Pyogenic vertebral osteomyelitis and iliopsoas muscle abscess caused by non-typhoidal Salmonella rapidly identified by genetic testing in an immunocompetent teenage boy without hemoglobinopathies

Abstract

Vertebral osteomyelitis caused by non-typhoidal Salmonella spp. is exceedingly rare, particularly among immunocompetent children. This report presents an unusual case of lumbar osteomyelitis and an iliopsoas muscle abscess caused by non-typhoidal Salmonella in an immunocompetent pediatric patient with a multidrug allergy. A 13-year-old boy presented with fever and lumbar pain. Diagnostic imaging revealed lumbar osteomyelitis and an iliopsoas abscess. Blood culture and initial iliopsoas puncture tissue sample test results were negative. Therefore, cefazolin was administered as empirical therapy for covering typical organisms such as Staphylococcus aureus and Streptococcus species causing vertebral osteomyelitis in healthy children. However, genetic testing of the biopsy sample of the vertebral tissue subsequently identified Salmonella spp. as the causative agent. Culture of the vertebral tissue also yielded Salmonella spp., with the O-antigen identified as type 4. Antibiotic selection was challenging because of the patient's drug allergies and age. Treatment was commenced with ceftriaxone and later changed to ampicillin owing to adverse drug reactions. The side effects, such as rash, fever, and nausea, persisted after switching to oral sulfamethoxazole-trimethoprim, which was later changed to amoxicillin. Although the treatment duration of vertebral Salmonella osteomyelitis in children is not standardized, we treated the patient for 9 weeks based on previously reported evidence. Rapid identification of the causative organism is important because vertebral osteomyelitis requires long-term treatment and treatment options may be limited, particularly in pediatric patients. Physicians should consider genetic testing to identify the causative organism of osteomyelitis.

Introduction

Salmonella is a gram-negative, rod-shaped bacterium belonging to the Enterobacteriaceae family [1]. The major species within the Salmonella genus are Salmonella enterica and S. bongori, with S. enterica further subdivided into six subspecies [1]. Salmonella spp. are classified into different serovars based on the presence of specific antigens such as O (somatic) and H (flagellar) [1]. Each serovar exhibits unique characteristics, including different host range and pathogenicity. For example, S. typhi is associated with typhoid fever, whereas S. enteritidis is commonly associated with foodborne gastroenteritis [1].

Non-typhoidal Salmonella (NTS) mainly causes gastroenteritis and may also cause severe invasive infections (extraintestinal infections), predominantly among children with hemoglobinopathies and immunocompromised individuals with conditions such as malignancies, human immunodeficiency virus infection, and sickle cell disease [2], [3], [4]. However, vertebral osteomyelitis due to Salmonella is very rare. There are no specific symptoms characteristic of Salmonella vertebral osteomyelitis, and as with vertebral osteomyelitis caused by other bacteria, fever and back pain are the main symptoms [5]. Furthermore, Salmonella osteomyelitis is reported in only 0.45 % of all osteomyelitis cases and comprises 0.7–0.8 % of all Salmonella infections [6], [7].

Regarding antibiotic selection for Salmonella infection, fluoroquinolones are a reasonable empiric choice for treating osteomyelitis and bacteremia caused by Salmonella spp [8]. However, quinolone use in children is generally limited owing to potential safety concerns, such as musculoskeletal effects (tendinitis and tendon rupture) and cartilage and joint developmental issues [9], [10]. This means that antimicrobial options for Salmonella infections in children are limited.

We report a case of lumbar osteomyelitis and iliolumbar muscle abscess in an immunocompetent pediatric patient, caused by S. enterica subspecies enterica (NTS). The patient experienced adverse reactions to multiple antimicrobials, including third-generation cephalosporins and trimethoprim-sulfamethoxazole (SXT). The presence of Marfan syndrome precluded quinolone use, further complicating the choice of therapeutic antimicrobials.

This case underscores the necessity of considering NTS alongside Staphylococcus aureus as potential causative agents in pediatric vertebral osteomyelitis despite its rarity in children. It also details treatment duration (both intravenous and oral antibiotic administration) for Salmonella vertebral osteomyelitis in this demographic. Furthermore, the utility of genetic tests in identifying osteomyelitis pathogens is described.

