Recurrent multidrug-resistant Salmonella enterica serovar Typhimurium bacteremia in a returned traveller

Abstract

This case report describes a 68-year-old male with recurrent multidrug-resistant Salmonella enterica serovar Typhimurium bacteremia acquired during travel abroad. He experienced a recurrence of bacteremia without a clear source and was successfully treated with 10 weeks of intravenous ertapenem. Post hoc genome sequencing revealed an isolate bearing class A, C, and D extended-spectrum β-lactamases (ESBLs). A review of English- and French-language literature since 2000 revealed eight publications that discussed recurrent S. enterica serovar Typhimurium bacteremia. Patients with multidrug-resistant S. enterica serovar Typhimurium should be monitored frequently for recrudescence, even in the absence of risk factors.

Case Presentation

A 68-year-old male was admitted to hospital with a 5-day history of subjective fever after returning 2 weeks prior from Kinshasa, Democratic Republic of the Congo (DRC) for 3 months. Prior to that, he spent 3 months in Gujarat, India. In both countries, the patient had stayed with relatives in urban centres. He denied myalgias, headache, nausea, vomiting, diarrhea, gastrointestinal symptoms, rash, or febrile illness while travelling abroad. His subjective fevers occurred daily and did not have a tertian or quartan pattern. He also endorsed malaise, intermittent chills, non-productive cough, and a decrease in his baseline functional ability. Review of animal exposures, recent hospitalizations, or sick contacts—including tuberculosis—was negative. He endorsed mosquito exposure in the DRC and was not on malaria prophylaxis.

His medical history was significant for malaria, diagnosed at age 35 years and treated with sulfadoxine/pyrimethamine. Other history included hypertension, dyslipidemia, chronic obstructive pulmonary disease, and previously treated prostate cancer on annual surveillance. His home medications included acetylsalicylic acid, candesartan, hydrochlorothiazide, pantoprazole, and fluticasone/vilanterol. He had a 35 pack-year smoking history but denied alcohol consumption or any recreational drug use.

On admission, the patient was afebrile (37.2°C) and vitals were otherwise unremarkable, except for hypotension (82/46 mmHg). A complete blood count, electrolyte panel, liver function tests, and lactate level were within normal limits; however, the C-reactive protein was elevated at 40.5 mg/L. The infectious differential diagnosis included malaria, dengue fever, rickettsial infection, and Salmonella. A chest radiograph was normal. Abdominal ultrasound did not show hepatosplenic abscesses or other foci of infection. He was HIV negative; thick and thin smears did not demonstrate Plasmodium spp. While a hemoglobinopathy work-up was not done, a thick and thin smear did not show evidence of erythrocyte sickling. The differential cell count revealed a normal lymphocyte count that did not trigger further work-up for immunodeficiencies. The patient was started empirically on ceftriaxone.

The patient became febrile on day 2 of admission, with a maximal temperature of 38.8°C. Three sets of blood cultures grew S. enterica serovar Typhimurium, with time-to-positivity of 14, 15, and 18 hours, respectively. At the local microbiology laboratory, S. enterica serovar Typhimurium identification was made using serologic methods and matrix-assisted laser desorption/ionization-time of flight mass spectrometry (MALDI-TOF MS). The clinical isolate agglutinated O poly-A-S antiserum but did not agglutinate with Vi antiserum; MALDI-TOF MS provided a 99.9% probability identification of the isolate as S. enterica subsp enterica. The positive blood cultures were reported to the local public health unit, and the samples were sent to the provincial public health laboratory for confirmatory testing and speciation. At the public health laboratory, serologic methods were used, which demonstrated that the clinical isolate could be classified as per the Kauffmann–White system as 1,4,[5],12:i:1,2—consistent with S. enterica serovar Typhimurium (1).

Given the positive blood culture results, the patient was placed in contact precautions, his antimicrobial therapy was broadened to IV piperacillin-tazobactam, and the Infectious Disease service was consulted. While on piperacillin-tazobactam, the patient’s symptoms—including his fever—persisted.

