Fax: (713) 792-8233.
Risk factors for infections with multidrug-resistant Stenotrophomonas maltophilia in patients with cancer†
Version of Record online: 8 MAY 2007
Copyright © 2007 American Cancer Society
Volume 109, Issue 12, pages 2615–2622, 15 June 2007
How to Cite
Ansari, S. R., Hanna, H., Hachem, R., Jiang, Y., Rolston, K. and Raad, I. (2007), Risk factors for infections with multidrug-resistant Stenotrophomonas maltophilia in patients with cancer. Cancer, 109: 2615–2622. doi: 10.1002/cncr.22705
Presented in part as an abstract at the 43rd Annual Meeting of the Infectious Diseases Society of America, San Francisco, California, October 6–9, 2005.
- Issue online: 4 JUN 2007
- Version of Record online: 8 MAY 2007
- Manuscript Accepted: 15 FEB 2007
- Manuscript Revised: 7 FEB 2007
- Manuscript Received: 7 DEC 2006
- Stenotrophomonas maltophilia infection;
- drug resistance;
- trimethoprim/sulfamethoxazole (TMP/SMX)
Stenotrophomonas maltophilia is responsible for an increasing number of infections, especially in hospitalized patients. Therapy options are limited and trimethoprim/sulfamethoxazole (TMP/SMX) is often the main treatment option for this infection. In the current study, the risk factors were determined for the emergence of multidrug-resistant (MDR) S. maltophilia.
A case-control study was conducted to determine risk factors for the development of MDR S. maltophilia in cancer patients. The case group was composed of patients treated at the University of Texas M. D. Anderson Cancer Center for MDR S. maltophilia between 1996 and 2004 (n = 54). Two control groups were used: patients at comparable risk for S. maltophilia (C-controls) and patients with S. maltophilia infection that was susceptible to TMP-SMX and at least 2 other antibiotics (ciprofloxacin, ceftazidime, amikacin, and ticarcillin/clavulanate) (S-controls).
When compared with C-controls, prior use of carbapenems or quinolones and admission to an intensive care unit within 30 days of isolation of the pathogen were found to be independently associated with MDR S. maltophilia infection (P < .02), as was an increased overall mortality rate (P = .04). When compared with S-controls, risk factors were history of S. maltophilia infection during the prior year and prior use of TMP-SMX (P = .015).
Judicious use of TMP-SMX, carbapenems, and quinolones is necessary to control the risk for MDR S. maltophilia infection. Cancer 2007. © 2007 American Cancer Society.
Stenotrophomonas maltophilia has emerged, with increasing incidence, as an important opportunistic nosocomial pathogen in patients with cancer and other critical conditions.1–3 It is most often associated with pneumonia, but it may also cause a wide spectrum of diseases.3, 4 Several studies from the University of Texas M. D. Anderson Cancer Center in Houston have shown that S. maltophilia infections were associated with a severe clinical course and high mortality rates.5–7 This increase in incidence of this infection may be a consequence of selective pressure caused by the overuse of broad-spectrum β-lactam antibiotics.8–10
A prominent feature of S. maltophilia is its resistance to multiple antibiotics, including β-lactam agents such as the carbapenems, cephalosporins, and aminoglycosides.11 Treating S. maltophilia infections is a challenge because of the organism's resistance to most antimicrobial agents and the variable antimicrobial susceptibility of different strains of the organism.12–15
The antimicrobial agent most frequently active against S. maltophilia is trimethoprim-sulfamethoxazole (TMP-SMX).4, 5, 16, 17 However, several studies have reported resistance to TMP-SMX.1, 13, 17–21 A previous surveillance study22 found that the rates of resistance to TMP-SMX ranged from 2% in Canada and Latin America to 10% in Europe, with rates ranging from 3% to 19% during the study period. Another agent, ticarcillin/clavulanate, has been found to be active in 60% to 70% of cases. Quinolones, aminoglycosides, and ceftazidime are also used, but are less active.17
We sought to identify risk factors associated with acquiring infections caused by multidrug-resistant (MDR) S. maltophilia through a case-control study of patients with cancer who had had MDR S. maltophilia infections, as well as examine mortality associated with the infection.
MATERIALS AND METHODS
We identified retrospectively all patients who had S. maltophilia infections between January 1996 and September 2004 at the University of Texas M. D. Anderson Cancer Center through searching the microbiology laboratory and the Infection Control Surveillance records. We performed 2 simultaneous case-control analyses after the Institutional Review Board approved the study.