Case

A 13-year-old Japanese boy presented at our hospital with fever, lumbar pain, and weight loss. Two months prior to admission, he experienced intermittent back pain, which worsened subsequently. Additionally, he developed intermittent fever and chills three days before the visit. Over the preceding two months, he lost 5 kg (11 lbs). The patient wore a brace at bedtime due to thoracic scoliosis caused by Marfan syndrome, although he experienced no pain from the scoliosis. He had no medical history of immunodeficiency disorders, such as chronic granulomatous disease or sickle cell disease. He had received the BCG vaccine for tuberculosis when he was 6 months old. Furthermore, he did not have a family history of tuberculosis, immunodeficiency, or hematological disorders.

Upon arrival, the patient had difficulty getting out of bed and walking owing to back pain. His height, weight, and body mass index were 173.2 cm, 46.0 kg (compared with from 51.0 kg 2 months previously), and 15.4 kg/m², respectively. On admission, he was alert, and his vital signs were as follows: blood pressure, 115/66 mmHg; heart rate, 110 bpm; body temperature, 39.2°C; and respiratory rate, 16 breaths per minute. Physical examination revealed no superficial lymph nodes. Cardiorespiratory, abdominal, limb, and skin examination results were normal. Mild knocking pain in the lumbar spine was noted. Although laboratory tests indicated a normal white blood cell count of 8200/μL (normal range: 3300–8600), neutrophils count was high at 81.4 %, C-reactive protein level was slightly elevated at 2.9 mg/dL (normal range: ≤0.04), and erythrocyte sedimentation rate was 26 mm/h (normal range: 0–10 mm/h). The HIV screening antibody test was negative, and serum immunoglobulin levels were within normal limits, with no abnormalities observed in serum complement levels or lymphocyte subsets. Magnetic resonance imaging (MRI) using short-tau inversion recovery (STIR) revealed a high signal intensity at the upper left margin of the second lumbar vertebra and the lower left margin of the first lumbar vertebra (Fig. 1A). Additionally, an MRI using STIR revealed a high-signal intensity area in the left iliopsoas muscle (Fig. 1B). Lumbar vertebral osteomyelitis and an abscess in the left iliopsoas muscle were suspected.

Fig. 1.

Magnetic resonance imaging findings of the lumbar spine and iliopsoas muscle. (A) Short-tau inversion recovery revealed a high signal at the upper left margin of the second lumbar vertebra and the lower left margin of the first lumbar vertebra. (B) Short-tau inversion recovery revealed a high-signal-intensity area in the left iliopsoas muscle.

Two sets of blood cultures were immediately obtained. Computed tomography-guided iliopsoas puncture was performed; however, no evidence of any organism on Gram staining of the tissue was obtained. The attending physician started 4.5 g/day intravenous cefazolin (CFZ) (100 mg/kg/day pediatric doses) as empirical therapy for targeting S. aureus and streptococci. However, the high fever and back pain persisted despite the use of antipyretic analgesics. The iliopsoas tissue culture and two sets of blood cultures were negative. Therefore, we performed a repeated biopsy of the vertebral tissue in addition to abscess drainage on the 7th day after hospitalization. Gram-stain of obtained vertebral tissue revealed the gram-negative rods (Fig. 2). Additionally, we used the blood culture identification panel 2 (BCID2) of the FilmArray® system (BioFire Diagnostics, Salt Lake City, UT, USA) to conduct genetic testing of obtained vertebral tissue according to methods utilized in previous studies to detect a causative pathogen [11]; BCID2 identified the causative pathogen as “Salmonella spp.” Culture of the vertebral tissue also yielded Salmonella spp., with the O-antigen identified as type 4. Identification of H antigens was not performed because our laboratory lacks the necessary tools. Therefore, we switched from CFZ to ceftriaxone (CRO) 2 g/day for definitive therapy. Additionally, a stool culture was performed, but Salmonella spp. was not detected. Subsequently, his back pain and fever gradually subsided. However, on the 10th day of hospitalization, the patient developed a fever over 40 °C, chills, fatigue, and a rash, thought to be caused by CRO. Antimicrobial susceptibility test results revealed that the NTS isolated from the vertebral tissue was susceptible to ampicillin (AMP), cefotaxime, ciprofloxacin, levofloxacin, and SXT, based on the Clinical and Laboratory Standards Institute M100-S26 (Table 1). Quinolone use in children is generally limited because of the risk of adverse musculoskeletal events such as tendinitis and tendon rupture and the risk of affecting cartilage and joint development. Therefore, we switched the treatment from CRO to AMP 8 g/day, which was continued until the 29th day of hospitalization. His fever subsided on the 19th day, and the back pain disappeared on day 25. In preparation for discharge, AMP, an intravenous antibacterial drug, was replaced with SXT (4 tablets/day), an oral antibacterial drug, on day 30. However, skin rash, fever over 38°C, and nausea appeared, which were thought to be caused by SXT. Therefore, SXT was replaced with amoxicillin (AMX) (1500 mg/day). Subsequently, the skin rash disappeared, and the patient became afebrile. The patient’s general condition improved, and he was discharged with AMX treatment on day 35 (Fig. 3). AMX 1500 mg/day was administered for 4 weeks. We had MRI confirmation that the patient's abscess was in remission, and the patient received antibiotics for a total of 9 weeks for osteomyelitis.

