- Split View
-
Views
-
Cite
Cite
Alessandra Micozzi, Mario Venditti, Monica Monaco, Alexander Friedrich, Fabrizio Taglietti, Stefania Santilli, Pietro Martino, Bacteremia Due to Stenotrophomonas maltophilia in Patients with Hematologic Malignancies, Clinical Infectious Diseases, Volume 31, Issue 3, September 2000, Pages 705–711, https://doi.org/10.1086/314043
- Share Icon Share
Abstract
Predisposing factors, clinical characteristics, and antimicrobial treatment of 37 hematology patients with Stenotrophomonas maltophilia bacteremia who were seen at the department of hematology of the University La Sapienza (Rome) from 1987 to 1996 were evaluated. The results were compared with a control group of patients with Pseudomonas aeruginosa bacteremia. Profound neutropenia was more prolonged in the S. maltophilia group (P=.025), severe cellulitis occurred only in S. maltophilia-infected patients (11 [30%]; P=.0002), and the bacteremia presented as breakthrough infection in 56% of the cases due to S. maltophilia (vs. only 24% of those due to P. aeruginosa; P=.002). Acute mortality rates associated with S. maltophilia and P. aeruginosa bacteremia were 24% and 21%, respectively. In both groups, profound neutropenia and hypotension at the onset of bacteremia, duration of profound neutropenia during bacteremia, severity-of-illness score ≥4, and inappropriate antibacterial treatment were factors significantly associated with death. Most S. maltophilia isolates were resistant to aminoglycosides, β-lactams, and ciprofloxacin. Cotrimoxazole and ticarcillin-clavulanic acid showed borderline activity. Prompt administration of in vitro-active antibiotics may improve the prognosis of S. maltophilia bacteremia, especially for immunocompromised patients, and novel drug combinations are needed for the treatment of severe infections.
Stenotrophomonas maltophilia is an aerobic ubiquitous gram-negative bacillus that has been isolated from humans, animals, soil, food, and pharmaceuticals [1]. Because of its limited invasiveness and low level of pathogenicity, it is rarely responsible for community-acquired serious infections [2]; rather, it is usually a commensal, a contaminant, or part of the endogenous flora of hospitalized patients [1].
In recent years, S. maltophilia has been increasingly reported as a cause of life-threatening infections, in particular in immunocompromised patients; antimicrobial pressure, indwelling venous catheterization, long-term hospitalization, aggressive cytoreductive treatments for malignancy, and neutropenia seem to be contributory factors [3–10]. The morbidity and mortality associated with severe S. maltophilia infections may be high for several reasons: (1) this species is inherently resistant, (2) its antimicrobial resistance may increase when patients who are colonized with S. maltophilia are treated with antibiotics over extended periods, and (3) the debilitated state of patients at risk may seriously complicate antimicrobial treatment [6, 11, 12].
For all these reasons, it seemed appropriate for us to gather and report additional data on the presumed emerging role and clinical significance of S. maltophilia in patients with hematologic malignancies. To this end, all cases of S. maltophilia bacteremia that occurred during a 10-year period at our facility were analyzed for predisposing factors, clinical presentation, outcome, and impact of antimicrobial therapy in a case-control study that also included patients with P. aeruginosa bacteremia.
Patients and Methods
Patients with hematologic malignancies who were treated at the department of hematology of the University La Sapienza (Rome) from 1987 to 1996 and whose blood cultures were positive for S. maltophilia were included in this retrospective study. The episodes of S. maltophilia bacteremia were compared with those due to Pseudomonas aeruginosa that occurred among our hematology patients during the same period (the control group). We selected the cases of P. aeruginosa bacteremia by occurrence simultaneously with or just before or after each case of S. maltophilia bacteremia.
At our institute, patients undergoing induction remission, consolidation chemotherapy, or autologous bone-marrow transplantation are placed in rooms with 2 beds, and almost all receive prophylaxis (oral quinolones for adults [13] and oral cotrimoxazole for children) when their granulocyte counts are <1000/mm3. Allogeneic bone-marrow recipients are placed in rooms with single beds, and reverse isolation procedures are performed for prevention of infection.