Cases and Controls
We used the criteria established by the U.S. Centers for Disease Control and Prevention to determine the cases for the study.23 Patients with colonization were excluded from the study. Stenotrophomonas maltophilia was considered MDR if it was resistant to TMP/SMX and to at least 2 other additional antibiotics that are often active against S. maltophilia: ciprofloxacin, ceftazidime, amikacin, or ticarcillin/clavulanate. We used 2 control groups with the ratio of cases to controls as 1:2 for each of the 2 control groups. The control groups were selected from patients who had been in the same place at the same time as the other cases. The first control group was a Comparable Risk Group (C-control) and consisted of 108 patients who were matched to the case patients on the basis of their underlying cancer, period of hospitalization, and duration of hospitalization; unlike the case patients, however, these patients were not infected or colonized by S. maltophilia. The other control group was a microbiologically susceptible group (S-control) and consisted of 108 patients who were matched to the case patients based on the site of infection. These patients had infections caused by S. maltophilia, but the organisms were susceptible to TMP-SMX and to at least 2 of the four antibiotics often active against S. maltophilia: ciprofloxacin, ceftazidime, amikacin, and ticarcillin/clavulanate. Antimicrobial susceptibility testing to determine MDR was performed according to National Nosocomial Infections Surveillance System guidelines and with the E-test for antimicrobial susceptibility.
Data Collected and Risk Factors Investigated
Data were collected from patient charts as well as computerized pharmacy, Infection Control, and microbiology databases. Information was collected regarding age; sex; cancer diagnosis; comorbid conditions; and history of surgery, radiation therapy, and chemotherapy. The following were considered comorbid conditions or risk factors if they had occurred within 30 days before the isolation of S. maltophilia: prior hospitalization before admission for the infection, a stay in an intensive care unit (ICU), neutropenia and its duration, severity of illness based on the Acute Physiology and Chronic Health Evaluation II (APACHE II) classification system, and history of steroid therapy. For the cases and control group S, data were collected on antibiotics given within the 90 days before the date of isolation of S. maltophilia; for control group C, data were collected on antibiotics given within the 90 days before the matched hospitalization date, including any days during the current hospitalization before the case patient's culture became positive.
Data were also collected on all patients to evaluate treatment success or failure and overall mortality rates. Treatment failure was defined as a clinical or microbiologic lack of response or recurrence of the infection or death of the patient resulting from the S. maltophilia infection.
Cases were compared with each of the 2 controls using the chi-square test or the Fisher exact test to assess the univariate association of categoric risk factors for acquiring MDR S. maltophilia. Those categoric risk factors were sex, underlying disease, chemotherapy, steroid use, and neutropenia within 30 days of isolation of S. maltophilia, and antibiotic use within 90 days of its isolation.
We used the Shapiro-Wilks test to determine the normality of the continuous risk factors, which were age at admission, APACHE II score, and duration of neutropenia. The 2-sample Student t-test or Wilcoxon rank sum test was used to compare the measures between case and control patients. If the distribution of a continuous variable departed significantly from normality we repeated the comparison using the nonparametric Wilcoxon rank-sum test.
Logistic regression models were used for further analyses. First, for factors that had a P-value < .10 on the 2-sample Student t- or Wilcoxon rank sum test, we performed a univariate analysis for each factor alone. We then performed multiple logistic models: initially, a full multiple logistic regression analysis, and then a backward selection method to determine the final (reduced) multiple model so that all factors remaining in the model were statistically significant at a 5% level (P < .05). Data analyses were performed using SAS software (version 9.1; SAS Institute Inc., Cary, NC).
We identified a total of 54 patients who had had infections caused by MDR S. maltophilia during the study period (case patients) and we included 108 control patients in each of the 2 control groups (C-control and S-control).
By univariate analysis, when cases were compared with C-control (patients with comparable risk factors but with no infection), the following factors were found to be significantly associated with acquiring MDR S. maltophilia infection: the use of any antibiotics within 90 days before the date of the S. maltophilia culture (85.2% vs 68.5%, respectively, for case and control patients; P = .023), specifically carbapenems (46.3% vs 25.0%; P = .006) and quinolones (66.7% vs 43.5%; P = .006). Another risk factor identified was ICU stay within 30 days before the date of the culture (29.6% vs 8.3%; P = .0004). The overall mortality rate was also significantly associated with the MDR infection (23.1% vs 10.9%; P = .046) (Table 1).