Fig. 2.

Gram staining of vertebral bone tissue. Black arrows indicate gram-negative rod.

Table 1.

Antimicrobial susceptibility of non-typhoidal Salmonella isolated in the present case.

Antibiotics	MIC (μg/mL)	Interpretation*
Ampicillin	≤ 8	S
Cefotaxime	≤ 1	S
Ciprofloxacin	≤ 0.063	S
Levofloxacin	≤ 0.125	S
Trimethoprim-sulfamethoxazole	≤ 38/2	S

Abbreviations: MIC, minimum inhibitory concentration; S, susceptible

^*Interpretation criteria are recommended by the Clinical and Laboratory Standards Institute M100-S23.

Fig. 3.

Clinical course of the present case.

Additional interviews revealed that the patient's family purchased eggs sold at the poultry farm once every 3 weeks. Therefore, it was suspected that the eggs consumed were contaminated with Salmonella, although none of the family members, including the patient himself, presented with diarrhea or other gastrointestinal symptoms in the 3 months before admission. This patient did not consume raw eggs. In addition, this patient had no history of contact with animals or pets (such as turtles) associated with Salmonella infections.

Discussion

In this case report, lumbar vertebral osteomyelitis and a psoas abscess in an immunocompetent teenager, caused by NTS, are described. Despite the initial sample and blood cultures yielding no microorganisms, genetic testing on the second sample swiftly identified the causative organism, enabling prompt targeted therapy. Treatment options were restricted due to adverse reactions to multiple antimicrobials and underlying Marfan syndrome.

Generally, NTS is an important bacterial causative agent of diarrheal diseases worldwide. This pathogen can lead to invasive diseases, such as bacteremia and meningitis; however, the occurrence of osteomyelitis is rare, manifesting in just 59 (0.76 %) of the 7779 patients with Salmonella infection in the United States [7]. The most common sites of osteomyelitis due to Salmonella spp. are long bones such as the humerus and femur [7]. Vertebral osteomyelitis is rare, accounting for 0.45 % of all cases of Salmonella osteomyelitis [12]. Moreover, vertebral Salmonella osteomyelitis is more commonly associated with the lumber spine than the cervical or thoracic spine [13]. Boguniewicz et al. conducted a comparison of musculoskeletal infections due to NTS and S. aureus in their case-control study of immunocompetent children aged 30 days to 18 years old and found that NTS is more implicated in spinal involvement than S. aureus (58.3 % vs. 28.3 %, odds ratio 5.32, 95 % CI:1.42–20.13) [14]. Therefore, although Salmonella vertebral osteomyelitis is a rare condition, clinicians need to consider the possibility of Salmonella as the causative organism when “lumbar” spinal osteomyelitis occurs in children because CFZ, which is used as empiric therapy for S. aureus in vertebral osteomyelitis, is ineffective against Salmonella.

Currently, there is no international consensus on the recommended duration of treatment for Salmonella vertebral osteomyelitis in pediatric patients [15]. A review of the available literature shows that several studies have made different recommendations. One review article on acute osteomyelitis in children described a minimum of 4–6 weeks [16]; however, the intravenous infusion and oral administration durations were not specified. One practice bone and joint infection guideline published in 2017 recommended a minimum treatment duration of 4–6 weeks, with 1–3 weeks of intravenous infusion and at least 3 weeks of oral administration. Wen et al. recommended 4–6 weeks of antibiotic treatment with at least 3 weeks of parenteral therapy [17]. Osteomyelitis caused by NTS in immunocompetent children is exceedingly rare, complicating traditional case-control study conduction. Therefore, using treatment durations from prior reports as guidelines is practical. Katzouraki et al.'s systematic review on NST vertebral osteomyelitis in immunocompetent children (spanning 20 articles from 1977 to 2020, covering 21 patients averaging 12.7 years [range, 2–18]) showed that a 9.6-week average antibiotic treatment led to positive outcomes in 20 of 21 (95.2 %) patients, predominantly affecting the lumbar L4/5 region [18]. Based on this previous investigation [18], our patient was treated with antimicrobials for 9 weeks. Because the number of studies reported to date is small, further evidence is required.