Surveillance cultures of oropharyngeal and rectal swabs, urine specimens, and stools are performed weekly. In some cases, blood cultures are performed during the 24–48 h after insertion of a central venous catheter or as part of a follow-up evaluation of apparently cured bloodstream infection, after discontinuation of antibiotic treatment. If fever occurs, all patients are thoroughly investigated for a potential infectious etiology, and broad-spectrum antibiotics are given empirically within 4–8 h in cases of granulocytopenia (<1000 granulocytes/mm3).
The majority of patients included in this study received antibacterial treatment according to the following current multicenter or local protocols. From November 1988 to August 1989, patients were randomly assigned to receive piperacillin plus amikacin or piperacillin plus amikacin plus teicoplanin [14]; from December 1989 to September 1990, to receive piperacillin-tazobactam plus amikacin or piperacillin plus amikacin plus teicoplanin [15]; from November 1990 to January 1992, to receive ceftazidime plus amikacin plus teicoplanin or vancomycin [16]; and from November 1993 to June 1994, to receive ceftazidime plus amikacin or meropenem [17]. From February 1995 to October 1996, all patients received ceftazidime plus amikacin plus teicoplanin [18]. Patients not included in the above-mentioned protocols usually received piperacillin or ceftriaxone plus amikacin. Records of the hospital pharmacy were reviewed to evaluate the cumulative days of therapy with the different antibiotics.
The microbiological investigations performed for all febrile patients, whether granulocytopenic or not, included ≥3 blood cultures and cultures of specimens from any suspected site of infection (such as urine, respiratory secretions, rectum, insertion site of a central venous catheter, and/or catheter tip). Patients with a central venous catheter had ≥1 blood sample drawn via a peripheral vein. Daily blood cultures were usually performed for the duration of fever and for 2–3 days after fever resolved.
Bacteremia was defined by the isolation of S. maltophilia from ≥1 culture of blood drawn before death. A case was considered clinically significant when signs or symptoms of systemic infection were observed concurrently with bacteremia or when S. maltophilia was isolated from blood and from a clinically evident local infection. Polymicrobial bacteremia was defined by the isolation of >1 organism in a single blood culture. Breakthrough bacteremia was defined as an episode of bacteremia occurring at least 48 h after the start of systemic antibiotic therapy. The source of bacteremia was clinically suspected when an active site of infection was documented concurrently with bacteremia. The bacteremia was considered catheter-related when it was associated with inflammation of the exit site or subcutaneous tunnel of a central venous catheter or with a positive culture of a catheter tip. Pneumonia due to S. maltophilia was suspected when S. maltophilia was isolated from blood or respiratory specimens; cough and chest pain occurred; and evidence of pneumonia was documented radiographically, after previously negative chest radiographic findings.
Clinical data recorded included demographics; underlying hematologic malignancies; administration of steroids; cytotoxic chemotherapy; granulocytopenia (total granulocyte count <1000/mm3); duration of granulocytopenia (calculated from the time when the granulocyte count decreased to 1000/mm3 until it returned to a level >1000/mm3); profound granulocytopenia (total granulocyte count <100/mm3); persistent profound granulocytopenia (in those episodes of septicemia in which a granulocyte count <100/mm3 did not return to a level >500/mm3 after 2 weeks); presence of an indwelling central venous catheter; oral prophylactic antibiotics; and prior or concomitant antibiotic therapy. Surveillance cultures to evaluate S. maltophilia colonization in patients with S. maltophilia bacteremia were also reviewed.
Clinical presentation, presence of mucocutaneous or soft-tissue infections, antibacterial treatment, and outcome were evaluated. The treatment of each infection was categorized as appropriate or inappropriate, depending on the in vitro susceptibility of the isolates. The severity of the illness at the onset of bacteremia was assessed by a grading system [19].
Twenty-seven strains of S. maltophilia remained available for further study. This included determination by the macrodilution method of the MIC of the following antibiotics: trimethoprim-sulfamethoxazole, piperacillin, piperacillin-tazobactam, ceftazidime, cefepime, amikacin, ciprofloxacin, meropenem, and ticarcillin-clavulanate. Agar disk diffusion tests were also performed for the determination of susceptibility to minocyclin. All susceptibility tests were performed according to recommendations of the National Committee for Clinical Laboratory Standards [20]. DNA ribotyping to distinguish the different strains extracted was performed according to the procedures of Ezaki et al. [21], and digestion was with the restriction enzyme BamHI [22].
