Introduction

Stenotrophomonas maltophilia (S. maltophilia, formerly called Xanthomonas maltophilia or Pseudomonas maltophilia, first identified in 1958) is a non-fermentative gram-negative bacillus (NF-GNB) [1]. In critically ill patients colonization and infection with NF-GNB are frequent and associated with poor outcomes [2, 3]. S. maltophilia has been recognized as an emerging nosocomial pathogen [4, 5]. Among a variety of different infection sites, the most frequent infections caused by this pathogen are pneumonia and bacteremia [5].

The therapy of S. maltophilia infections is challenging because this pathogen is characterized by resistance to many antibiotics. This extended antibiotic resistance to extended-spectrum penicillins, third-generation cephalosporins, carbapenems, and aminoglycosides is due to a number of antibiotic resistance determinants including reduced outer membrane permeability, beta-lactamases and enzymes inactivating aminoglycosides as well as multidrug efflux pumps [6–8]. The bacteriostatic drug trimethoprim-sulfamethoxazole has been the antibiotic of choice for this pathogen [4]. Variable susceptibility rates have been reported for the quinolone ciprofloxacin [5]. In vitro studies revealed promising results regarding susceptibility of S. maltophilia to moxifloxacin [4, 5, 9, 10].

In oncology patients, studies have characterized risk factors for S. maltophilia colonization, infections and bacteremia [11–15]. In patients with cystic fibrosis, S. maltophilia has been identified as an emerging pathogen for infections of the respiratory tract [16, 17]. It has been shown that infections related to S. maltophilia are more frequent in patients with neutropenia and immunosuppression as well as burn patients and organ transplant recipients [11, 14, 18–22]. The majority of studies on S. maltophilia infections investigated S. maltophilia bacteremia and catheter-associated S. maltophilia infections [23–26]. However, despite the increasing frequency of NF-GNB infections in intensive care unit (ICU) patients, the role of S. maltophilia colonization or infection of the respiratory tract in these patients is not well investigated [27, 28].

This retrospective study was conducted in critically ill medical ICU patients with S. maltophilia colonization or infection of the respiratory tract. The aim of the study was to investigate clinical characteristics in these patients and antibiotic susceptibility of S. maltophilia as well as to identify risk factors for S. maltophilia infection (pneumonia) and ICU mortality.

Materials and methods

Study design, setting, and patients

This retrospective study was conducted in three medical ICUs (25 beds) of a German university hospital. We analyzed the medical records of all patients admitted to the medical ICUs between November 2005 and December 2009. We identified 64 patients with S. maltophilia isolated from the respiratory tract (S. maltophilia isolated from microbiologic culture from bronchoalveolar lavage (BAL) or tracheal aspiration). The study was approved by the local ethics committee. Need for informed consent was waived for this retrospective analysis.

Definitions

For definition of pulmonary infection (i.e. pneumonia) associated with S. maltophilia we used modified criteria proposed by the US Centers for Disease Control and Prevention. S. maltophilia-associated pneumonia was diagnosed by a positive microbiologic culture from BAL or tracheal aspiration accompanied by radiographic signs of pulmonary infection (presence of new or increasing infiltrates on chest radiograph) and at least two of the following clinical criteria of pulmonary infection: abnormal temperature (>38.3°C or <35.0°C), abnormal leucocyte count (leucocyte count >10,000/μL or increase in leucocyte count > 25% or leucocyte count <5,000/μL), or macroscopically purulent tracheal secretions. S. maltophilia colonization was defined as a positive microbiologic culture for S. maltophilia without the above-mentioned clinical signs for pulmonary infection/pneumonia.

Immunosuppression was defined as presence of neutropenia (neutrophil granulocytes < 1,500/μL), medical immunosuppression or HIV infection.

To quantify the severity of pulmonary dysfunction, the lung injury score (LIS) was calculated as reported previously [29].