On day 5 of admission, the clinical isolate was found to be a multidrug-resistant, extended-spectrum β-lactamase (ESBL)-producing strain of S. enterica serovar Typhimurium only susceptible to carbapenem antibiotics. Minimum inhibitory concentrations of the clinical isolate and methods used to determine antimicrobial susceptibility are outlined in Table 1. Double disk diffusion method confirmed an ESBL-producing strain.

Table 1:

Antimicrobial susceptibilities of the patient’s clinical S. enterica serovar Typhimurium isolate

Antimicrobial	Method of susceptibility testing	CLSI M100 breakpoint (µg/mL)*	Isolate MIC	Interpretation
Ampicillin	VITEK 2	≤8	≥32 µg/mL	Resistant
Azithromycin†	Agar dilution	N/A‡	≤8 mg/L	N/A‡
Ceftriaxone	VITEK 2	≤1	≥64 µg/mL	Resistant
Ciprofloxacin	ETEST	≤0.25	2 µg/mL	Resistant
Ertapenem	VITEK 2	≤0.5	≤0.5 µg/mL	Susceptible
Trimethoprim-sulfamethoxazole	VITEK 2	≤2/38	≥320 µg/mL	Resistant

Note: VITEK 2, ETEST (bioMérieux, Marcy-l’Étoile, France)

* CLSI M100 Performance Standards for Antimicrobial Susceptibility Testing, 30th edition (19)

† Azithromycin MIC was determined by the provincial public health laboratory

‡CLSI M100 (30th ed.) does provide clinical breakpoints for macrolides based on limited data for S. enterica serovar Typhi only; these breakpoints are not interpretable for S. enterica serovar Typhimurium.

CLSI = Clinical and Laboratory Standards Institute; MIC = Minimum inhibitory concentration

The patient’s prior antimicrobial therapy was discontinued, and he was broadened to intravenous ertapenem 1 g every 24 hours. Computed tomography (CT) of the chest, abdomen, and pelvis with IV contrast did not demonstrate aortitis or vertebral foci of infection, but did show mild distension and edematous wall thickening of the gallbladder, as well as inflammatory changes in the adjacent fat (Figure 1). A surgical opinion did not recommend cholecystectomy. He had a normal transthoracic echocardiogram and no foci of metastatic infection. Magnetic resonance imaging (MRI) of the thoracic spine was performed due to mid-thoracic back pain that developed during the patient’s first 7 days of admission; this was negative for osteomyelitis. Blood cultures became sterile on hospital day 7, after receiving three doses of ertapenem.

Figure 1:

Computed tomography of the chest, abdomen, and pelvis with IV contrast obtained on day 5 of admission, illustrating inflammatory changes adjacent to the gallbladder neck (circled)

Outcome and Follow-up

The patient was discharged 12 days after admission and was afebrile with negative blood cultures, but he still had intermittent, migratory pain at the level of T8–9. He continued IV ertapenem through a peripherally inserted central catheter (PICC) at home with a planned treatment duration of 4 weeks.

At initial follow-up 1 week post-discharge, he denied constitutional symptoms, but had ongoing, intermittent pain at the level of T8-9 that migrated to his left ribs. His rib pain was reproducible to palpation; however, his back pain was not. A radiograph of his ribs did not illustrate any focus of infection or osteolytic lesion.

By his second follow-up visit, he had completed 4 weeks of IV ertapenem. At the conclusion of his antimicrobial therapy his symptoms had improved substantially. Follow-up blood cultures were arranged to be drawn 1 week after cessation of ertapenem.

His repeat blood cultures off antibiotics were positive for recurrent S. enterica serovar Typhimurium infection. A CT chest–abdomen–pelvis with oral and IV contrast did not demonstrate any source for his recurrent bacteremia, including no aortitis or vertebral foci. The inflammatory changes in the gallbladder seen initially were not visible on this follow-up imaging. He was restarted on ertapenem 1 g IV q24h for an additional 6 weeks.