|Characteristics||Cases (n = 54)||Controls (n = 108)||P|
|Median age, y (range)||55||(20–80)||60||(18–91)||.48|
|Chemotherapy in prior mo||18||(33.3)||38||(35.2)||.82|
|Neutropenia in prior mo||15||(27.8)||19||(17.6)||.132|
|Steroid use in prior mo||24||(44.4)||40||(37)||.362|
|ICU stay in prior mo||16||(29.6)||9||(8.3)||.0004|
|Hospitalization in prior mo||11||(20.4)||36||(33.3)||.09|
|Median APACHE II score (range)||12||(6–25)||12||(4–25)||.67|
|Median duration of hospitalization, d (range)||14||(1–88)||13||(1–75)||.17|
|Antibiotic use in prior 90 d||46||(85.2)||74||(68.5)||.02|
When cases were compared with S-control (patients whose S. maltophilia infections were susceptible to antibiotics), the following risk factors were found to be significantly associated with acquiring MDR S. maltophilia infection: prior ICU stay (29.6% vs 13.0% for cases and controls, respectively; P = .01), history of S. maltophilia infection within the prior year (25.9% vs 10.2%; P = .009), and TMP-SMX use during the prior 90 days (35.2% vs 20.4%; P = .041). However, S-control patients were more likely than the cases to have undergone chemotherapy during the month before the infection (50.0% vs 33.3%; P = .044), to have had a prior hospitalization during the month before the infection (45.4% vs 20.4%; P = .002), and tended to have been treated less often with cephalosporins within the 90 days before the infection (25.9% vs 35.2%; P = .07). The overall mortality rate was comparable between the cases and the S-control patients (22.2% vs 17.3%; P = .45), and response to treatment was also comparable between these 2 groups (64% vs 76.0%; P = .12) (Table 2).
|Characteristics||Cases (n = 54)||Controls (n = 108)||P|
|Median age (range)||55||(20–80)||59||(16–82)||.79|
|Chemotherapy in prior mo||18||(33.3)||54||(50)||.04|
|Neutropenia in prior mo||15||(27.8)||29||(26.9)||.90|
|Steroid use in prior mo||24||(44.4)||60||(55.6)||.18|
|ICU stay in prior mo||16||(29.6)||14||(13)||.01|
|Hospitalization in prior mo||11||(20.4)||49||(45.4)||.002|
|Median APACHE II score (range)||12||(6–25)||11||(6–25)||.38|
|Median duration of hospitalization, d (range)||14||(1–88)||19||(2–187)||.98|
|S. maltophilia infection in prior y||14||(25.9)||11||(10.2)||.009|
|S. maltophilia isolation site||.99|
|Radiation in prior mo||6||(11.1)||3||(2.8)||.0611|
|Surgery in prior mo||7||(13)||10||(9.3)||.47|
|Antibiotic use in prior 90 d||46||(85.2)||92||(85.2)||.99|
|Treatment success||32 (/50)||(64)||73 (/96)||(76)||.12|
|Recurrence||12 (/53)||(24.5)||22 (/106)||(20.8)||.59|
Multiple Logistic Regression Analyses
Using C-control (Table 3) the following variables were found to be independently associated with MDR S. maltophilia infection: prior treatment with carbapenems within 90 days before the date of culture (odds ratio [OR] of 4.70; 95% confidence interval [95% CI], 1.63–13.50 [P = .004]), prior use of quinolones within 90 days before the date of culture (OR of 3.19; 95% CI, 1.25–8.14 [P = .015]), and an ICU stay within the 30 days before the date of culture (OR of 5.98; 95% CI, 1.72–20.82 [P = .005]). Prior hospitalization within 30 days before the current admission was less likely to be associated with MDR S. maltophilia infection (OR of 0.20; 95% CI, 0.06–0.64 [P = .007]). In the multiple logistic regression analysis using S-control (Table 4) the following variables were found to be independently associated with MDR S. maltophilia infection: history of infection with S. maltophilia within the previous year (OR of 3.35; 95% CI, 1.29–8.74 [P = .013]) and prior treatment with TMP-SMX within 90 days before the date of culture (OR of 3.46; 95% CI, 1.46–8.18 [P = .005]). Prior use of cephalosporin and prior hospitalization were less likely to be associated with MDR S. maltophilia infection (Table 4).
|Characteristics||No.||Final multiple model|
|Odds ratio||95% CI||P|
|Carbapenems in prior 90 d||.0041|
|Quinolones in prior 90 d||.0151|
|ICU stay in prior 30 d||.005|
|Hospitalization in prior 30 d||.0068|
|Characteristics||No.||Final multiple model|
|Odds ratio||95% CI||P|
|Cephalosporins in prior 90 d||.0167|
|TMP-SMX in prior 90 d||.0047|
|S. maltophilia infection in prior y||.0132|
|Hospitalization in prior 30 d||.0022|
In this study, we found that a number of risk factors are associated with the acquisition of MDR S. maltophilia infections, including the prior use of TMP-SMX, carbapenems, and quinolones, as well as prior ICU stay and prior infection with S. maltophilia.
