Fatal hemorrhagic pneumonia in patients with hematologic diseases and Stenotrophomonas maltophilia bacteremia: a retrospective study

Background Fatal hemorrhagic pneumonia is one of the most severe manifestations of Stenotrophomonas maltophilia (SM) infections. Here, we aimed to investigate the clinical characteristics of SM bacteremia and to identify the risk factors of hemorrhagic pneumonia caused by SM in patients with hematologic diseases. Methods The clinical records of 55 patients diagnosed with hematologic diseases and SM bacteremia were retrospectively reviewed. We compared patients’ clinical characteristics and outcomes between the hemorrhagic pneumonia group and non-hemorrhagic pneumonia group. Results Twenty-seven (49.1%) patients developed hemorrhagic pneumonia. The overall mortality rate of SM bacteremia was 67.3%. Hemorrhagic pneumonia (adjusted HR 2.316, 95% CI 1.140–4.705; P = 0.020) was an independent risk factor of 30-day mortality in hematological patients with SM bacteremia. Compared with the non-hemorrhagic pneumonia group, patients in the hemorrhagic pneumonia group were older and showed clinical manifestations as higher proportions of isolated SM in sputum culture, neutropenia and elevated procalcitonin (PCT). Multivariate analysis showed that neutropenia, high levels of PCT, prior tigecycline therapy within 1 month were independent risk factors associated with hemorrhagic pneumonia. Conclusions Neutropenia, high level of PCT and prior tigecycline therapy within 1 month were significant independent predictors of hemorrhagic pneumonia in hematologic patients with SM bacteremia. Due to no effective antibiotics to prevent hemorrhagic pneumonia, prophylaxis of SM infection and its progression to hemorrhagic pneumonia is particularly important.


Introduction
Stenotrophomonas maltophilia (SM) is a nonfermentative gram-negative bacterium without highly virulent widely found in nature, and rarely causes infection in the normal immune population [1]. However, SM has been becoming a common cause of opportunistic infections in immunocompromised patients, especially in patients diagnosed with hematologic diseases, and can cause serious infections, such as bacteremia or hemorrhagic pneumonia [2][3][4][5][6].
Due to the intrinsic resistance of SM to many broadspectrum antimicrobial agents, it is difficult to determine the appropriate antibiotic treatment before the SM is detected [2,7]. Therefore, previous studies have shown a very poor prognosis in the cases of SM bacteremia, and the overall mortality was ranged from 35 to 75% [8][9][10][11][12]. Noteworthy, in patients with hematologic malignancies and/or hematopoietic stem cell transplantation (HSCT) recipients, SM can cause fatal hemorrhagic pneumonia, and the mortality is nearly 100% [10,[12][13][14][15][16][17]. At present, only a few reports have performed a statistical analysis to identify the risk factors of hemorrhagic pneumonia caused by SM [10,12].
In this study, we identified cases of SM bacteremia in patients with hematologic disorders and performed a retrospective review of these patients' clinical characteristics and outcomes. Further, we attempted to confirm the risk factors of SM bacteremia with hemorrhagic pneumonia.

Patients
This study was conducted at a 318-bed hematology department, the first affiliated hospital, College of Medicine, Zhejiang University, in China. From January 2015 to July 2019, all patients who were diagnosed with hematologic diseases and had positive blood cultures for SM were included in the study, with each patient being included only once at the time of initial blood culture. All patients were categorized into 2 groups: hemorrhagic pneumonia group and non-hemorrhagic pneumonia group. Clinical and microbiological characteristics were compared between the two groups to identify the risk factors for hemorrhagic pneumonia in SM bacteremia patients.

Data collection
We retrospectively reviewed the electronic medical records for all enrolled patients and extracted the clinical data. The data included the following: age, gender, underlying hematologic disease, therapeutic treatment for hematologic diseases, sputum culture, polymicrobial bacteremia, laboratory tests, antibiotics susceptibility testing results, prior antibiotic therapy within 1 month, indwelling central venous catheter, inadequate initial empirical antimicrobial treatment and mortality.
Definitions SM bacteremia was defined as one or more positive blood cultures with clinical symptoms of infection. SM bacteremia with hemorrhagic pneumonia was diagnosed as follows:1) SM isolated from a blood culture, with or without SM in sputum culture at the same time, 2) newly detected imaging findings by chest X-ray or computed tomography (CT) scan when the blood culture was positive, 3) with pulmonary hemorrhage symptoms, such as continuous blood in sputum or hemoptysis. Polymicrobial bacteremia was defined as the isolation of another pathogen within 24 h of index SM isolate, satisfying the Centers for Disease Control and Prevention (CDC) criteria for blood stream infection (BSI) [18]. Neutropenia was defined as an absolute neutrophil count of < 500/uL at the onset of bacteremia. Adequate initial antimicrobial treatment was defined as the administration of at least one intravenous antibiotic to which the microorganism was susceptible in vitro within 72 h of the onset of SM bacteremia. Onset of SM bacteremia was defined as the date when the blood culture was collected.

