INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Pseudomonas aeruginosa is one of the three most common causes of pneumonia in patients with artificial airway (intubated or tracheostomized) (1) and the leading etiology among fatal episodes (2). In addition, this species is the most frequently isolated pathogen in ventilated patients with respiratory superinfections (3). A recent retrospective study (4) using sputum or blood cultures suggested that repeated episodes of P. aeruginosa infection were found in a high proportion of patients who survived a first episode of pneumonia caused by this organism.

Phenotypic markers that distinguish between strains of P. aeruginosa, such as pyocin production, pyocin susceptibility, serotyping, antibiotyping, and phage susceptibility have long been employed in the investigation of point source outbreaks of P. aeruginosa (5). More recently, the effectiveness of epidemiologic typing of microorganisms has been expanded by introducing the use of molecular analysis of microbial DNA. Use of these novel techniques in conjunction with conventional investigation could provide new insights into the current epidemiology of pneumonia in ventilated patients.

In pneumonia caused by P. aeruginosa, persistent isolation of strains from respiratory airways after several days of therapy seems a frequent finding (6). Consequently, our hypothesis was that recurrent episodes of pneumonia caused by P. aeruginosa in ventilated patients would be due predominantly to a reactivation of the previous strain (relapse) rather than to exogenous reinfection. Thus, the aim of this study was to determine whether ventilated patients presenting recurrent episodes of pneumonia caused by P. aeruginosa acquired new strains, or retained their prior clones despite appropriate antibiotic therapy. Once established, this information may contribute to improving our control efforts.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Patient Population and Bacterial Strains

In May 1993, a prospective study was begun at the Hospital de Sabadell to follow-up all ventilated patients who met strict criteria for pneumonia due to P. aeruginosa until death or hospital discharge. The study lasted 30 mo. Twenty-six of the patients were part of a study evaluating attributable mortality (7). Several clinical variables, including underlying disease, comorbidities, clinical signs, days of intubation, as well as antibiotic regimens and sensitivity patterns, were recorded in a standardized form. The APACHE II score for each patient was calculated (8) in the initial 24 h of ICU admission and when a diagnosis of pneumonia was made.

Definitions

The diagnosis of pneumonia was considered when new and persistent pulmonary infiltrates not otherwise explained appeared on chest X-rays. Moreover, at least two of the following criteria were also required: (1) fever  38° C; (2) leukocytosis  10,000 per mm³; and (3) purulent respiratory secretions. A pneumonia was considered ventilator-associated when its onset occurred after 48 h of mechanical ventilation and was judged not to have been incubating before starting mechanical ventilation (9). Fiberoptic bronchoscopic examination using a protected specimen brush or bronchoalveolar lavage was performed on each of these episodes within the first 12 h after the development of a new pulmonary infiltrate. A Pseudomonal etiology was considered when a significant count ( 1,000 CFU/ml in protected brush specimens or  10,000 colony-forming unit (CFU)/ml in bronchoalveolar lavage) of the species was isolated.

Combination therapy was defined as the use of two or more antibiotics to which the P. aeruginosa was susceptible in vitro, at the time that bronchoscopy was performed. All patients received appropriate combination antibiotic therapy based on susceptibility testing of the isolated microorganism and therapy was prolonged at least 2 wk after the bronchoscopic procedure. Adult respiratory distress syndrome (ARDS) was defined according to standard definitions (10). Clinical resolution was defined if the patient had complete resolution of all signs and symptoms of pneumonia along with improvement, or lack of progression, of all abnormalities on the chest radiograph (11).

Recurrent pneumonia caused by P. aeruginosa was defined as follows: (1) occurrence at least 72 h after clinical resolution; (2) positive bronchoscopic quantitative culture for P. aeruginosa; (3) evidence for a new infiltrate on the chest roentgenogram; (4) two of the following: (a) fever  38° C; (b) white blood cell (WBC) count  10,000/mm³; or (c) purulent respiratory secretions; and (5) absence of evidence of a new extrapulmonary source of infection.

Microbiologic Processing and PFGE

In patients in whom ventilator-associated pneumonia was suspected, bronchoscopy, processing of samples, cultures and identification of bacteria were performed as previously described (7). P. aeruginosa strains from bronchopulmonary specimens containing significant counts ( 1,000 CFU/ml in protected brush specimens or  10,000 CFU/ml in bronchoalveolar lavage) were collected and stored at 70° C. Episodes with other etiologic diagnoses were excluded. No follow-up cultures were performed after treatment ended. Antimicrobial susceptibility was determined using a commercial microdilution broth method (PASCO, Detroit, MI), following the manufacturer's recommendations. The minimal inhibitory concentration (MIC) interpretative standards for the susceptibility categories considered were those given by the National Committee for Clinical Laboratory Standards (NCCLS) (12). Moderately susceptible and resistant bacteria were both regarded as resistant in this study.

