Salmonella enterica subspecies arizonae infection of adult patients in Southern Taiwan: a case series in a non-endemic area and literature review

Background The majority of Salmonella arizonae human infections have been reported in southwestern United States, where rattlesnake-based products are commonly used to treat illness; however, little is known in non-endemic areas. We reviewed and analyzed the clinical manifestations and treatment outcomes in adult patients with S. arizonae infection at our institution. Method A retrospective study was conducted at a regional teaching hospital in southern Taiwan from July 2007 to June 2014. All adult patients diagnosed with S. arizonae infections and treated for at least three days at Chia-Yi Christian Hospital were included. Patients were followed till discharge. Results A total of 18 patients with S. arizonae infections (median age: 63.5 years) were enrolled for analysis, of whom two thirds were male. The three leading underlying diseases were diabetes mellitus, peptic ulcer disease and malignancy. Ten patients had bacteraemia and the most common infection focus was the lower respiratory tract. Most of the patients (72.2%) received third-generation cephalosporins as definitive therapy. In contrast, ampicillin-based regimens (accounting for 45.2%) were the major treatment modalities in previous reports. The crude in-hospital mortality was 5.6%, which was much lower than what was previously reported (22.7%). Conclusions Though uncommon, there were cases of S. arizonae infections in Taiwan. Patients receiving third-generation cephalosporins treatment had better prognosis compared with those treated with ampicillin-based regimen.

Cold-blooded animals are the usual habitats of S. arizonae, especially in reptiles or rattlesnake-based products. Other animals, including poultry, rats, and dogs [6,9,17], have also been involved in human infection by S. arizonae. According to previously published articles [1,10,13,14,16,[25][26][27][28][29]31], the geographic distributions of S. arizonae human infections are mainly located in south-western United States, where the use of rattlesnake-associated products to treat a wide variety of illnesses is popular among Mexican-Americans populations. Around 20 pediatric patients below 18 years of age suffering from S. arizonae infection have been reported in the literature [9,17,21,23]. Little is known about S. arizonae infection of adult patients in Asian countries, including Taiwan. The aim of this study was to analyze all adult patients with S. arizonae infection at a regional teaching hospital in southern Taiwan and to perform a literature review on similar patients.

Methods
From July 2007 to June 2014, all adult patients (≥18 years) diagnosed with S. arizonae infection and treated for at least 3 days at Chia-Yi Christian Hospital (CYCH, a regional teaching hospital with a capacity of 1000 beds in southern Taiwan) were retrospectively enrolled. S. arizonae infection was defined as positive cultures of any kind of clinical specimens, including blood, pleural effusion, ascites, urine, sputum, stool, pus and bone, for S. arizonae plus the presence of signs of systemic or local infection. All types of specimens were collected, transported, and processed according to the suggestion described previously [33], and then inoculated into the corresponding culture media for subsequent incubation [34]. The blood culture system at CYCH was Bactec FX system (Becton, Dickinson and Co. [BD], Franklin Lakes, NJ). All bone tissues were collected by bone biopsy or surgical procedures. Potential pathogens were identified by Vitek version 2.0 (bio Merieux Suisse S.A., Geneva, Switzerland), and Salmonella isolates were confirmed by serologic testing (Difco TM Antiserum Solutions). The culture-positive cases were identified by reviewing microbiology records at CYCH.
A standardized case report form was used to collect the demographics, clinical and laboratory data and treatment outcomes. Patients who used H 2 -receptor antagonists or proton pump inhibitors within one month prior to admission were defined as having peptic ulcer disease. Leukocytosis was defined as white blood cell counts exceeding 10 K/μl and thrombocytopenia was defined as platelet count below 150 K/μl. The infection foci of non-bacteraemic patients and patients with secondary bacteraemia were determined if there was a presence of clinical symptoms or signs of infection and isolation of S. arizonae from related clinical specimens. Bacteraemia without an obvious infection source or related to intravascular catheter infection or vascular lesions was classified as primary bacteraemia.
The antimicrobial susceptibilities to chloramphenicol, ciprofloxacin, trimethoprim/sulfamethoxazole, ampicillin and ceftriaxone were determined using the disk diffusion method according to the recommendation of the Clinical and Laboratory Standards Institute (CLSI) [35]. The results were also interpreted using the criteria suggested by CLSI [36]. Antimicrobial agents given before the susceptibility results were defined as empirical therapy, whereas definitive therapy was defined as effective antibiotic therapy prescribed according to the results of final blood cultures and susceptibility testing. The study was approved by the ethics review boards of the hospital [Chia-Yi Christian Hospital-Institutional Review Board (CYCH-IRB) No. 104035, 06/30/2015]. The IRB waived informed consent due to the retrospective study design and the research posing no more than minimal risk.
Continuous variables were described as medians with interquartile ranges (IQR) and categorical variables were described as percentage.