Univariate analysis of the contingency data was done with use of the 2-tailed χ2 or Fisher's exact test.
Results
From 1987 to 1996, admissions to the hematology department of the University La Sapienza progressively increased, from 556 admissions in 1987 to 840 in 1996. After a new ward and a hematology emergency unit were created, the total number of beds increased from 38 to 58. During this period, we identified 44 S. maltophilia isolates from among 3284 blood culture isolates recovered from patients with hematologic malignancies (1.3%); as shown in table 1, the annual frequency of S. maltophilia blood isolation remained constant during this period, ranging from 0.2% in 1996 to 2% in 1993. During the same period, P. aeruginosa isolations from blood ranged from 3.5% in 1994 to 9.2% in 1993.
Analysis of the hospital pharmacy records showed that from 1989 to 1996, there were 2910 cumulative days of therapy with carbapenems (268 days of therapy with meropenem and 2641 days with imipenem), compared with 61,384 cumulative days of therapy with β-lactam antibiotics (27,931 cumulative days with ceftriaxone, 9344 with ceftazidime, and 24,109 with piperacillin).
Among 7 cases, S. maltophilia was isolated from a single blood culture, and in these cases a review of the clinical record showed neither fever nor any other sign or symptom of infection. Moreover, clearance of bacteremia occurred without any antibacterial treatment and/or prophylaxis. Blood cultures in these cases were performed after insertion of a central venous catheter or as part of a follow-up evaluation of apparently cured bloodstream infection after discontinuation of antibiotic therapy. In these cases, S. maltophilia was considered a contaminant, and the study therefore analyzed only the 37 remaining cases.
Table 2 compares demographic characteristics for patients with S. maltophilia versus P. aeruginosa bacteremia. Sex distribution, age, underlying hematologic malignancies, intensive antineoplastic chemotherapeutic treatment, and use of steroids were comparable in the 2 groups of patients. The number of patients with neutropenia, the degree and duration of neutropenia, the mean duration of hospitalization before pathogen isolation, as well as the presence of an indwelling central venous catheter were the same in both groups. However, the duration of profound neutropenia (number of days with <100 neutrophils/mm3) was greater for patients with S. maltophilia septicemia (P=.025).
Prophylaxis with quinolones was given to 24 patients (65%) with S. maltophilia bacteremia and 15 (40%) with P. aeruginosa bacteremia. Similar proportions of patients received cotrimoxazole prophylaxis (8% and 11%). The presence of an indwelling central venous catheter was somewhat more frequent in patients with S. maltophilia bacteremia (65%) than in those with P. aeruginosa bacteremia (40%), although the difference was not statistically significant.
Table 3 shows the clinical findings at the onset of bacteremia. Fever and hypotension appeared more frequently among patients with S. maltophilia bacteremia, but the difference was not significant. The illness score was ≥4 for 9 patients (24%) infected with S. maltophilia and for 8 (22%) infected with P. aeruginosa. No statistically significant differences were observed in the number of neutropenic patients, the degree or duration of neutropenia, or the mean duration of hospitalization before septicemia. Central venous catheter-related bacteremia occurred in 16% of patients with S. maltophilia bacteremia versus 8% of those with P. aeruginosa. Mucocutaneous or soft-tissue infections were observed in 11 patients in the S. maltophilia group but none in the P. aeruginosa group (P=.0002). In 6 cases the infection was cellulitis located at the catheter exit site (in 3, it extended to the subcutaneous tunnel); in 1 it was extended perianal cellulitis; and in 1 it was infiltrated ulcers of the gingiva, buccal mucosa, and pharynx.
Three patients had embolic lesions consisting of multiple, disseminated, hard, subcutaneous nodules ranging in size from 1×1 to 5×6 cm. Pulmonary infiltrates were documented in 7 patients (19%) with S. maltophilia bacteremia (the organism was also isolated from respiratory specimens from 4 of these 7) and in 3 (8%) with P. aeruginosa bacteremia. Although not significant, a trend toward a slower clearance of bacteremia was apparent among patients infected with S. maltophilia, whose blood cultures remained positive during antibiotic therapy for a mean duration of 5 days (vs. 3.5 days in the P. aeruginosa group).
