Risk factors and antimicrobial resistance profiles of Pseudomonas putida infection in Central China, 2010–2017

Abstract The aim of this study was to analyze the risk factors, clinical features, and antimicrobial resistance of Pseudomonas putida (P putida) isolated from Tongji Hospital in Wuhan, China. The data of 44 patients with P putida infections were retrospectively reviewed in this study. All cases of P putida strains were detected by the clinical laboratory of Tongji Hospital in the period of January 2010 to December 2017. Antimicrobial susceptibility testing was conducted using Kirby-Bauer method. Forty-four effective strains of P putida were isolated, including 32 inpatients and 12 outpatients. The 32 inpatients cases were obtained from various departments, which were urosurgery wards (n = 5, 15.6%), pediatrics wards (n = 4, 12.5%), hepatic surgery wards (n = 4, 12.5%), among others. The isolates had been discovered from urine specimens (28.2%), blood specimens (21.9%), sputum specimens (12.5%), and so on. Twenty-five patients had histories of catheterization before the isolation of P putida. Twenty-four patients were in immunocompromised states, 5 patients had undergone surgery, catheterization and were taking immunosuppressive therapy simultaneously. Polymicrobial infections were found in some P putida cases, especially Stenotrophomonas maltophilia, Pseudomonas aeruginosa, and Escherichia coli. All the patients had treated by antimicrobial before culture. Multi-drug-resistant strains were detected in 75% of P putida isolates. The P putida strains were resistant to trimethoprim/sulfamethoxazole (97.7%), aztreonam (88.6%), minocyline (74.3%), ticarcillin/clavulanic acid (72.7%), and sensitive to amikacin (86.4%), imipenem (62.8%), gentamicin (56.8%). Catheterization or other invasive procedures, immunocompromised states, and underlying diseases increased the risks of P putida infections. Moreover, the P putida strains were highly resistant to trimethoprim/sulfamethoxazole, aztreonam, minocyline, ticarcillin/clavulanic acid.


Introduction
Pseudomonas putida, a specialized aerobic organism of the fluorescent group of Pseudomonas species, is a pathogenic bacterium of fish which can also colonize the human throat. [1][2][3] It can be widely found in inanimate hospital surfaces and moist environments because of its strong tolerance to hard living conditions. [3,4] Moreover, P putida can cause infections in hospitals because of its various infection and transmission routes. [5] However, compared with other Pseudomonas species, it was previously thought to be of low pathogenicity. Previous studies have shown that P putida was sensitive to most antimicrobial agents, so clinical cases caused by P putida were uncommon. [6] In recent years, the isolation rate of P putida has been rising yearly, and the emergence of multi-drug-resistant (MDR) strains, even extensively drug-resistant strains (XDR) of P putida had became a cause for concern. [7,8] At present, there are few articles in the literature-most case reports are related to the infections and antimicrobial resistance of P putida, making it difficult for us to analyze the clinical features and the prevalence of P putida resistance. To further understand the infection profiles of P putida and its resistance to common antimicrobials in recent years, we reviewed 44 cases of P putida infected during January 2010 to December 2017 in a large teaching hospital in central China.

Clinical specimen and information collection
Forty-four cases of P putida-infected patients (including outpatients) were identified during January 2010 to December 2017 through a review of the clinical microbiology laboratory records in Tongji Hospital, Huazhong University of Science and Technology, a comprehensive healthcare organization also served as education facility for both Department of Healthcare and Education in Wuhan, China. Then, the inpatients' data, including the age, sex, distribution of wards, underlying diseases, comorbidities, indwelling devices, co-pathogens, drug resistance, and administering of antimicrobial before culture, were collected from the electronic medical records of Tongji Hospital. Finally, we analyzed the clinical features, risk factors, and antimicrobial resistance of the data and finally draw conclusions in the following parts.