Microbiological testing

Antibiotic susceptibility testing of S. maltophilia isolates was performed according to Clinical and Laboratory Clinical Standards Institute standards, using a 0.5 McFarland inoculum on Mueller-Hinton plates. The following disk diffusion breakpoints (< = resistant; > = susceptible) were used: polymyxin B 300 μg 11/12 mm, ceftazidime 30 μg 14/18 mm, ciprofloxacin 5 μg 15/21 mm, moxifloxacin 5 μg 17/21 mm, trimethoprim-sulfamethoxazole 1.25/23.75 μg 10/16 mm, and tigecycline 15 μg 14/19 mm. Antibiotic susceptibility testing was not always performed for all antibiotic agents. Therefore, data on susceptibility to moxifloxacin and trimethoprim-sulfamethoxazole are missing in one patient each, data on susceptibility to polymyxin B are missing in three patients, and data on susceptibility to tigecycline are missing in 17 patients.

Statistical analysis

All statistical tests were conducted two-sided in an explorative manner on a 5% level of significance. Analyses were performed using PASW statistics, version 18.0 (SPSS Inc., Chicago, Illinois, USA). Due to violations of the normal distribution continuous variables were compared using the nonparametric Mann-Whitney-U test. Categorical variables were analyzed using the chi-squared test or Fisher’s exact test when appropriate. Multivariate logistic regression analyses were performed to identify risk factors for S. maltophilia pneumonia and ICU mortality. Variables considered to be a risk factor in advance or showing a p-value <0.100 in the univariate analysis were entered into the multivariate model. Stepwise forward variable selection was used to define the final independent risk factors. Adjusted odds-ratios and 95% confidence intervals are reported for logistic regression analyses. To identify independent risk factors for S. maltophilia pneumonia, the following factors were included in multivariate regression analysis: age, neutropenia, medical immunosuppression, immunosuppression (as defined in the methods section), COPD, nicotine abuse, development of renal insufficiency, presence of endotracheal tube, tracheostomy, ICU mortality, hospital mortality, Simplified Acute Physiology Score II (SAPS II score), Sequential Organ Failure Assessment (SOFA) score, leucocyte count on the day of S. maltophilia isolation, and duration of mechanical ventilation before S. maltophilia isolation. Factors included in the multivariate binary logistic regression analysis regarding risk factors associated with ICU mortality were S. maltophilia-related pneumonia, gender, admission from another hospital, liver cirrhosis as reason for ICU admission, liver failure on ICU admission, renal failure during ICU stay, liver failure during ICU stay, SAPS II score, SOFA score, presence of endotracheal tube, and need for mechanical ventilation when S. maltophilia was isolated.

Survival rates were estimated according to Kaplan-Meier curves and compared between groups by log-rank tests. Descriptive statistics for categorical factors are given by absolute and relative frequencies. Descriptive statistics for continuous factors are given by median, range and interquartile range (IQR) while the latter equals the span between the 25th and 75th percentile.

Results

Patients’ characteristics

Sixty-four critically ill ICU patients with S. maltophilia isolated from the respiratory tract were enrolled in this study. Basic demographic data and patients’ characteristics are stated in Table 1.

Table 1 Demographic data and patients’ characteristics
Full size table

ICU treatment

The reasons for ICU admission and interventions during ICU treatment are shown in Table 2.

Table 2 Intensive care unit treatment
Full size table

S. maltophilia-related colonization or infection

The median time between ICU admission and first isolation of S. maltophilia from the respiratory tract was 10.0 days (range, 0–55 days; IQR, 2.5–14.5 days) (Table 2). Before the first isolation of S. maltophilia patients had been treated in the hospital for a median time period of 15.0 days (range, 2–77 days; IQR, 8.0–25.5 days) without differences between the patients with S. maltophilia infection or colonization (median, 15.5 days; range, 2–77 days; IQR, 7.5–30.5 days compared to median, 14.5 days; range, 2–62 days; IQR, 10.5–24.0 days, respectively; p = 0.823). At the time of first isolation of S. maltophilia from the respiratory tract, 36 patients (56%) fulfilled the criteria for the diagnosis of pulmonary infection/pneumonia, whereas the diagnosis of S. maltophilia colonization of the respiratory tract was made in 28 patients (44%). Serum levels of the laboratory infection parameter C-reactive protein (CRP) were significantly higher in patients with respiratory tract infection compared to patients with colonization of the respiratory tract (median, 13.1 mg/dL; range, 1.3–43.0 mg/dL; IQR, 5.6–20.4 mg/dL compared to median, 5.4 mg/dL; range, 0.8–29.6 mg/dL; IQR, 2.4–15.6 mg/dL, respectively; p = 0.029).