Twenty-two cases of S. maltophilia bacteremia (59%) presented as breakthrough infections, and in 20 cases the ongoing antibiotic treatment was with a broad-spectrum antibacterial combination; only 9 (24%) of the cases of P. aeruginosa bacteremia were breakthrough infections (P=.002). In both groups, no breakthrough bacteremia developed during carbapenem treatment. Two patients in the S. maltophilia group had received imipenem within the 2 weeks before the onset of bacteremia.
In 6 cases (16%) S. maltophilia bacteremia was polymicrobial (in 3 cases there was mixed infection with gram-positive cocci; in 2, with gram-negative bacilli; and in 1, with a gram-negative anaerobic bacillus). P. aeruginosa bacteremia was polymicrobial in only 1 case (2.7%).
Overall, including the polymicrobial cases (none of these associated with a fatal outcome), 9 patients infected with S. maltophilia died within 14 days of the onset of septicemia (acute mortality rate, 24%), and 3 patients died later during the same hospitalization, after resolution of bacteremia (crude mortality rate, 32%). No difference was observed between these rates and those associated with P. aeruginosa septicemia (acute mortality, 10 [21%] of 37 patients; crude mortality, 12 [27%] of 37).
Univariate analysis (table 4) was performed to assess risk factors of acute mortality in patients with S. maltophilia bacteremia or P. aeruginosa bacteremia. In both groups, profound neutropenia and hypotension at the onset of bacteremia, duration of profound neutropenia during bacteremia, severity of illness score ≥4, and administration of inappropriate antibacterial treatment were factors significantly associated with acute mortality. Moreover, the administration of remission induction or reinduction intensive chemotherapy was associated with acute mortality among patients with S. maltophilia bacteremia, as was pneumonia in patients with P. aeruginosa bacteremia.
All patients with S. maltophilia bacteremia received antibacterial treatment, which in 73% of cases was a β-lactam-plus-aminoglycoside combination. Evaluation of the treatment of each infection showed that 15 patients received antibacterials to which the strain was susceptible, and 14 (93%) of them survived, whereas 22 patients received antibacterials without in vitro activity, and only 11 (50%) of them survived (P=.006). Administration of inappropriate treatment was eventually recognized as occurring more frequently for patients with S. maltophilia bacteremia (22 patients [59%]) than for patients with P. aeruginosa bacteremia (14 patients [38%]; P=.05).
Removal of the infected central venous catheter was required for 5 patients with S. maltophilia bacteremia and 1 patient with P. aeruginosa bacteremia. As shown in table 5, most S. maltophilia isolates were resistant to aminoglycosides, ciprofloxacin, and β-lactam antibiotics. Ticarcillin-clavulanate and piperacillin-tazobactam had the best in vitro activity; in fact, 92% and 63%, respectively, of the 26 isolates tested were susceptible to these combinations. Cotrimoxazole was active against 42% of the isolates and minocycline against 100%.
Twenty-seven S. maltophilia strains were analyzed by DNA ribotyping; BamHI restriction enzyme was sufficient to discriminate among the strains. Restriction-fragment length polymorphism (RFLP) analysis showed that only 2 isolates had identical patterns; all others were different.
Discussion
Recent reports indicate that the incidence of serious infections due to S. maltophilia is increasing as a consequence of larger populations of patients with predisposing factors, such as aggressive therapies for malignancies, immunosuppression, neutropenia, antimicrobial pressure, and long-term presence of venous access devices. In this retrospective study, we did not observe an increased incidence of S. maltophilia bacteremia among our patients with hematologic malignancies receiving antineoplastic-intensive chemotherapy. Our policy to limit the administration of carbapenems, used only when a severe infection due to a gram-negative multiresistant strain is documented, might explain this epidemiological feature at our institution [10].
To identify specific risk factors for S. maltophilia bacteremia, we included in our study a control group of patients with bacteremia due to P. aeruginosa, which at our institution remains the nonfermenting gram-negative bacillus most frequently isolated from blood and from severe infections.