Genomic DNA for pulsed field gel electrophoresis (PFGE) was prepared as previously reported (13, 14). Chromosomal DNA plugs were incubated with XbaI (Pharmacia Biochemicals, Uppsala, Sweden). Restriction fragments were separated with an orthogonal field alternation electrophoresis apparatus, Gene Navigator (Pharmacia LKB Biotechnology, Sweden), through 1% agarose gel for 20 h run at a constant voltage of 250 V. The pulse time was 20, 10, and 2 s for 3, 15, and 2 h, respectively.

PFGE chromosomal fingerprints were compared, and a bar-code format translation of chromosomal fingerprint profiles was made with a Bio Image system (Millipore, Ann Arbor, MI). The similarities of the fragment length patterns between two strains were scored by the Dice coefficient (15).

To interpret the DNA fragment patterns generated by PFGE and to transform them into epidemiologically useful information, we used the guidelines proposed by Tenover and colleagues (16). In brief, isolates were designated genetically indistinguishable if their restriction patterns had the same number of bands and if these bands appeared to be the same size. Isolates were considered to be closely related or possibly related when the patterns of two of them differed by no more than three bands (i.e., a single genetic event) or by four to six bands (i.e., two independent genetic events), respectively. Finally, two isolates were considered unrelated if their PFGE patterns differed by changes consistent with three or more independent genetic events (generally seven or more band differences). The origin of a new isolate was considered as a reinfection when it was classified as unrelated. The remaining isolates were considered as relapses of the same strain.

Statistical Analysis

Continuous variables were compared with t tests and one-way analysis of variance (ANOVA) or the nonparametric counterpart (Mann-Whitney U test). A two-tailed Fisher exact test was used to compare differences between groups for discrete variables.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

During the study period, 37 intubated patients had P. aeruginosa pneumonia at the medical-surgical intensive care unit (ICU) of Hospital de Sabadell. Thirty-three (89.1%) survived the initial week after pneumonia diagnosis, of whom six (18.1%) had more than one episode caused by the same species. The main demographic data of the study population are shown in Table 1. No patients were neutropenic (< 4,000/ mm³) or had cystic fibrosis. Mean ± SD period between recurrences and the prior episode of pneumonia was 23.6 ± 16.6 d. One recurrence was identified (Figure 1, strain 2) while the prior episode was still being treated, after 9 d of ciprofloxacin (400 mg/8 h) plus amikacin (1 g/24 h). It was documented 7 d after presenting clinical resolution of the prior episode, in the form of severe hypoxemia along with clinical signs of recurrence (fever, 23,000 WBC/mm³, new radiologic infiltrate and a shift in respiratory secretions). Strain 2 was isolated from a bronchoscopic sample and the patient died 1 d later with refractory hypoxemia.

View larger version (17K): [in this window] [in a new window]   Figure 1.   PFGE patterns of P. aeruginosa strains from five ventilated patients with recurrent episodes of pneumonia. Serial sets of isolates are shown. Left column shows the DNA size markers. Other lines show serial P. aeruginosa isolates from five ventilated patients with recurrent episodes of pneumonia. Lanes 1, 2: Patient A; lanes 3-8: Patient B; lanes 9-11: Patient C; lanes 12-14: Patient D; lanes 15, 16: Patient E.

Patients with recurrent episodes had a significantly higher proportion of ARDS (83.3% versus 22.2%, p = 0.02) before recurrence. The other clinical and microbiologic variables evaluated, including persistence of the pathogen after 72 h of therapy or an initial high count (defined by one log above the classic threshold, indicating a large bacterial burden), did not distinguish between patients with recurrences and those without (Table ).

Two patients died as a direct consequence of the recurrent pulmonary infection and two others of complications related to their underlying disease. Two more were discharged alive from the ICU, but only one (Patient D) was discharged from the hospital. He carried a tracheostomy and developed two recurrences (60 d and 98 d after the first episode) when hospitalized in a general ward. Six months after the last episode (at home), he showed airway colonization by an unrelated strain of P. aeruginosa (PFGE pattern not shown).

The study population was constituted by six patients, and sets of multiple isolates were available from five of them (nine recurrences). Detailed information on susceptibility data for the first and subsequent strains of these five patients and prior antibiotic exposure is presented in Table 2. Sequential isolates from the remainder were not stored. The 16 isolates of P. aeruginosa were analyzed by chromosomal fingerprinting based on PFGE (Figure 1). Patients A (strains 1 and 2) and E (strains 15 and 16) had one recurrence due to closely related and indistinguishable isolates, respectively. Patients C (strains 9-11) and D (strains 12-14) had two recurrences due to indistinguishable isolates. Patient B had a first episode caused by two different clones of P. aeruginosa (3 and 4) followed by a second episode with strains 5 (indistinguishable isolate from strain 4) and 6 (unrelated strain). This patient developed a third episode caused by strain 7 (indistinguishable from strain 6) plus Stenotrophomonas maltophilia and a fourth episode caused by strain 8 (indistinguishable from strains 4 and 5). Most PFGE patterns from the five patients were different from one another (Figure 1).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Two previous retrospective studies (4, 6) have reported the incidence of recurrent pneumonia caused by P. aeruginosa in intubated patients, showing a considerable disparity in the incidence of this complication (3% versus 50%). However, phenotypic patterns provide limited information for identifying common clones. Antibiogram typing results should be interpreted with caution, since unrelated strains may exhibit the same antibiogram, and changes in susceptibility may occur during episodes of infection. The present study is unique in that it elucidates the precise epidemiologic relations determined by interpretation of chromosomal fingerprinting based on PFGE in this population. Similar methods have been used to determine the source of other bacterial species in the respiratory tract of immunocompromised patients (17, 18).