At his final visit, after a total of 10 weeks of parenteral therapy, he was back to his functional baseline. Antimicrobial therapy was discontinued, and follow-up blood cultures taken 1 week after ertapenem cessation did not demonstrate growth. He was discharged from our care.

Discussion

Invasive non-typhoidal Salmonella (iNTS) is a problem of epidemic proportions, causing approximately 535,000 cases worldwide in 2017 (2). The emergence of ESBL resistance is a growing concern among imported cases in Canada. While the frequency of ceftriaxone resistance among endemic non-typhoidal Salmonella has decreased from 5.7% in 2012 to 3.5% in 2017 (3), the emergence of ESBLs among imported cases in Canada remains a growing concern as it casts doubts on the use of third-generation cephalosporins as empiric treatment for imported cases. Public Health Agency of Canada data illustrate the swift increase in multidrug- and extensively drug-resistant Salmonella infections (3). Among all non-typhoidal Salmonella isolates from any clinical source in Canada, 2017 statistics show 15% resulted from S. enterica serovar Typhimurium infections, with 6% of total clinical non-typhoidal Salmonella isolates being identified from blood (3). Of non-typhoidal Salmonella isolates from any source, 13% were resistant to three or more classes of antimicrobials, with increasing resistance rates to nalidixic acid, a quinolone used to predict in vitro ciprofloxacin activity (from 6.3% in 2012 to 19% in 2017), and streptomycin (from 9% in 2012 to 18% in 2017) (3).

Post hoc genome sequencing revealed the extent of antimicrobial resistance harboured by our patient’s clinical isolate of S. enterica serovar Typhimurium (for the whole genome sequence in FASTA format, see Supplemental Appendix 1). This was further confirmed using the Salmonella In Silico Typing Resource (SISTR), which predicts Salmonella serovars from whole genome assemblies using the antigen gene and core genome multilocus sequence typing (cgMLST) (4). SISTR called this isolate as “Salmonella enterica subsp enterica serovar Typhimurium” with MLST 313, Serovar (antigen) Typhimurium, Serogroup B, H1 = i, H2 = 1,2. These in silico results corroborate the serologic identification at the provincial public health laboratory. SISTR contains information on 22 isolates of MLST 313, but only one had associated epidemiological data point sampled from Malawi on an unknown date. As such, we are unable to make any predictions regarding the country of acquisition of this strain during the patient’s travels.

Table 2 outlines 38 resistance genes identified by running sequence data against the Comprehensive Antibiotic Resistance Database (CARD) using the Resistance Gene Identifier software (5). Of note, class A (TEM-1, CTX-M-15) and class D (OXA-1) ESBLs, as well as the class C ampC1 β-lactamase, were identified. The presence of mdtK, CRP, acrAB fluoroquinolone efflux pumps also contributed to the underlying extensive drug resistance pattern.

Table 2:

Antimicrobial resistance genes identified from sequenced S. enterica serovar Typhimurium clinical isolate according to the CARD Resistance Gene Identifier software (5)

