Stenotrophomonas maltophilia (formerly: Pseudomonas or Xanthomonas maltophilia) is an aerobic non-fermenting Gram-negative bacillus (NGNB) that can be found ubiquitously in the environment (1). Next to Pseudomonas aeruginosa and Acinetobacter spp., the organism is considered the third most frequent nosocomial pathogen among non-fermentative bacteria (2, 3).

Pneumonia and bloodstream infection (BSI) are the most common clinical manifestations of S.maltophilia infections. Less frequently, S.maltophilia can cause urinary tract infections, cholangitis, peritonitis, wound infections, eye infections, arthritis, meningitis, and endocarditis (4, 5). Patients with hematologic malignancies are at high risk for S.maltophilia infection because of chemotherapy-induced neutropenia and immunodeficiency. Frequent exposure to broad-spectrum antibiotics and the presence of central venous catheters further enhance the risk of S.maltophilia infection (6, 7). The rate of S.maltophilia BSI among BSIs in this patient population has been reported to be as high as 60% (8–11).

Treatment of S.maltophilia infection can be difficult because of the organisms inherent resistance to a variety of antibiotics (12, 13). Trimethoprim-sulfamethoxazol (TMP-SMX) is the drug of choice, and fluoroquinolones are the proposed alternative. Similar to the treatment of Pneumocystis jirovecii pneumonia, up to five-fold higher than regular doses of TMP-SMX are recommended for severe infections (5, 14). Thus, the therapeutic options for S.maltophilia infections are quite different from those available for other NGNB, and appropriate antimicrobial therapy is often delayed through ineffective treatment during initial empirical therapy (15). Accordingly, mortality rates are high in immunocompromised and critically ill patients (11, 16), with 30-day mortality rates of S.maltophilia BSIs ranging from 11% to 53% (8, 11–13, 17–19).

While series of adult cancer patients with invasive S.maltophilia infections have been published in regular intervals, few reports exist for pediatric patients with cancer and/or allogeneic hematopoietic cell transplantation (HCT) (20–22). We therefore analyzed the incidence, genetic relatedness, clinical course and outcomes of invasive S.maltophilia infections observed during the past ten years at our institution, a high volume European pediatric cancer center with an active allogeneic HCT program.

Stenotrophomonas maltophilia is a non-fermentative, Gram-negative bacillus that has emerged as important nosocomial pathogen in immunocompromised and critically ill patients (16, 27). Published experience in pediatric patients with cancer and/or allogeneic HCT is limited to two separate studies reporting on a total of 24 S.maltophilia BSIs (22, 28) and several larger pediatric series that include a relevant proportion of patients with hematological malignancies or solid tumors (24, 25, 29, 30) ( Table 2 ). In the study presented here, S.maltophilia accounted for 1.8% of all BSIs and for 7% of those caused by Gram-negative rods. Invasive S.maltophilia infection was associated with a diagnosis of acute leukemia and/or allogeneic HCT, or immunodeficiency, and occurred in a setting of impaired host defences, defective mucocutaneous barriers, indwelling central venous catheters, treatment with broad-spectrum antibacterial agents, and admission to the intensive care unit. Four patients died in direct relationship to the infection, including three patients with pneumonia and pulmonary hemorrhage ( Figure 3 ), which has been reported to be associated with S.maltophilia infection and status post allogeneic HCT (29, 44, 45). Similar to others (28), we found a numerical increase in S. maltophilia infections over time. Molecular typing of a limited number of blood culture isolates, however, confirmed that isolates were genetically not related and suggests the absence of a nosocomial outbreak (46).

The exact route of acquisition of S.maltophilia often remains unknown. Nevertheless, isolation of the organism from mucosal surfaces of the respiratory and/or the lower gastrointestinal tract may herald later infection as many patients with S.maltophilia BSIs were reported to be colonized prior to infection (5, 17). Indeed, the oral microbiome has recently been described as a potential reservoir, and real-time monitoring of the oral S.maltophilia relative abundance has been suggested to identify patients at risk for invasive infection (30). In our limited cohort, concomitant colonization was detected in the majority of cases with invasive S. maltophilia infection, but overall, there was no apparent relationship between pharyngeal or respiratory colonization and invasive infection.

Similar to previous reports (20), the majority of invasive S. maltophilia infections in our cohort was associated with indwelling central venous catheters. Six of the ten patients were treated at the intensive care unit and four were on invasive ventilation. Intensive care, mechanical ventilation, and/or central venous catheterization have been identified as risk factors for S. maltophilia BSI and/or dismal outcome ( Table 2 ). Several studies suggest a survival benefit for removal of indwelling central venous catheters (8, 9, 17, 47–49), and international guidelines strongly recommend prompt catheter removal in S. maltophilia associated BSIs (50), independent on whether the catheter is considered the source of the infection or being colonized secondary to ongoing bacteremia.

Patients with S.maltophilia BSIs often have polymicrobial infections (5), and their relative frequency in children seems to be higher as observed with Pseudomonas aeruginosa (35). In the cohort presented here, concomitant BSI occurred in 30% of patients with S.maltophilia infection, which is below the rate in previous series of pediatric patients (31, 33). Bacteria most commonly recovered in temporal context with S.maltophilia were coagulase-negative Staphylococcus and Enterococcus spp (8, 45). It remains unclear whether the detection of S.maltophilia is a consequence of appropriate antimicrobial therapy for other BSIs or whether the concurrent invasive infections simply reflect the sum of immunodeficiency in the affected patients.

