Methods.Peking Union Medical College Hospital (PUMCH), an 1800-bed hospital, consists of 2 campuses that house different medical departments. We retrospectively reviewed all culture-proven cases of listeriosis occurring at PUMCH between 1999 and 2011. Point estimates and 95% confidence intervals are presented.

Results.There were 38 patients with listeriosis: 5 neonatal, 8 maternal, and 25 nonmaternal. The median age of the adult nonmaternal patients was 47 (range, 18–79) years with a female predominance (72%). Forty percent (n = 10) had an underlying rheumatic disease. Forty-four percent of cases (n = 11) were healthcare-associated infections occurring a median of 20 (range, 3–44) days after hospital admission. Only 2 of the 11 healthcare-associated cases clustered in space and time. One healthcare-associated case occurred in a patient receiving KHI-272 therapy, an oral, irreversible dual EGFR/HER2 inhibitor. The neonatal and maternal listeriosis cases were similar to those reported in the literature.

Gastroenteritis, bacteremia, and meningitis are the most common manifestations of listeriosis. Because L. monocytogenes has a strong predilection for elderly and immunocompromised persons [15–18], results in poor fetal outcomes [19–21], exhibits poor response to third-generation cephalosporins, and is associated with a high mortality rate, it has become an increasingly important emerging infectious disease [22].

In the United States, L. monocytogenes is the fourth causative microorganism of bacterial meningitis [23]. Among persons aged >65 years, L. monocytogenes is the third leading pathogen [24, 25]. Most listeriosis cases have been reported from industrialized Western countries. Reports from East Asia and developing countries are scarce [26, 27].

Our goal was to retrospectively review all culture-proven cases of listeriosis at Peking Union Medical College Hospital (PUMCH) since 1999 and describe the clinical characteristics and outcomes of the infected patients.

We retrospectively identified all patients with L. monocytogenes infections based on a list generated from an electronic database in the clinical microbiology laboratory at PUMCH. All positive culture results for L. monocytogenes diagnosed at PUMCH since 1999 are stored in the database. We included all cases from January 1999 to October 2011. Clinical data from the identified cases were abstracted from the medical records. These data included demographic characteristics, comorbidities, known risk factors (immunosuppressive therapy, dietary history, travel, and exposures), the sites from which the organism was isolated, clinical presentation, laboratory data, type of antimicrobial therapy, duration of hospitalization, and outcomes.

The diagnosis of listeriosis was based on one of the following: isolation of L. monocytogenes from normally sterile clinical specimens (eg, cerebrospinal fluid [CSF], blood, amniotic fluid, uterine swab); isolation of L. monocytogenes from nonsterile specimens (eg, rectal swab, tracheal swab); and histopathology compatible with listeriosis [22]. Cases were categorized as neonatal, maternal, or nonmaternal infections. All maternal cases were in pregnant women who had L. monocytogenes isolated from cultures of normal sterile body sites or vaginal swab [19]. Healthcare-associated cases were defined as onset of listeriosis symptoms >48 hours after admission for medical conditions other than listeriosis.

It is intriguing to speculate whether these healthcare-associated cases were the result of in-hospital acquisition, or whether this was the result of colonization. Until this retrospective case series was conducted, we had absolutely no insight about the frequency of these healthcare-associated cases. The cases were not clustered in time or space so they did not elicit additional surveillance. Although we could not perform PFGE on the specimens from our study, the majority did not cluster in time or space, suggesting that a common source was unlikely. Investigators have recognized for >20 years that L. monocytogenes can be carried in the gastrointestinal tract [29–31]. Listeria monocytogenes can be isolated in the stool of 1%–10% of the population, where it can persist without causing symptoms [32]. Using repeated sampling, Listeria can be detected in the feces of nearly 70% of healthy nonpregnant individuals and 44% of pregnant women [21, 31]. MacGowan et al found that Listeria was isolated from 5.6% (10/177) of renal transplant recipients on 1 or more occasions over the period of a year; moreover, >1 species or serovar of listeria can be isolated from 40% of fecal carriers, and no cases of clinical infection occurred in any fecal carriers [33]. Fecal, cervicovaginal, and oropharyngeal carriage of L. monocytogenes has been reported as a possible predisposing factor for perinatal listeriosis [34, 35]. In one study conducted by Schuchat et al [36], asymptomatic carriage of the illness-associated strain of L. monocytogenes was identified in nearly one-fifth of household contacts of patients with sporadic listeriosis, and no cases of secondary disease were detected within households in this study. Their findings suggest that gastrointestinal carriage of pathogenic strains of L. monocytogenes is not uncommon in contacts of cases, underscoring the critical role that host susceptibility plays in determining whether illness occurs following exposure to this organism. All of our cases of healthcare-associated listeriosis had severe underlying immunosuppression. Besides immunosuppression, many of our patients had underlying diseases involving the gastrointestinal tract, or their therapy could impact the integrity of the intestinal mucosa. So, the role that gastrointestinal colonization of Listeria played in the pathogenesis of these healthcare-associated infections warrants further study.