Bacterial identification and the antimicrobial susceptibility testing
The bacterial culture procedures were followed by the "National Clinical Laboratory Operation Regulations" (Version 3) and the kit instructions. In addition, we used the Vitek II Compact Automated System (BioMé roués, France) and the Bruner Maldi-Tof MS System Mass Spectrometer (Bruker GmbH, Germany) to identify the P putida strains. Antimicrobial susceptibility was determined for all isolates by the disk diffusion testing (no inhibition zone). Pseudomonas aeruginosa ATCC27853 and Escherichia coli ATCC25922 were used as reference strains for quality control. Inhibition zone diameters were measured and interpreted according to Clinical and Laboratory Standards Institute guidelines criteria. The final results showed sensitive (S), intermediate (I), and resistant (R). The antimicrobial agents involved were as follows: trimethoprim/sulfamethoxazole, ciprofloxacin, gentamicin, amikacin, imipenem, ceftazidime, aztreonam, piperacillin, cefoperazone/sulbactam, levofloxacin, cefepime, piperacillin/tazobactam, meropenem, minocyline, ticarcillin/clavulanic acid, cefoperazone, tobramycin. The susceptibility disc was provided by OX-OID Company. All reagents were qualified before use.

Specimen source and the distributing of P putida
A total of 44 effective strains of P putida were isolated from 32 inpatients and 12 outpatients. The clinical data of the 12 outpatients were not available because they have no records in the electronic medical system. Only the first bacterium episode for each patient was included in the analysis. The clinical data of the 32 inpatients were listed below (Table 1).
Klebsiella spp and Legionella pneumophila were also detected in some cases. In addition, 6 patients (18.8%) detected with >2 pathogens.
The most common clinical manifestation of P putida infection was fever. Patients also showed frequent urination, burning with urination, abdominal pain, diarrhea, tachypnoea, cough, headaches, and among others. An increased white blood cell count, the elevated levels of interleukin-6, procalcitonin, and C-reactive protein were mainly found in laboratory analysis. After an effective treatment, there were no deaths in our study.

Prevalence of P. putida resistance
The 44 strains of P putida (including 12 strains from outpatients) were tested for susceptibility to 17 commonly used antimicrobials. The P putida strains were resistant to trimethoprim/