Results
A total of 485 adult patients with Salmonella species infection were identified during this seven-year period, and only 4.7% (23/485) of those patients suffered from S. arizonae infection. Five among them were excluded because they received treatment for less than 3 days at CYCH and did not follow-up subsequently. For the five eliminated patients, the median age was 65 years (IQR, 60-72 years). Four of them had primary bacteraemia and one had pneumonia. The demographics, clinical features, laboratory data and treatment outcome of the enrolled 18 patients are showed in Table 1. Of these 18 patients, the median age was 63.5 years, ranging from 27 to 81 years, with 12 men and six women. All patients lived in Chiayi City, except one, who lived in Yunlin County. All patients had various underlying diseases, including endocrine diseases in 12 (66.7%; diabetes mellitus in ten, and another diseases in two), peptic ulcer disease in eight (44.4%), malignancy in eight (34.8%; hepatocellular carcinoma in four, lung cancer in three, and colon cancer in one), hypertension in seven (38.9%), liver cirrhosis in six (33.3%), chronic viral hepatitis in six (33.3%; hepatitis B virus in five, and hepatitis C virus in two), chronic kidney disease in three (16.7%), autoimmune diseases in one, and acquired immunodeficiency syndrome in one (5.6%).
Blood cultures were obtained in all 18 patients, and bacteraemia was identified in ten (55.6%), including primary bacteraemia in two, both with mycotic aneurysm and secondary bacteraemia in eight. The most common infection foci of patients with secondary bacteraemia and non-bacteraemic clinical syndromes were lower respiratory   Table 2. Patients in the present study were either elderly male or had various underlying conditions (Table 1) which would compromise their cell-mediated immunity. In the literature (Table 2), the most commonly reported comorbidities associated with S. arizonae infections were immunocompromised status, including connective tissue diseases under steroid therapy (40.9%), malignancy (27.3%) and acquired immunodeficiency syndrome (13.6%) [15,17]. Our case series produced similar findings. Additionally, we identified three more associated host underlying conditions, including type 2 diabetes mellitus, liver cirrhosis, and peptic ulcer disease. Uncontrolled diabetes and liver cirrhosis have been shown to cause impairment of humoral-and cellmediated immunity, which play important roles in clearing Salmonella [37]. Therefore, these two diseases would reasonably predispose to the development of S. arizonae infection. Moreover, patients with peptic ulcer diseases received acid suppressants to treat their diseases. Usage of acid suppressants would not only decrease the acidity of gastric juice, which in turn might result in intestinal bacterial overgrowth, facilitate bacterial translocation from the intestine and lead to infection via the gastro-intestinal route [38], but also reduce the gastric acid barrier with subsequent infection despite a lower inoculum of bacteria. Similar findings have been previously reported by Wu et al. [39].
A variety of clinical manifestations were displayed in our series, including enterocolitis, bacteraemia, vascular infection and localized infections. The rank order of infection syndromes in our reports was bacteraemia, intraabdominal infection, and pulmonary infection. In contrast, bacteraemia [1, 7, 10, 13-15, 17-19, 21, 22, 26, 28, 30-32], intra-abdominal infection [4, 6, 10-12, 14, 17, 21, 24, 27, 29, 30, 32, 40] and bone or joint infection [8,9,11,12,14,20,21,24] were the most common clinical manifestations in previously-reported patients ( Table 2). Much fewer patients with bone and joint infection (16.7%) were noted in our present study. Interestingly, both of the two patients in our series diagnosed as S. arizonae related mycotic aneurysm with bacteraemia had a past history of hypertension. This is similar to the result by Wang JY et al., who demonstrated that hypertension was the major factor predisposing to S. choleraesuis mycotic aneurysm [41]. Importantly, our study is the first one to demonstrate that soft tissue could be the infection focus of S. arizonae, which was observed in two of our patients. Overall, the difference in clinical manifestations between our study and prior reports might be due to the small number of patients enrolled in every study. Therefore, further study is needed to clarify whether geographic variance or other factors were associated with the difference of clinical syndromes.
All S. arizonae isolates collected in the present study were susceptible to all five of the recommended anti-Salmonella agents, including chloramphenicol, ciprofloxacin, trimethoprim/sulfamethoxazole, ampicillin, and ceftriaxone. These five antimicrobial agents have the ability to penetrate host cells, which is crucial in killing intra-cellular pathogens, such as Salmonella. All of our patients received at least one of these five agents as their definite therapy, among which third-generation cephalosporins were prescribed for the majority of patients (72.2%, Table 1). Instead, ampicillin was usually chosen as the backbone of the treatment modality (45.2%) in previously published articles ( Table 2). Although the difference was not statistically significant (p = 0.1 by chi-square test), the crude in-hospital mortality rate was only 5.6% in the present study, which was much lower than that (10/44, 22.7%) of the prior reports (Table 2). In particular, 6 of the 10 fatal patients in prior reports received ampicillin as their treatment against S. arizonae infection. Therefore, a third-generation cephalosporin could potentially be a better choice for treating S. arizonae infection compared to ampicillin. However, further investigation is needed. One patient in our study received piperacillin/tazobactam, which is not the recommended antibiotic as the definitive treatment against Salmonellae , and had a favorable outcome. Using piperacillin/tazobactam for Salmonella infection was rarely reported in previous articles. Bell SD demonstrated that minimal inhibitory concentration of piperacillin/tazobactam for S. arizonae isolated from this study was ≦4 μg/mL [4], implicating the bacterium susceptible to it in-vitro. Gerada et al. reported a liver transplant recipient with Salmonella related infectious aortitis and bacteraemia, who responded well to piperacillin/tazobactam treatment [42]. Thus, piperacillin/ tazobactam might be considered as one of the therapeutic options for S. arizonae infection.
During the 1980s, S. arizonae infection became an important issue in public health due to the emergence of many severe infection cases, and its association with extensive use of rattlesnake-based products [10,19,28] mainly in areas with large Mexican-American populations in southwestern United States [1,7,13]. Approximately 70% of the 44 patients reported in previous studies (Table 2) mentioned a history of exposure to reptiles, especially snakes. Rattlesnake-based products are not common in Taiwan and we could not identify any specific animal contact history from the medical records of those patients. It is important to identify whether natural habitats of S. arizonae are present and whether S. arizonae infection is one of the zoonoses in Taiwan. Further studies are needed to identify the possible sources of this infection.

Conclusions
Our study showed that S. arizonae infection, although uncommon, is present in Taiwan, an area outside of typical endemic areas. In addition to previously reported risk factors, usage of acid suppressants, such as proton pump inhibitors and H 2 blockers, might also predispose to S. arizonae infection. To treat S. arizonae infection,