Blood culture and drug sensitivity test
The BacT/Alert 3D (bioMérieux, France) automated blood culture system was used. Species identification was performed by the automated Vitek 2 system with the matched gram-negative (GN) test card (bioMérieux, France). Antimicrobial susceptibility was tested with the agar dilution method. The antibiotics tested included trimethoprim-sulfamethoxazole (TMP-SMX), cefoperazone/sulbactam, minocycline, levofloxacin and tigecycline. The MICs of antimicrobial agents were interpreted according to the 2018 Clinical and Laboratory Standards Institute (CLSI) criteria except for tigecycline [19]. The interpretation of tigecycline was performed based on the breakpoints published by the Food and Drug Administration (FDA) (available at https://www.fda.gov/drugs/ development-resources/tigecycline-injection-products), and the relevant criteria for Enterobacteriaceae was used. For Enterobacteriaceae isolates, ≤2 / ≥8 μg/ml were the sensitivity/resistance MIC breakpoints against tigecycline based on FDA criteria. Additionally, Cefoperazone/sulbactam susceptibility was based on the breakpoints for cefoperazone alone against Enterobacteriaceae (susceptible 16/8 mg/L; intermediate, 32/16 mg/ L; and resistant, 64/32 mg/L) according to the criteria for susceptibility testing by CLSI [20].

Statistical analysis
All statistical analyses were performed using SPSS version 22.0 for Windows. The independent sample t-test was used for continuous variables and the Chi-square test or Fisher's exact test was used for categorical variables. Cox's proportional hazard model with forward selection was used to find out which independent risk factors were predicting 30-day mortality. All variables with P-value of less than 0.10 in univariate analyses were enrolled in further multivariate analysis. The survival curves were performed by the Kaplan-Meier method. Univariate and multivariable logistic regression was used to evaluate the risk factors of hemorrhagic pneumonia in hematologic patients with SM bacteremia. Variables with P-value of less than 0.10 in univariate analyses were included in a multivariable logistic regression analysis. In addition, thrombocytopenia and prolonged activated partial thromboplastin time (APTT) which are considered as the potential variables associated with bleeding were also enrolled in the multivariate model. Results from the univariate and multivariate analysis are expressed as a hazard ratio (HR) or an odds ratio (OR) and 95% confidence interval (CI). Statistical tests were two-sided, and significance was defined as P-value < 0.05.
However, SM was only identified in 14 (25.5%) patients from sputum culture simultaneously.
The cohort was categorized into hemorrhagic pneumonia group and non-hemorrhagic pneumonia group (27 cases and 28 cases, respectively). The comparisons of clinical characteristics between the two groups were listed in Table 1. The median age was 56 years old in the hemorrhagic pneumonia group, which was older than the 48.5 years observed in the non-hemorrhagic pneumonia group (P = 0.049). There were more patients (40.7%) isolated SM from sputum in the hemorrhagic pneumonia group. The proportions of neutropenia and elevated procalcitonin (PCT) were significantly higher in the hemorrhagic pneumonia group than the nonhemorrhagic pneumonia group (P = 0.016 and P = 0.003, respectively). There were no significant differences for patients' gender, polymicrobial bacteremia, anemia, platelet count, coagulation function, underlying hematologic diseases, C-reactive protein level, inadequate initial antimicrobial treatment and prior courses of chemotherapy between the two groups.

Risk factors associated with mortality
The overall mortality was 67.3% (37/55), and all patients died within 30 days after SM bacteremia onset in nonsurvivor group. It was noteworthy that 20 (36.4%) patients died before obtained the positive blood culture reports, and all these patients had hemorrhagic pneumonia. The 30-day mortality rate was significantly higher in patients with hemorrhagic pneumonia than those without hemorrhagic pneumonia (85.2 and 50.0%, respectively). In this study, 35 (63.6%) patients received inadequate initial antimicrobial treatment. The 30-day mortality rate was 77.1% in patients who received inadequate initial antimicrobial treatment, which was significantly higher than that in patients treated with adequate initial antimicrobial (50.0%).