Antimicrobial resistance gene	Resistance mechanism	Antimicrobial classes affected	% Identity of region matching CARD reference sequence	% Length of CARD reference sequence
AAC(6’)-Ib-cr	Inactivation	Fluoroquinolones, aminoglycosides	100	107.5
AAC(6’)-Iaa	Inactivation	Aminoglycosides	100	100
mdsB	Efflux	Penicillins, cephalosporins, carbapenems, monobactams	100	100
mdsA	Efflux	Penicillins, cephalosporins, carbapenems, monobactams	100	100
golS	Efflux	Penicillins, cephalosporins, carbapenems, monobactams	100	100
sdiA	Efflux	Cephalosporins, fluoroquinolones, glycylcycline, tetracycline, rifamycin	100	100
sul2	Target replacement	Sulfonamides	100	100
TEM-1	Inactivation	Penicillins, cephalosporins, monobactams	100	100
CTX-M-15	Inactivation	Cephalosporins	100	100
OXA-1	Inactivation	Penicillins, cephalosporins	100	100
QnrB1	Target protection	Fluoroquinolones	100	100
catI	Inactivation	Phenicols	100	99.1
mdsC	Efflux	Penicillins, cephalosporins, carbapenems, monobactams	100	96.4
MdtK	Efflux	Fluoroquinolones	99.8	100
APH(3”)-lb	Inactivation	Aminoglycosides	99.6	100
APH(6)-ld	Inactivation	Aminoglycosides	99.6	100
aadA	Inactivation	Aminoglycosides	99.6	98.5
EF-Tu(R234F)	Target alteration	Elfamycins	99.5	96.4
dfrA14	Target replacement	Diaminopyrimidines	99.4	101.2
CRP	Efflux	Penicillins, fluoroquinolones, macrolides	99.0	100
bacA	Target alteration	Peptide antibiotics	97.1	100
cpxA	Efflux	Aminoglycosides	96.9	100
GlpT(E448K)	Target alteration	Fosfomycin	96.9	100
baeR	Efflux	Aminoglycosides	96.7	100
msbA	Efflux	Nitroimidazoles	96.4	100
soxR	Target alteration, efflux	Penicillins, cephalosporins, fluoroquinolones, glycylcycline, tetracycline	96.0	98.7
UhpT(E350Q)	Target alteration	Fosfomycin	95.7	100
emrB	Efflux	Fluoroquinolones	95.5	100
soxS	Target alteration, efflux, reduced permeability	Penicillins, cephalosporins, fluoroquinolones, monobactams, glycylcycline, cephamycins, tetracycline	95.3	100
marA	Efflux, reduced permeability	Penicillins, cephalosporins, fluoroquinolones, carbapenems, monobactams, glycylcycline, tetracycline	95.2	100
H-NS	Efflux	Penicillins, cephalosporins, fluoroquinolones, macrolides, tetracycline	94.9	100
acrB	Efflux	Cephalosporins, fluoroquinolones, glycylcycline	94.7	100
tet(A)	Efflux	Tetracyclines	94.4	94.1
emrR	Efflux	Fluoroquinolones	93.1	100
marR	Target alteration, efflux	Penicillins, cephalosporins, fluoroquinolones, glycylcycline, tetracycline	92.4	100
ampH	Inactivation	Cephalosporin	91.7	100
acrA	Efflux	Penicillins, cephalosporins, fluoroquinolones, glycylcycline	91.7	100
kdpE	Efflux	Aminoglycosides	91.7	100
ampC1	Inactivation	Penicillins, cephalosporins	72.9	99.4

CARD = Comprehensive Antibiotic Resistance Database

Our patient’s case was among the one-third of iNTS bacteremias without an obvious risk factor such as active malarial or HIV infection, active malignancy, or immunosuppression (6). There is extensive literature demonstrating a link between HIV infection and the establishment of intracellular niches for persistent iNTS infection (7–10), primarily driven through inhibition of anti-LPS IgG production as well as the loss of IL-17-producing CD4+ T cells in the gut mucosa (11). In our HIV-negative patient, given that we were unable to find a source for his bacteremia, a similar mechanism of intracellular niche establishment in the gastrointestinal tract may have driven recrudescence and relapsing bacteremia soon after antibiotics were stopped at 4 weeks of therapy.

In patients with iNTS, typical foci causing recrudescence include vertebral, renal, endovascular, or gastrointestinal sources. In our patient’s initial imaging, there was concern for a potential nidus of infection in the gallbladder, given its thickened appearance and adjacent fat stranding. A surgical opinion did not recommend source control at the time given the lack of clinical syndrome consistent with cholecystitis, normal liver enzymes over the course of admission to hospital, and lack of inflammatory gallbladder findings on follow-up imaging during recrudescence. As a result of the aforementioned factors, we do not feel that the gallbladder represented a feasible nidus of infection resulting in recrudescence, and was an incidental finding on the initial CT imaging.