Data regarding pulmonary function and mechanical ventilation

Data regarding respiratory function and mechanical ventilation are shown in Table 3. A significantly higher LIS at the time of first S. maltophilia isolation was observed in patients with S. maltophilia-related pneumonia compared to patients with S. maltophilia colonization.

Table 3 Pulmonary function and mechanical ventilation
Full size table

Risk factors associated with S. maltophilia pneumonia

In univariate analysis, factors associated with S. maltophilia pneumonia (in contrast to S. maltophilia colonization of the respiratory tract) were: higher SOFA score on the day of first S. maltophilia isolation (p = 0.002), immunosuppression (p = 0.004), neutropenia (p = 0.006), higher SAPS II score on the day of first S. maltophilia isolation (p = 0.020) and higher serum CRP levels on the day of first S. maltophilia isolation (p = 0.029) (Tables 1 and 3). In multivariate binary logistic regression analysis, higher SOFA score on the day of first S. maltophilia isolation (p = 0.009; odds-ratio, 1.2; 95% confidence interval, 1.0–1.4) and immunosuppression (p = 0.014; odds-ratio, 4.9; 95% confidence interval, 1.4–17.4) were identified as independent risk factors associated with S. maltophilia pneumonia.

Outcome of patients with S. maltophilia colonization and infection

Patient outcome is presented in Table 4. Kaplan-Meier analysis demonstrated a significantly higher ICU mortality within a follow-up of 28 days in patients with S. maltophilia-related pneumonia (p = 0.040; Fig. 1). Patients with S. maltophilia-associated pneumonia had a significantly higher hospital mortality than patients with S. maltophilia colonization of the respiratory tract (p = 0.018) (Table 4).

Table 4 Intensive care unit and hospital mortality
Full size table
Fig. 1
figure 1

Kaplan-Meier survival curve for patients with Stenotrophomonas maltophilia pneumonia and pulmonary colonization. Kaplan-Meier survival curve for patients with Stenotrophomonas maltophilia (S. maltophilia) pneumonia (grey line) and pulmonary colonization (black line); p = 0.040

Full size image

Risk factors for ICU mortality

In univariate analysis, ICU mortality was significantly associated with higher SAPS II score on the day of first S. maltophilia isolation (p < 0.001), higher SOFA score on the day of first S. maltophilia isolation (p < 0.001), development of renal failure during the ICU stay (p = 0.006), need for mechanical ventilation on the day of first S. maltophilia isolation (p = 0.007), cerebral/neurological dysfunction during the ICU stay (p = 0.008), hepatic dysfunction on ICU admission (p = 0.010), development of liver failure during the ICU stay (p = 0.011), higher LIS on the day of first S. maltophilia isolation (p = 0.012), lower pO2/FiO2 ratio on the day of first S. maltophilia isolation (p = 0.025), and female gender (p = 0.029) (Tables 1, 2 and 3). Multivariate binary logistic regression analysis identified SOFA score on the day of first S. maltophilia isolation (p = 0.001; odds-ratio, 1.4; 95% confidence interval, 1.1–1.6) and development of renal failure in the course of the ICU stay (p = 0.021; odds-ratio, 4.8; 95% confidence interval, 1.3–18.4) as independent risk factors associated with ICU mortality.

Polymicrobial infections

In 33 (52%) of the 64 patients, other bacteria were isolated from the respiratory tract at the same time as S. maltophilia or within 7 days after the first isolation of S. maltophilia. The most frequently isolated co-pathogens were Enterococcus species (17 patients), Coagulase-negative Staphylococcus species (10 patients), Escherichia coli (7 patients), Viridans group Streptococci (6 patients), Pseudomonas aeruginosa (4 patients), Enterobacter cloacae (4 patients), Klebsiella pneumoniae (3 patients), Staphylococcus aureus (3 patients), Enterococcus faecium (3 patients) and Serratia marcescens (2 patients). Presence of a polymicrobial infection was significantly associated with hospital mortality (mortality 19/33 patients (58%) in patients with polymicrobial infection compared to 10/31 patients (32%) in patients without polymicrobial infection; p = 0.042).