Previous antimicrobial therapy within 1 month before BSI onset
All patients in this study were administrated with antibiotics before SM bacteremia onset, either due to a previous other infection or as an empirical therapy for unknown origin infection ( Table 3). The majority of patients received carbapenems (46 cases, 83.6%) and/or antifungal drugs (45 cases, 81.8%). But there were only a few patients who had treated with TMP-SMX (3 cases, 5.5%). It was to be observed that a total of 23 (41.8%) cases received tigecycline, and tigecycline was more frequently prescribed in the hemorrhagic pneumonia group (59.3 and 25%, respectively, P = 0.010).

In vitro antimicrobial susceptibility
The antimicrobial susceptibilities of SM isolate in vitro are shown in Fig. 2. All patients were tested for TMP-SMX sensitivity, with a sensitivity rate of 89.09%. Susceptibility to levofloxacin, minocycline, cefoperazone/ sulbactam was 83.3% (40/48), 100% (46/46) and 77.8% (35/45), respectively. There was no significant difference in susceptibility to four antibiotics mentioned above between the hemorrhagic pneumonia group and the nonhemorrhagic pneumonia group. In the present study, only six cases performed susceptibility tests to tigecycline, and the susceptibility is 83.3% (5/6).

Risk factors associated with hemorrhagic pneumonia
For patients with SM bacteremia, once hemorrhagic pneumonia develops, the patients' conditions always