We searched PubMed using MeSH terms for recurrence, Salmonella, and bacteremia in order to determine the optimal length of therapy for iNTS bacteremia. Our search revealed 67 articles, of which we reviewed only those published after 2000 (n = 10). Of 10 retrieved articles, two were excluded for describing patients without bacteremia. Characteristics of eight reviewed studies (8–10, 12–17) are summarized in Table 3. The decision to restrict reviewed studies to the past 20 years was due to drastic changes in Salmonella antimicrobial susceptibility patterns from the period of 1980–1999 versus the period from 2000 onward. Our literature review did not demonstrate a strong foundation of evidence for the treatment length or recrudescence rate in patients without risk factors for infection.

Table 3:

Articles (n = 8) outlining duration of therapy and recrudescence rate in recurrent S. enterica serovar Typhimurium bacteremia

Author (year published)	Publication type	No. of patients (n)	Study setting	Patient demographics	Duration of therapy	Recrudescence rate
Forrest (2009)	Case series	26	USA	Adults with HIV	28 days	15% at 14–180 days post-discharge
Gordon (2002)*	Case series	100	Malawi	Adults with HIV	14 days	43% among survivors at 23–186 days
Gordon (2003)*	Case series	100	Malawi	Adults with HIV	14 days	43% among survivors at 23–186 days
Gordon (2008)	Review	N/A	UK	Adults with immunocompromise	N/A	N/A
Guery (2018)	Case series	15	France	Adults with sickle cell disease	N/A	N/A
Habib (2005)	Case series	26	Singapore	10 with malignancy 7 with immunocompromise	N/A	N/A
Kankwatira (2004)	Review	N/A	N/A	Adults with HIV in Africa	10 days	N/A
Lim (2001)	Case series	50	Singapore	Adults with lupus	21 days	12%
McCollister (2016)	Case report	1	USA	Adult with immunocompromise	42 days	100%

* Duplicate publication

In uncomplicated iNTS bacteremia without persistently positive blood cultures or a risk of an endovascular focus of infection, the Infectious Diseases Society of America (IDSA) guidelines for the prevention of HIV-associated opportunistic infection recommend treating patients with CD4 counts >200/mm³ for 10 to 14 days (18). The initial treatment of choice is a fluoroquinolone, with acceptable alternatives being trimethoprim-sulfamethoxazole, ceftriaxone, or cefotaxime (18). Unfortunately, both antimicrobial susceptibility and genome sequencing revealed that none of the guideline options would have been appropriate empiric therapy for our patient. Frequent monitoring for recrudescence is indicated in the HIV-positive subpopulation, as this can present as a myriad of clinical syndromes including bacteremia or infections of the skin and soft tissue, urinary tract, bone, lung, and meninges. Our patient’s case highlights the need to also monitor closely for recrudescence in HIV-negative patients with extensively resistant iNTS, as well as to tailor antimicrobial therapy to the resistance patterns at the country of origin for an imported infection.

Conclusion

We describe a case of recurrent multidrug-resistant S. enterica serovar Typhimurium bacteremia, initially acquired from travel in a patient with no obvious risk factors. Genome sequencing revealed an extensively resistant strain of S. enterica serovar Typhimurium harbouring resistance to more than six antimicrobial classes through mechanisms including class A, C, and D ESBLs. Our patient’s case highlights the need to consider carbapenems as empiric therapy for iNTS bacteremia acquired outside Canada, raises the prospect that patients with extensively resistant iNTS bacteremia that lack risk factors should be monitored closely for recrudescence, and outlines that further study should be taken to determine the optimal duration of therapy.

Funding Statement

Genomic analysis was funded by the Canadian Institutes of Health Research (PJT-156214 to AGM).

Funding:

Genomic analysis was funded by the Canadian Institutes of Health Research (PJT-156214 to AGM). Computer resources were supplied by the McMaster Service Lab and Repository computing cluster, funded in part by grants to AGM from the Canadian Foundation for Innovation (34531). The authors declare no additional funding from a specific grant for this research from any funding agency in the public, commercial, or not-for-profit sectors.

Disclosures:

The authors have nothing to disclose.

Consent:

The authors obtained patient consent for the publication of this article.

Peer Review:

This manuscript has been peer reviewed.