Neutropenia was confirmed as a risk factor for S. maltophilia bacteremia. Even though at the onset of bacteremia the majority of patients in both groups were neutropenic and the total durations of neutropenia were similar, patients infected with S. maltophilia had neutrophil counts <100/mm3 for significantly more days (P=.025). Hospitalization was confirmed to be a predisposing factor for development of S. maltophilia infections; although the difference between those infections and the cases of P. aeruginosa bacteremia is not statistically significant, we observed that more cases of S. maltophilia bacteremia developed during hospitalization. Moreover, S. maltophilia-infected patients had been hospitalized for longer periods before isolation of the pathogen. The role of antibiotic pressure as a predisposing factor is suggested by the high rates of antibacterial resistance found among our strains and by the fact that the majority of cases of S. maltophilia bacteremia (59%, vs. 24% of cases of P. aeruginosa bacteremia; P=.002) involved patients who were already receiving antibacterial treatment.
The origin of S. maltophilia remains unknown. One of our patients might have been carrying a strain at the time of admission to the hospital, and the combined action of antibiotic pressure plus profound and prolonged immunodepression may have favored the development of infection. On the other hand, for some patients the environment might have represented the source of nosocomial infection. The great diversity of our isolates, as determined by ribotyping studies, seems to support the former hypothesis.
The respiratory tract [2] and gastrointestinal tract [7] have been suggested as reservoirs. We documented pharyngeal colonization in only 3 cases, but we did not look for S. maltophilia in the stool specimens of these patients. A central venous catheter was documented as the primary source of infection in only 6 patients (16%). Hypotension and fever at the onset of bacteremia occurred similarly in the 2 groups, but morbidity associated with S. maltophilia appeared to be somewhat higher than that associated with P. aeruginosa; in particular, severe cellulitis occurred in 30% of patients with S. maltophilia bacteremia, compared with none in the other bacteremia group (P=.0002). Even though we did not find a significant difference with the P. aeruginosa bacteremia group, among S. maltophilia–infected patients there was a trend toward more central venous catheter placements, more catheter-related bacteremia, and more frequent documentations of cellulitis at the catheter exit site.
Antibacterial therapy for S. maltophilia infections is problematic; in fact, most S. maltophilia strains are resistant to the majority of agents used for the empirical treatment of febrile neutropenia or for documented infections due to gram-negative organisms (aminoglycosides, extended-spectrum penicillins, and third-generation cephalosporins) [8–11]. Of particular concern, we observed among our S. maltophilia isolates a 58% rate of resistance to cotrimoxazole, which is one of the agents recommended for treatment of severe infections [5, 6]. Ticarcillin-clavulanic acid was the most active drug, followed by piperacillin-tazobactam: both of these agents showed significant inhibition against about two-thirds of our isolates.
In our study, fewer patients with S. maltophilia bacteremia were appropriately treated (41%) than were patients with P. aeruginosa bacteremia (62%; P=.05). Moreover, we also observed, as reported in other studies, a significant association between survival and administration of an antimicrobial agent to which the strain was susceptible in vitro [5, 6]. These data may explain, at least in part, the high acute and crude mortality rates among S. maltophilia-infected patients, which in this study were similar to those among patients with P. aeruginosa bacteremia.
In conclusion, S. maltophilia bacteremia in patients with hematologic malignancies seems to be a serious complication of profound and persistent neutropenia and broad-spectrum antibiotic pressure. Morbidity and mortality are high; however, the prognosis may be improved by prompt administration of in vitro-active antibiotics. Recent reports suggest that newer quinolones (particularly in combination with other agents) [23] or the combination of polymyxin B, rifampin, and β-lactams [24] or that of piperacillin-tazobactam [25] could have a therapeutic role. High resistance rates to cotrimoxazole and the borderline activity of ticarcillin-clavulanic acid, which have been the agents of choice, suggest the urgent need for a novel drug combination for the treatment of serious S. maltophilia infections.
References
Presented in part at the 37th Interscience Conference an Antimicrobial Agents and Chemotherapy, Toronto, 28 September–1 October 1997 (abstract J-208).
Comments