After the discovery of these nonclustered healthcare-associated cases, we have implemented a more aggressive approach: all healthcare-associated cases will be thoroughly investigated for both prehospital and in-hospital exposures. We are also saving all bacterial isolates for DNA fingerprinting. This more aggressive approach may help us better define the source of these infections.

Among the healthcare-associated listeriosis cases, one patient with diffuse metastatic breast cancer experienced sudden onset of fever, oral ulcers, and diarrhea after 3 days of HKI-272 treatment (Table ​(Table4,4, patient 26). Blood culture yielded L. monocytogenes. The HKI-272 therapy was discontinued and antibiotic treatment was initiated, and the patient fully recovered. HKI-272, also known as neratinib [37], is an oral, irreversible dual EGFR/HER2 inhibitor for breast and non-small-cell lung cancer. Phase 1 and 2 studies reported gastrointestinal adverse events, including diarrhea (89%), nausea (29%–64%), and vomiting (23%–50%). Approximately 30% of patients required discontinuation or dose reduction due to severe diarrhea. Cases of listeriosis were reported among patients undergoing therapy with other biologic agents such as infliximab (antitumor necrosis factor agents) [38–41], etanercept (a tumor necrosis factor antagonist) [42], and trastuzumab (a monoclonal antibody against the HER2 receptor) [43].

Forty percent of our cases had underlying rheumatologic diseases. This proportion is higher than what was previously reported in the literature [38]. Although PUMCH does not specifically specialize in the treatment of rheumatic diseases, we do have a large population of such patients. Persons of Asian descent have a higher incidence of SLE, compared with other races [44–46]. Given the paucity of published reports on L. monocytogenes from East Asia, this may explain the higher incidence among patients with rheumatic diseases in our report. This may also have impacted the sex distribution of cases. Traditionally, L. monocytogenes has been reported more often among men than women. The male to female ratio in our study was 1:1.8. This may reflect the increased predisposition of rheumatic diseases among women [47–49].

Comorbidity plays a very important role in the prognosis of listeriosis [18]. Eighty-one percent of 225 patients with listeriosis studied in France had a predisposing immunocompromising condition, whose severity was the major prognostic factor [17]. In our population, 92% of nonmaternal listeriosis cases were immunosuppressed.

Our cases of infant listeriosis mirrored the cases reported in the literature, as did their outcomes. We did not observe any late-onset cases of infant listeriosis, as reported by other authors [9, 22, 50–52]. Similarly, the characteristics of our maternal listeriosis were similar to those reported in the literature.

This study has several limitations. First, it is a retrospective assessment over a protracted timespan. As such, we were unable to obtain specimens for molecular testing, and we were unable to clarify additional issues relating to certain in-hospital epidemiological exposures. Second, it consists of a relatively small sample size, and our findings may not be necessarily generalizable to other populations or settings. Third, cases of listeriosis in China are not routinely reported to public health authorities. As such, the epidemiology of listeria is not well defined. Our case series reflects a selection bias toward hospitalized (ie, sicker) patients and may not reflect the overall epidemiology of listeria.