Antibiotic treatment

Fifty-eight (91%) of the 64 patients had been treated with at least one antibiotic agent before S. maltophilia was isolated from the respiratory tract. Thirty-seven patients (58%) had been exposed to carbapenems prior to first isolation of S. maltophilia. Extended-spectrum penicillins (piperacillin/sulbactam) had been used in 34 patients (53%). Macrolids had been administered in 32 patients (50%). Twenty-five patients (39%) had been treated with vancomycin, 11 (17%) with second and third generation cephalosporins and 11 (17%) with ciprofloxacin. Only two patients (3%) had been treated with moxifloxacin before the first isolation of S. maltophilia from the respiratory tract.

In patients exposed to carbapenems prior to isolation of S. maltophilia from the respiratory tract, S. maltophilia-related pneumonia was significantly more often diagnosed compared to patients without exposure to carbapenems (pneumonia in 25/37 patients (68%) in patients exposed to carbapenems compared to 11/27 patients (41%) in patients not exposed to carbapenems; p = 0.033).

Antibiotic susceptibility

Antibiotic susceptibility is depicted in Fig. 2. In 84% of cases S. maltophilia was susceptible to moxifloxacin at the time of the first isolation of S. maltophilia from the respiratory tract. In 24 patients with S. maltophilia initially susceptible to moxifloxacin, S. maltophilia was isolated at least once more from the respiratory tract at a later time point. In 50% of cases S. maltophilia was then no longer susceptible to moxifloxacin.

Fig. 2
figure 2

Antibiotic susceptibility of Stenotrophomonas maltophilia. Antibiotic susceptibility of 64 Stenotrophomonas maltophilia isolates from the respiratory tract of 64 critically ill patients. Antibiotic susceptibility is depicted in % of isolates that were resistant, intermediate or sensitive to the antibiotic agent. Antibiotic agents: CEF ceftazidime, CIP ciprofloxacin, PXB polymyxin B, MOX moxifloxacin, TIG tigecycline, SXT trimethoprim-sulfamethoxazole

Full size image

The rate of antibiotic susceptibility to ciprofloxacin was significantly lower in patients treated with ciprofloxacin prior to first isolation of S. maltophilia (4/11 patients (36%)) compared to patients not treated with ciprofloxacin before S. maltophilia isolation (40/53 patients (75%)) (p = 0.027).

Discussion

Sixty-four critically ill patients with S. maltophilia pneumonia or pulmonary colonization were included in this retrospective cohort study.

In the present study the patients’ characteristics are different compared to previously presented studies that were conducted in interdisciplinary ICUs. The medical ICU patients included in our study showed higher SAPS II score, higher Acute Physiology and Chronic Health Evaluation II (APACHE II) score, and the duration of the ICU stay was longer compared to previous data [27, 28].

In multivariate binary logistic regression analysis, higher SOFA score on the day of first S. maltophilia isolation and immunosuppression were identified as independent risk factors for the development of S. maltophilia pneumonia. These findings are in accordance with previous data showing that immunosuppression and a debilitated clinical state are associated with S. maltophilia infections [5, 22].

Although the over-all ICU mortality rate was not significantly different between patients with S. maltophilia pneumonia or colonization, patients with pneumonia caused by S. maltophilia showed a significantly higher ICU mortality during a 28-day follow-up and higher in-hospital mortality. In accordance, in previous studies S. maltophilia infection has been shown to be associated with bad outcomes and high mortality rates [27, 30].