Discussion
The cases of human infections of P putida had been first reported from blood during 1980 and 1985 in 15 patients with cancer. [9,10] After that, patients with pneumonia, catheter-related bloodstream infections, acute cholecystitis, cholangitis, tonsillitis, thrombophlebitis, skin, and soft tissue infections have been ever reported to be infected with P putida, [5,[11][12][13] Most of studies have shown that P putida, which acted as an opportunistic pathogen, [14,15] often infected patients who were in an immunocompromised state and had a poor physical condition. [10,[16][17][18] Besides, P putida was ever considered as a bacterium with low toxicity, weak pathogenicity, showed a high susceptibility to many antimicrobials, and finally had a good prognosis. However, recent studies indicated that the mortality rate in P putida-infected patients with underlying disease was high (40%), [11,19] which has gradually aroused clinician's concern. Despite the fact that this organism can cause healthcare-related infections, clinical data on P putida infections are relatively lacking owing to the rarity. To date, the literature about P putida-related infections were mostly case reports, and few large case series were found, thus making it hard to analyze the clinical characteristics and the prevalence of P putida resistance.
In this study, we collected 44 cases of P putida infections, which might contained the largest number of P putida infection cases in the literature until now. Among the 32 inpatient cases, most were elderly or children, 24 inpatients (75%) were in immunocompromised states (including tumor, cirrhosis, taking immunosuppressive agents after transplantation, and so on), 25 inpatients (78.1%) had a history of catheterization or catheter insertion (especially indwelling urinary catheter) before the isolation of P putida, which had the same trends with previous studies. [10,11,19,20] Besides, one of the other main ways of P putida invasion was through bloodstream infection. Our results showed that the bloodstream infection of P putida mainly occurred in patients with organ transplants, hematologic diseases, and tumors. In these patients, the therapeutic procedures were required for primary diseases, as well as the poor conditions of the patients significantly increased the risks of P putida infection. As a result, implementing aseptic precaution, enhancing the immunity of patients, and blocking the infection route (device removal) were necessary for reducing infection risk of P putida and shortening the duration of hospitalization during the treatment of susceptible individuals or application of invasive procedures.
As for detection methods of P putida, at early time, the classic strategy for bacterial identification was based initially on fast and simple tests, and performed by using either commercial kits such as miniaturized biochemical tests (API analysis) or automated systems. After that, the use of protein profiles obtained by Matrix-Assisted Laser Desorption Lonization Time of Flight Mass Spectrometry directly from colonies was successfully proposed and developed gradually. Although molecular biology developed in recent years enabled rapid bacterial identification using polymerase chain reaction (PCR), which was one of the most sensitive test, the cost and workload requirements currently preclude their routine use. In our research, we used the Vitek II Compact Automated System and the Bruker Maldi-Tof MS System Mass Spectrometer to identify the P putida strains which identification results were reliable.
In previous reports, clinical isolates of P putida showed low resistance to most antimicrobials. For example, Sader et al, reported that from 1997 to 2003, the resistant rates of P putida to levofloxacin and ciprofloxacin were 20.2% and 21.7%, respectively. [21] Afterward, P putida isolates were usually reported increasing resistance to common antimicrobials, including carbapenem. Our study showed that its resistance rates to trimethoprim/sulfamethoxazole were up to 90%, and quinolones and cefoperazone/sulbactam were >50%. However, Table 2 Antimicrobial resistance of P putida strains to 17 common antimicrobials.    [22] was widely applied in the clinical and agricultural fields, causing the extensive resistance to various bacteria (included P putida certainly). In this study, the rate of resistance of P putida strains to trimethoprim/sulfamethoxazole was >97%, which was consistent with the previous studies. [1,22] In addition, it can be seen from the collected cases that P putida maintained a higher sensitivity to imipenem and amikacin compared with other antimicrobials. Thus, imipenem and amikacin can be used as references for clinical practice. Furthermore, in our study, 24 of the 32 (75%) hospitalized patients were infected with MDR strains of P putida, and the MDR strains showed a broadly resistance trend to 17 common antimicrobial. Among the 24 inpatients, 2 children with acute lymphoblastic leukemia developed resistance to all antimicrobials, which increased the difficulty of treatment, duration of hospital stay, economic burden, and the morbidity of patients. Therefore, when it came to pathogenic infections, a rational selection of antimicrobial agents was critical for patients. Besides, multiantimicrobial combinations and effective surveillance of resistance will reduce the generation of drug-resistant strains and finally improve the prognosis of patients. [20] There are some limitations associated with our study. First, organism identification was identified by using an automated system and was not performed by genotypic-based methods. Along similar lines, demonstration of antimicrobial resistance genes was not performed, and characterization of resistance profiles was carried out based upon disk diffusion data only. However, according to Jacquier et al, 8 of 9 P putida isolates identified by 16S rRNA gene sequencing were confirmed accurately by using the Vitek II Compact Automated System and there was no misidentification. [23] The common microbials mentioned in this article, such as P putida, were not difficult to identify. It must be kept in mind that this manuscript was not intended to provide a detailed microbiological analysis but rather was meant to be a broad survey of the isolation patterns and susceptibility profiles of P putida. Second, this study was retrospective and had a limited number of cases. Because of its retrospective nature, it was not possible to confirm the pathogenic role of all of the identified isolates, and we failed to exclude factors associated with other pathogens in polymicrobial infection. Moreover, we could not fully avoid contaminants of P. putida from other sources, such as the environment and endogenous sources. Furthermore, this study was a single-center; the results obtained from this study were not generalizable enough. Ideally, a multicenter study is essential and meaningful for future research to determine whether these results represent a local or global phenomenon. Despite these limitations, this study provided risk factors, clinical characteristics, and antimicrobial susceptibility of P putida infection.

Conclusion
This study demonstrated that P putida infections, mostly presented as polymicrobial infections, were predisposed to patients with underlying diseases, immunocompromised state, a history of catheterization, or other invasive procedures. The P putida strains had showed high resistance rates to most antimicrobials, such as trimethoprim/sulfamethoxazole, aztreonam, minocyline, ticarcillin/clavulanicacid, cefoperazone/ sulbactam, ciprofloxacin, cefoperazone, and so on.