The unique aspect of our study was that we used 2 control groups: 1 consisting of patients who had susceptible S. maltophilia infections and the other consisting of patients who were at high risk for the infection but who did not have it. We used these 2 control groups because there has been some concern that the conventional method for selecting control patients in case-control analyses is not sufficient, and might lead to an overestimate the impact of antibiotic use as a risk factor. For example, in a review of case-control studies that analyzed the risk factors for infections resistant to antibiotics, Harris et al.24 noted that biased estimates might result if patients with antibiotic-susceptible microorganisms were used as controls. And in a study of risk factors for nosocomial infections caused by antibiotic-resistant bacteria, Paterson25 noted that the optimal control group would be hospitalized patients with the same potential exposure to the antibiotic-resistant bacteria that the case patients had. Thus, not only would the control patients be hospitalized during the same period as the case patients, but they would also be hospitalized for at least the same duration as the case patient: from the time of admission of the case patient to the first isolation of the resistant organism in question.25
S. maltophilia isolates are intrinsically resistant to many commonly used antimicrobial agents. They produce diverse drug-hydrolyzing enzymes such as L1, a zinc-dependent metalloenzyme, and L2, a cephalosporinase, which are able to destroy important β-lactam antibiotics such as carbapenems and cephalosporins.19, 26, 27S. maltophilia isolates also quickly develop resistance to fluoroquinolones by mutations in outer-membrane proteins. Recently, Zhang et al.28 demonstrated the involvement of efflux mechanisms in acquired MDR in S. maltophilia.
The use of carbapenems has also been established as a risk factor for S. maltophilia infection. In a previous case-control study conducted at the University of Texas M. D. Anderson Cancer Center,29 the case patients infected by S. maltophilia had been treated with imipenem 10 times more frequently than the uninfected control patients. In the current study, when MDR S. maltophilia cases were compared with control patients who were at a comparable risk, it is possible that the use of carbapenems and quinolones contributed to the emergence of S. maltophilia at our hospital. In a more recent study also conducted at the University of Texas M. D. Anderson Cancer Center, Ohmagari et al.30 demonstrated that treatment with carbapenems for more than 7 days was associated with the emergence of MDR Pseudomonas aeruginosa.
A number of studies have pointed to an association between ICU stays and the subsequent development of infections with MDR organisms. In 2 studies17, 31 ICU stay was found to play a role in increased incidence, morbidity, and mortality due to S. maltophilia. One was a prospective, multicenter observational study by Muder et al.,17 with 91 cases of bacteremia due to S. maltophilia; the other was a study by Wang et al.31 assessing the characteristics of 50 episodes of S. maltophilia bacteremia in 48 patients. In the current study, an ICU stay within 30 days before the isolation of S. maltophilia was found to be significantly associated with the emergence of MDR S. maltophilia and had the highest OR of all the risk factors. The reasons behind this emergence were likely the use of multiple antibiotics, the presence of multiple comorbid conditions, and the impaired immunity.
Our study also pointed to the role of quinolones in causing the increased incidence of S. maltophilia infections. Quinolones are among the most commonly used antibiotics for this infection. In a previous case-control study, Spanik et al.32 found that the prophylactic use of quinolones in patients with cancer had a role in causing bacteremia due to MDR gram-negative bacilli in neutropenic cancer patients, including S. maltophilia.
The use of TMP-SMX has been identified as a potential risk factor for the emergence of TMP-SMX-resistant S. maltophilia. Vartivarian et al.13 observed that the overall usage of antibiotics, therapeutically and particularly prophylactically, in patients with cancer affected the overall susceptibility patterns of S. maltophilia isolates and that the same isolates became more susceptible to TMP-SMX when the usage of that antibiotic decreased in the hospital. The rates of resistance to TMP-SMX increased when it was used for selective decontamination of the bowel, but those rates decreased when ciprofloxacin was substituted; however, the rates of resistance to ciprofloxacin increased with its use.13
Conversely, however, we have to point out that in our study TMP-SMX was identified as a risk factor of MDR S. maltophilia infection only from the comparison that chose patients infected with susceptible S. maltophilia as the control group (S-control), not from the comparison of case-patients vs patients with comparable risk for the infection but without it (C-control). Therefore, it should be noted that the effect measure of TMP-SMX in our study is somewhat biased.24
Ohmagari et al.30 found that a history of a P. aeruginosa infection during the preceding year was independently associated with the emergence of MDR P. aeruginosa infections. In addition, a case series by Harris et al.33 showed that MDR P. aeruginosa emerged in a stepwise manner after a patient's exposure to antipseudomonal antibiotics and caused adverse outcomes.33 We believe that in our study a similar mechanism likely accounted for the increased incidence of MDR S. maltophilia infections in patients infected with the TMP-SMX-susceptible strain the preceding year.