The definition of recurrent pneumonia was adapted from criteria previously defined by Silver and coworkers (4). We also required the presence of positive quantitative bronchoscopic cultures and a longer period after clinical resolution. Our definition resulted in a higher specificity. It should be noted that it was not necessary to be off antibiotics in order to define recurrence and that Patient A developed a fatal recurrence when he was still being treated with antibiotics. It may be questioned whether this episode truly represented a clinical recurrence or actually represented persistence of infection. The original episode of pneumonia presented a rapid clinical resolution with a fulminant return of clinical signs and a diagnostic culture that seems more consistent with the first interpretation. In spite of this consideration, this observation suggests that interventions based on longer duration of therapy probably represent a wrong approach to prevent clinical recurrences. On the other hand, Patient D developed clinical recurrences 60 d and 98 d after the first episode, when he was hospitalized in another hospital area. These findings suggest that development of clinical recurrences may be more associated with patient-related risk factors, such as interference with host defense mechanisms, rather than with pathogen-related factors.

Molecular typing of the different strains isolated from patients with multiple episodes of pneumonia due to P. aeruginosa differentiates between relapse or reinfection. If there is considerable reinfection as a result of ongoing transmission of an exogenous strain, traditional infection control measures should be reinforced. In contrast, if there are relapses, the focus of control should be based on a different policy. Although our study was not designed to evaluate persistent colonization because follow-up cultures were not performed when treatment ended, our observations suggest that relapse is the most important mechanism of clinical recurrence, at least in our institution. The six patients with clinical recurrences received appropriate combination therapy based on susceptibility testing of the isolated organism, suggesting that, even in absence of follow up cultures, relapse of the original strain occurred in spite of combination therapy with active agents. Patient D illustrates that relapse after a very prolonged period of resolution, even in other hospital areas, is possible. Interestingly, Niederman and associates (19) showed persistence of P. aeruginosa in the lower respiratory tract in up to 50% of patients with artificial airways, despite no clinical evidence of infection, and reported that all patients with persistent tracheal colonization had a concomitant pulmonary disease. Brewer and colleagues (6) found similar results in their Pseudomonas sp. ventilator-associated pneumonia patients. Data of Silver and coworkers (4) showed that patients with chronic obstructive pulmonary disease or chronic pulmonary disease were much more likely to develop recurrent episodes. Our results also stress the importance of the presence of lung injury to predispose to persistent colonization. Further studies should explore the molecular mechanisms that account for persistent colonization, since it should be possible to intervene in the factors mediating clonal persistence so as to eliminate such infections.

Some limitations of our investigation should be noted. The incidence of clinical recurrences may be undervalued, since some of the patients with nonrecurrent episodes had other episodes of pneumonia in which the bacteriologic diagnosis remained uncertain, in spite of the use of invasive procedures. Moreover, strain 6 seems to indicate that current diagnostic procedures were not sufficiently sensitive to the prior episode. This is also true of strain 4, which recurred in Patient B (strain 8) after a third episode of pneumonia in which it could not be isolated. These observations highlight the difficulty of identifying the responsible pathogen in ventilated patients with pneumonia and should be borne in mind for therapeutic decisions in patients with delayed resolution.

In addition, it was difficult to discern other significant distinctions between patients with or without clinical recurrences because of the size of the population involved. Other variables might prove to be significant in larger studies. Finally, a potential bias associated with selection of a specific population in our institution or with presence of an epidemic clone in others cannot be ruled out, a possibility that precludes the generalization of our findings. In spite of these limitations, our findings suggest that, in absence of important violation of barrier control measures, cross-infection is an infrequent cause of clinical recurrence in patients with artificial airways. Certainly, our conclusions are limited to P. aeruginosa, but this does not reduce the implications of our findings, because this etiology is the leading cause of recurrent pneumonia (3).

In summary, our study suggests that recurrent episodes of pneumonia caused by P. aeruginosa in patients with artificial airways occur due to relapses. Our findings support a shift in the current approach of prevention based on traditional measures to control transmission, suggesting that the focus of control should be on improved identification (and direct intervention) of persistent carriers.