Discussion
In the present study, we found two main results. First, the mortality rate of SM bacteremia with hemorrhagic pneumonia was significantly higher than those without hemorrhagic pneumonia. Hemorrhagic pneumonia was an independent risk factor of 30-day mortality in hematologic patients with SM bacteremia. Second, neutropenia, high level of PCT and prior tigecycline therapy were found to be independent risk factors associated with hemorrhagic pneumonia in hematologic patients with SM bacteremia. Hemorrhagic pneumonia is one of the most severe manifestations of SM infections. Almost all patients die within a few days after the onset of hemoptysis and before the isolate is detected from blood or sputum culture. Previous literature has proven that inadequate initial antimicrobial treatment is an independent adverse prognostic factor for SM bacteremia [21], which was confirmed again in our study. Nevertheless, the rate of receiving the adequate initial antimicrobial treatment was similar between hemorrhagic pneumonia group and non-hemorrhagic pneumonia group. This indicates that hemorrhagic pneumonia caused by SM is a severe condition and difficult to save a patient once it develops despite adequate initial antimicrobial treatment. In our study, there were 4 patients with successful treatment  Fig. 2 In vitro antimicrobial susceptibility for SM isolates outcomes after early combination therapy with TMP/ SMX, cefoperazone/sulbactam, moxifloxacin, and/or tigecycline, and rapid recovery of neutropenia. There are only two patients who were diagnosed with hematologic malignancies and hemorrhagic pneumonia due to SM infection and successfully treated after early combined therapy with TMP/SMX, polymyxin, and/or moxifloxacin were reported in the previous study [22]. Both patients had neutropenia during infection, and neutropenia resolved within 1 week after SM bacteremia onset. Rapid recovery of neutropenia depending on hematologic disorder itself and supportive treatment such as granulocyte colony stimulating factor (G-CSF), might be a key factor in reducing mortality of hemorrhagic pneumonia. However, there is still no effective therapy that could prevent life-threatening hemorrhagic pneumonia. The prophylaxis of SM infection and its progression to hemorrhagic pneumonia is thus crucially important. It is necessary to clarify risk factors for hemorrhagic pneumonia. A previous review summarized the clinical information of 30 cases of hemorrhagic pneumonia caused by SM [10]. The authors found that severe neutropenia (< 100 uL) was the most important risk factor for hemorrhagic pneumonia induced by SM. In the present study as well, 24 of the 27(88.9%) SM bacteremia patients with hemorrhagic pneumonia were neutropenic (< 500 uL).
This maybe explain why SM bacteremia is likely to occur in immunocompromised patients, and why hemorrhagic pneumonia induced by SM is always developed in patients with hematologic malignancy. Neutropenia due to the hematologic disorder itself or treatment for the hematologic disorder may be an important factor for susceptibility to SM bacteremia and progression to hemorrhagic pneumonia.
PCT is a prohormone of calcitonin, which has a short half-life and is highly upregulated during the acute phase of sepsis. PCT is valuable and sensitive in predicting bacteremia and distinguishing Gram-negative sepsis in hematologic patients with febrile neutropenia [23][24][25][26][27]. Its concentration is markedly associated with the severity of infection [28]. The biological action of PCT during severe infection is largely unknown. PCT might be a mediator of inflammation, which can be increased by overproduction of early proinflammatory cytokines, such as interleukin (IL)-1β, tumor necrosis factor (TNF)-α and IL-6. The high level of PCT indicated severity of infection, and might be associated with excessive inflammation which can cause vasculitic damage [29]. This maybe explain why patients with elevated PCT was more likely to progress and develop hemorrhagic pneumonia.
Tigecycline is a potential alternative antimicrobial for SM infection, and the susceptibility to tigecycline is  [30,31,32]. However, our study showed that prior tigecycline therapy was an independent risk factor for developing hemorrhagic pneumonia caused by SM, and empirical antimicrobial treatment with tigecycline cannot reduce the mortality rate (63.6%). The reason may be due to the side effects of tigecycline. It was reported by Windhorst et al. that StmPr1, a protease of the subtilase family secreted by SM, can degrade the collagen and fibroblasts in vitro. This process might lead to tissue invasion, destruction of the alveolar micro vessels and hemorrhage [33]. It was previously reported that the use of tigecycline might induce coagulopathy usually manifested as prolongation of PT and APTT and a reduction in the fibrinogen level in a dose-dependent manner [34]. The most severe adverse event is bleeding. This view was confirmed by our study again, which showed a higher incidence of prolonged APTT in patients received prior tigecycline therapy (87.0% vs 56.3%, P = 0.015, χ 2 = 5.908). Furthermore, the majority of enrolled patients occurred thrombocytopenia concurrently in our study. If destruction of alveolar micro vessels by protease secreted by SM, coagulative dysfunction induced by tigecycline, and thrombocytopenia happen at the same time, the probability of hemorrhagic pneumonia occurring might increase. On the other hand, it was identified that the resistance to tigecycline was easily selected during exposure to this antimicrobial and cross-resistance to other antimicrobials, such as aztreonam and quinolones, was also presented [35]. Unfortunately, in our study, there were only six cases finished susceptibility test to tigecycline in vitro, and only one strain showed resistance. The rate of resistance to TMP-SMX was higher in the prior tigecycline therapy group (17.4% vs 6.3%), which was no significantly statistical difference (P = 0.223). However, there may be some bias based on the nature of retrospective study. Tigecycline is a broadspectrum antibiotic used to treat infections that do not respond to first-line treatments. It might be indicated that these patients with prior tigecycline use had suffered from severe infections in a short time before SM bacteremia onset. It was unknown whether this previous infection was associated with hemorrhagic pneumonia. We need to verify the potential association between hemorrhagic pneumonia and prior tigecycline use by expanding the case samples or in a SM-infected mouse model.
The study still has several limitations. First of all, 59.3% of cases did not obtain a positive result of SM in sputum culture in the hemorrhagic pneumonia group. It might lead to the argument about whether SM was the true pathogen of hemorrhagic pneumonia. For this special group of patients, sputum culture was difficult to carry out under certain conditions, such as the sudden appearance of massive hemoptysis. Besides, the majority of these patients had thrombocytopenia, so it was unable to perform bronchoalveolar lavage fluid culture to improve the positive rate of culture. According to our definition, the blood cultures of all patients were positive for SM. Then combined with clinical and imaging findings, we reasonably identified that hemorrhagic pneumonia was caused by SM. On the other hand, the antimicrobial susceptibilities of SM isolates were insufficient. There were only a few cases completed the susceptibility test to tigecycline. Consequently, it is difficult to confirm the relationship between the susceptibility to tigecycline and the development of hemorrhagic pneumonia.

Conclusions
Our study suggested that SM bacteremia with hemorrhagic pneumonia developed in patients with hematologic diseases was a condition with rapid progression and high mortality. In addition, neutropenia, high level of PCT and prior tigecycline therapy within 1 month were independent predictors of hemorrhagic pneumonia in hematologic patients with SM bacteremia, and thus patients with these conditions should be treated carefully. Futher studies are needed to explore effective approaches for prevention of the lifethreatening hemorrhagic pneumonia in these patients.