In the current study, prior cephalosporin use was less likely associated with MDR S. maltophilia infection. This could be explained by the other findings of the study. According to multivariate analysis, cases of MDR S. maltophilia infections were more likely to have experienced a prior S. maltophilia infection during the preceding year and were more likely to have received prior TMP/SMX, which may be attributed to less use of cephalosporins in this group. Prior hospitalization was also found to be less likely associated with MDR S. maltophilia infections. The reason for its ‘protective’ effect in our study remains to be investigated.
In the current study, the most common source of isolation of S. maltophilia was the respiratory system, a finding noted in other studies as well, including the SENTRY antimicrobial surveillance program.22 In that study, pneumonia-causing S. maltophilia isolates were 4 times more prevalent than bloodstream infection isolates of S. maltophilia. The isolation rates by geographic area per year in patients with pneumonia varied from 1.2% to 4.8%.22 Other studies34, 35 have observed that S. maltophilia pneumonia can be complicated by bacteremia, septic shock, and multiple-organ-dysfunction syndrome.
Outcomes for patients with S. maltophilia infections are also associated with various risk factors. In a study of 99 cases, Morrison et al.36 observed a greater likelihood of death in patients in whom the isolates were polymicrobial (50%) than in patients with pure S. maltophilia isolates (30%). Stenotrophomonas maltophilia is often part of a polymicrobial infection,16 and in fact 46% of the S. maltophilia infections in our study were polymicrobial. In the patients infected with MDR S. maltophilia the overall mortality rate was significantly higher than that in the patients in control group C. In a multivariate analysis study by Vartivarian et al.37 it was shown that the presence of severe shock at onset of infection and delay in appropriate therapy were significant predictors of poor outcome. In what to our knowledge is 1 of the few studies that evaluated morbidity and mortality rates in patients with TMP-SMX-resistant S. maltophilia infections, Tsiodras et al.38 reported a crude mortality rate of 20% in the patients with a microbiologically confirmed infection, a rate similar to that seen in our study. According to Muder et al.17 survival rates tend to be higher in patients with S. maltophilia infections who are treated with more than 1 active agent than in those treated with single agents. The use of combination therapy with colistin, rifampin, and aztreonam has been found efficacious.39, 40
Our study had several limitations. First, it is retrospective in nature. Second, here have been changes in antimicrobials in the formulary or changes in the infection control practice and procedures over the 8 years. Third, our study lacks a comparison of patients infected with susceptible MDR S. maltophilia (S-control) vs uninfected control patients (C-control) that would be helpful for us to have a better knowledge about the risk factors of MDR S. maltophilia infection.
In conclusion, the current study investigated the risk factors associated with the acquisition of MDR S. maltophilia infections in patients with cancer, which is a patient population that is highly likely to be exposed to prophylactic antibiotics, such as quinolones and TMP/SMX, and empiric antibiotics, such as carbapenems and quinolones. We identified the use of carbapenems, quinolones, TMP/SMZ, prior ICU stay, and history of prior infection with S. maltophilia as risk factors. Judicious prophylactic and empiric use of quinolones, TMP/SMX, and carbapenems is necessary, especially in this high-risk cancer patient population.
We thank Cai Wu and Frank Hung in the Department of Pharmacy and Suzy Wallace in the Department of Biostatistics and Applied Mathematics for help in data collection and matching.
- 4Susceptibilities of non-Pseudomonas aeruginosa gram-negative nonfermentative rods to ciprofloxacin, ofloxacin, levofloxacin, D-ofloxacin, sparfloxacin, ceftazidime, piperacillin, piperacillin-tazobactam, trimethoprim-sulfamethoxazole, and imipenem. Antimicrob Agents Chemother. 1996; 40: 772–775., , , .
- 22Emerging importance of multidrug-resistant Acinetobacter species and Stenotrophomonas maltophilia as pathogens in seriously ill patients: geographic patterns, epidemiological features, and trends in the SENTRY Antimicrobial Surveillance Program (1997–1999). Clin Infect Dis. 2001; 15;32 suppl 2( ): S104–113., , , et al.
- 37Outcome of Stenotrophomonas maltophilia infections in cancer patients. In: Program and Abstracts of the 37th Annual Meeting of the Interscience Conference on Antimicrobial Agents and Chemotherapy (abstract J-92), September 28-October 1, 1997, Toronto, Ontario, Canada., , .