In eight of the ten cases, S.maltophilia infection occurred as breakthrough infection in patients receiving broad-spectrum antibacterial agents. Prior use of carbapenems has been repeatedly described as a risk factor for S.maltophilia infection (12, 17, 44, 51–53), and cumulative carbapenem use has been identified to be associated with S.maltophilia in leukemia patients with altered oral microbiome (30). Similarly, in the majority of studies in pediatric patients investigating factors related with outcome, prior use of carbapenems was associated with dismal outcome of S.maltophilia BSI (22, 35, 39) ( Table 2 ). As a consequence, clinicians should be aware that breakthrough infection with S.maltophilia may occur in severely ill patients being treated with carbapenems.

Antibacterial therapy for S.maltophilia infections is challenging because most clinical isolates are resistant to agents commonly used for empirical treatment of febrile neutropenia or documented infections by Gram-negative organisms, including extended-spectrum penicillins, third-generation cephalosporins, carbapenems, and aminoglycosides (53). In addition, current recommendations for treatment are based on historical evidence, case series, and in vitro susceptibility data rather than pharmacokinetic/pharmacodynamic considerations and results of controlled clinical trials (5). In general, TMP-SMX is the drug of choice for infections by susceptible S.maltophilia isolates based on its potent in vitro activity and documented clinical efficacy (10, 46). Nevertheless, susceptibility varies between geographic regions and resistance is an emerging threat (2, 5, 10, 16, 54, 55). Alternatives to treatment with TMP-SMX include fluoroquinolones and tigecycline (12, 13). However, in contrast to TMP-SMX, clinical breakpoints for these agents have not been defined, which makes a valid interpretation of in vitro susceptibility testing results with regards to the prediction of clinical efficacy difficult. In our study, seven of ten S.maltophilia initial isolates from patients with invasive infections were susceptible in vitro to TMP-SMX, and in one of these patients, isolates became resistant during treatment. Apart from primary or secondary resistance, further concerns in immunocompromised patients with cancer and/or allogeneic HCT include the myelotoxicity of therapeutic doses of TMP-SMX (18) and the widespread use of low and intermittent doses of the agent for antibacterial or anti-Pneumocystis prophylaxis that may result in the selection of resistant S.maltophilia strains (17). Indeed, based on emerging resistance, it has been suggested by individual experts to consider escalating therapies in immunocompromised or critically ill patients (49, 56). Previous observations on the use of fluoroquinolones against invasive S.maltophilia infections have demonstrated comparable patient survival relative to TMP-SMX (12, 13, 57), and quinolone prophylaxis in adult cancer patients has been associated with a reduced incidence of invasive S.maltophilia infections (58). Nevertheless, quinolone monotherapy for S.maltophilia BSIs should be critically reflected (19), as rapid emergence of resistance to these agents has been observed both in vitro and in vivo (5, 59).

Considering the small number of patients, the 30-day mortality rate of 30% in patients with invasive S.maltophilia infections in our study is in line with 30-day mortality rates of S.maltophilia BSIs of 33% and 38% reported by others (17, 60). In pediatric studies, the reported all-cause mortality rates in patients with S.maltophilia infections range from 12.5% to 61% with an attributable mortality of 0% to 18%, respectively (22, 31, 33, 37–39, 41) ( Table 2 ). Many studies across all age groups have reported risk factors for mortality associated with S.maltophilia BSIs including prolonged hospitalization prior to BSI onset, previous exposure to antimicrobial agents, use of indwelling medical devices, a compromised health status, complex medical care, granulocytopenia and/or transplantation, and inappropriate therapy (5, 8–10, 15, 17, 18, 36, 48, 61–64). While the limited number of patients included precluded robust statistical assessments, the presence of pneumonia and admission to the intensive care unit were both significantly associated in explorative analyses with mortality in the present study ( Supplementary Table 1 ). Nevertheless, as it is often difficult to distinguish between colonization and infection, identification of risk factors for mortality is ultimately limited to BSIs and may not consider the full spectrum of diseases caused by the organism (7).

To conclude, as reflected in this limited series of heterogenous patients, defined therapeutic strategies for invasive S.maltophilia infections in immunocompromised pediatric patients, including those with cancer and/or allogeneic HCT, so far lack uniformity but remain an important goal. Clinicians should be aware that breakthrough infections by S.maltophilia may occur during the administration of broad-spectrum antibiotics, especially following carbapenem use, and that these infections may be associated with fulminant and fatal pulmonary hemorrhage, in particular in allogeneic HCT patients (29). Detection of BSI by S.maltophilia should prompt the removal of indwelling central venous catheters and the immediate initiation of therapeutic doses of TMP-SMX. Initial combination with second generation fluoroquinolones and tigecycline until return of resistance testing and achievement of a stable clinical response may be considered in view of the high case fatality rates.

The data analyzed in this study is subject to the following licenses/restrictions: Clinical data. Requests to access these datasets should be directed to (andreas.groll@ukmuenster.de).

Written informed consent was obtained from the individual(s), and minor(s)’ legal guardian/next of kin, for the publication of any potentially identifiable images or data included in this article.

Retrospective analysis of single-center data and literature review was conducted by SKZ, supported by HH and with clinical input by CR, AH, KM, AR, and AHG. Identification and susceptibility testing by NJF, whole genome sequencing based typing was performed and analyzed by SK. Statistical analysis was performed by AR, SKZ and AHG. Manuscript was written by SKZ and was edited by SK and AHG. All authors contributed to the article and approved the submitted version.

The study was funded by internal resources.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.