Mechanical ventilation was needed in 51 patients (80%) in our study. Although more patients with S. maltophilia pneumonia were mechanically ventilated in the present study compared to patients with colonization of the respiratory tract (86% vs. 71%), that difference did not reach statistic significance. However, in our study a significantly higher LIS score and significantly higher serum CRP levels at the time of first S. maltophilia isolation were observed in patients with S. maltophilia pneumonia compared to patients with S. maltophilia colonization. The high proportion of patients requiring mechanical ventilation reflects the critical illness of the enrolled patients and is probably due to the fact that this study focused on patients with pulmonary S. maltophilia colonization and infection. However, the proportion of intubated and mechanically ventilated patients is comparable to that in the studies by Nseir et al. and Barchitta et al. both investigating S. maltophilia-associated infections in intensive care patients [27, 31]. As expected, in other studies characterizing both normal ward and ICU patients with S. maltophilia bacteremia the proportion of mechanically ventilated patients was markedly lower [24, 26, 32]. In our study the duration of mechanical ventilation before and after the isolation of S. maltophilia was not significantly different in patients with S. maltophilia pneumonia compared to patients with S. maltophilia colonization of the respiratory tract. In previous studies duration of mechanical ventilator support was revealed as a risk factor for the development of pneumonia and acquisition of S. maltophilia in general [28, 31].

In one-third of the patients in our study, bacterial co-pathogens were isolated from the respiratory tract when S. maltophilia colonization/infection was diagnosed. This frequency of polymicrobial infections is in accordance with previously reported data in patients with S. maltophilia pneumonia and S. maltophilia bacteremia [11, 24, 32, 33]. One study investigating the impact of positive S. maltophilia blood cultures in a tertiary care hospital in Singapore revealed markedly higher rates of polymicrobial bacteremia (77%) [34].

In accordance with previous data, the majority of patients (91%) in our study had been treated with at least one antibiotic prior to isolation of S. maltophilia. Several studies have demonstrated an association of exposure to broad-spectrum antibiotics with S. maltophilia infection [5, 26, 35]. In our study more patients (nearly 60%) had been treated with carbapenems before S. maltophilia was isolated compared to previous studies [26, 35]. The use of carbapenems has been previously shown to be associated with acquisition of S. maltophilia in several studies [35, 36]. In contrast, some studies did not identify exposure to carbapenems as a risk factor for S. maltophilia infection, whereas the use of broad-spectrum antimicrobial drugs other than carbapenems was associated with acquisition of S. maltophilia [11, 28, 37]. It has been hypothesized that exposure to broad-spectrum antimicrobial drugs in general might be more important than exposure to any single antibiotic [5]. In our study, S. maltophilia pneumonia was significantly more often diagnosed in patients who had been exposed to carbapenems compared to patients without exposure to carbapenems.

S. maltophilia usually shows antibiotic resistance to extended-spectrum penicillins, carbapenems, and aminoglycosides [6, 8]. In accordance with previous data, almost all S. maltophilia isolates in our study (97%) were susceptible to trimethoprim-sulfamethoxazole [26, 35, 38, 39]. However, there is evidence that the resistance rates to this drug are increasing [5, 40]. Besides trimethoprim-sulfamethoxazole, the most effective antibiotics against S. maltophilia in our study were the glycocycline antibiotic tigecycline, the quinolone moxifloxacin, and polymyxin B. Regarding treatment of S. maltophilia with ciprofloxacin, variable antibiotic susceptibility has been reported [5]. Only 69% of tested S. maltophilia isolates tested susceptible to the quinolone ciprofloxacin in our study. In accordance with previous data only about 50% of S. maltophilia isolates were susceptible to ceftazidime [5, 28].

This study was conducted retrospectively in a limited number of patients in three medical ICUs of a single university hospital in Germany. Therefore, the results and especially the data on antibiotic susceptibility of S. maltophilia may not apply for other hospitals and other geographic areas. Since molecular typing of S. maltophilia isolates is not performed in clinical routine we can not provide data on potential transmission of S. maltophilia isolates in our study. Moreover, our study is lacking a matched control group of patients without S. maltophilia colonization/infection.

In conclusion, the present study demonstrates that the severity of multiple organ dysfunction syndrome (reflected by a higher SOFA score) and immunosuppression are independent risk factors for S. maltophilia pneumonia in medical ICU patients with S. maltophilia isolated from the respiratory tract. S. maltophilia was previously considered to be a pathogen with limited pathogenicity. However, according to our data, patients with S. maltophilia pneumonia have a significantly higher ICU mortality within a follow-up of 28 days, hospital mortality and LIS compared to patients with S. maltophilia colonization. Regarding treatment of S. maltophilia infections, prospective controlled trials are needed to obtain in vivo data on the effectiveness of different antimicrobial agents.