goering R. V., Mickelsen P. A., Murray B. E., Persing D. H., Swaminathan B.Interpreting chromosomal DNA restriction patterns produced by pulsed-field gel electrophoresis: criteria for bacterial strain typing. J. Clin. Microbiol.33199522332239
27. Bonten M. J. M., Gaillard C. A., van Tiel F. H., Smeets H. G. W., van der Geest S., Stobberingh E. E.The stomach is not a source for colonization of the upper respiratory tract and pneumonia in ICU patients. Chest1051994878884
28. Centers for Disease Control1988 CDC definitions for nosocomial infections. Am. Rev. Respir. Dis.139198910581059
29. Gruner E., Kropec A., Huebner J., Altwegg M., Daschner F.Ribotyping of Pseudomonas aeruginosa strains isolated from surgical intensive care patients. J. Infect. Dis.167199312161220
30. Griffith S. J., Nathan C., Selander R. K., Chamberlin W., Gordon S., Kabins S., Weinstein R. A.The epidemiology of Pseudomonas aeruginosa in oncology patients in a general hospital. J. Infect. Dis.160198910301036
31. Blanc D. S., Petignat C., Janin B., Bille J., Francioli P.Frequency and molecular diversity of Pseudomonas aeruginosa upon admission and during hospitalization: a prospective epidemiologic study. Clin. Microbiol. Infect. Dis.41998242247
32. Bergmans D., Bonten M., van Tiel F., Gaillard C., London N., van der Geest S., de Leeuw P., Stobberingh E.Value of phenotyping methods as an initial screening of Pseudomonas aeruginosa in epidemiologic studies. Infection251997350354
33. Austin D. A., Bonten M. J. M., Slaughter S., Weinstein R. A., Anderson R. M.Vancomycin-resistant enterococci in intensive care hospital settings: transmission dynamics, persistence and the impact of infection control programs. Proc. Natl. Acad. Sci. U.S.A.96199969086913
34. Brun-Buisson C., Legrand P., Rauss A., Richard C., Montravers F., Besbes M., Meakins J. L., Soussy C. J., Lemaire F.Intestinal decontamination for control of nosocomial multiresistant Gram-negative bacilli: study of an outbreak in an intensive care unit. Ann. Intern. Med.1101989873881
35. Godard J., Guillaume C., Reverdy M. E., Bachmann P., Bui-Xuan B., Nageotte A., Motin J.Intestinal decontamination in a polyvalent ICU: a double-blind study. Intens. Care Med.161990307311
36. Bonten M. J. M., Gaillard C. A., de Leeuw P. W., Stobberingh E. E.Role of colonization of the upper intestinal tract in the pathogenesis of ventilator-associated pneumonia. Clin. Infect. Dis.241997309319
37. Pugin J., Auckenthaler R., Lew D. P., Suter P. M.Oropharyngeal decontamination decreases incidence of ventilator-associated pneumonia: a randomized, placebo-controlled, double-blind clinical trial. J.A.M.A.265199127042710
38. Rodriguez-Roldan J. M., Altuna-Cuesta A., Lopez A., Carrillo A., Garcia J., Lean J., Martinez-Pellus A. J.Prevention of nosocomial lung infection in ventilated patients: use of an antimicrobial pharyngeal nonabsorbable paste. Crit. Care Med.18199012391242
39. Vollaard E. J., Clasener H. A. L.Colonization resistance. Antimicrob. Agts. Chemother.381994409414
40. National Committee for Clinical Laboratory Standards. 1993. Methods for dilution antimicrobial susceptibility tests for bacteria that grow aerobically. Approved Standards M-A3, Villanova, PA.
Correspondence should be addressed to Marc J. M. Bonten, Department of Internal Medicine, Division of Infectious Diseases and AIDS, University Hospital Utrecht, Heidelberglaan 100, 3584 CX Utrecht, The Netherlands. E-mail: m.bonten

Presented in part at the 37th Interscience Conference on Antimicrobial Agents and Chemotherapy, Toronto, Canada, September 1997 (abstracts Nos. J-40 and J-42).


No related items
American Journal of Respiratory and Critical Care Medicine

Click to see any corrections or updates and to confirm this is the authentic version of record