One of the problems in diagnosing osteomyelitis is that conventional culture methods require several days to culture the causative organisms, and even when blood and tissue cultures are combined, the positivity rates of cultures are relatively low, at approximately 40–70 %, according to previous bone and joint infection studies [2], [3]. Although expensive and not available in all facilities, genetic tests have recently become available for detecting causative pathogens for osteomyelitis [19]. Genetic testing is beneficial for selecting suitable antimicrobial agents and preventing unnecessary drug use. Targeted therapy, in this case, commenced prior to culture and antimicrobial susceptibility results. Hence, if available, aggressive use of genetic testing is advised in severely ill patients with worsening symptoms post-empirical treatment initiation or for early causative organism identification, as illustrated here.

Although Marfan syndrome is not a classic immunodeficiency, its skeletal and connective tissue abnormalities may predispose to vertebral infections. Thoracic scoliosis can alter spinal mechanics and impair local blood flow, potentially creating a niche for bacterial seeding during transient bacteremia. Salmonella is known to target areas of vascular or structural abnormality, such as atherosclerotic vessels and aneurysms, suggesting that abnormal bony anatomy may also be vulnerable to hematogenous infection [20], [21]. Additionally, Marfan-associated tissue fragility and delayed healing due to fibrillin-1 mutations may compromise local defense [22]. Spinal instability, which is more frequent in Marfan patients with scoliosis, may lead to micro-injuries or necrotic foci, further lowering resistance to infection [23]. Although direct cases are rare, analogous reports show that Marfan patients with dural ectasia or meningoceles have an increased risk of spinal infections [24], [25]. These examples support the plausibility that thoracic scoliosis in Marfan syndrome could act as a local risk factor for vertebral osteomyelitis, including from non-typhoidal Salmonella.

This case report has several limitations. First, we cannot explain why NTS vertebral osteomyelitis occurred in this case. Immunocompromised children and those with underlying conditions such as hemoglobinopathies are particularly vulnerable to NST-invasive diseases, including osteomyelitis; however, the present case did not have those factors. Second, the route of Salmonella transmission in this case remains unclear. While we presumed that the eggs consumed by the patient may have been the source of infection, the estimated prevalence of Salmonella in commercially available eggs in Japan is extremely low, at 0.003 % [26]. Third, the patient was treated with antimicrobials for 9 weeks based on a previously reported investigation [18]; however, the appropriate duration of treatment was unknown. Osteomyelitis due to NST is rare, and the number of studies reported to date is small; thus, further evidence is warranted.

In conclusion, our case of pyogenic spondylitis in an immunocompetent child, caused by Salmonella with a left psoas muscle abscess, presumably resulted from contaminated chicken egg consumption. The patient's allergies to third-generation cephalosporins and SXT drugs, coupled with the infeasibility of using quinolones due to Marfan syndrome, complicated antimicrobial agent selection.

The key takeaway from this case is that pediatric NST osteomyelitis is rare in immunocompetent children (<1 %), but Salmonella more commonly causes pyogenic spondylitis in the lumbar spine than S. aureus. Thus, aggressively identifying the pathogen in lumbar spondylitis cases in children is crucial, as therapeutic antimicrobial agents are limited, and Salmonella has a natural resistance to CFZ, the empirical therapy for lumbar osteomyelitis.

Author’s contributions

All authors made a significant contribution to the work reported, whether in the conception, study design, execution, acquisition of data, analysis, and interpretation, or all these areas; took part in drafting, revising, or critically reviewing the article; gave final approval of the version to be published; agreed on the journal to which the article has been submitted; and agreed to be accountable for all aspects of the work.

CRediT authorship contribution statement

Hirai Jun: Writing – review & editing, Writing – original draft, Investigation, Conceptualization. Mori Nobuaki: Writing – review & editing, Conceptualization. Sakanashi Daisuke: Writing – review & editing. Shibata Yuichi: Visualization. Asai Nobuhiro: Data curation. Hagihara Mao: Writing – review & editing. Mikamo Hiroshige: Writing – review & editing, Supervision.

Ethics approval and informed consent

Written informed consent was obtained from the patient’s mother to publish this case report and the accompanying images. The present case did not require ethics committee approval based on the Japanese Ethical Guidelines for Clinical Research to publish the case details.

Funding

The authors and co-workers did not receive any specific funding.

Declaration of Competing Interest

None

Data availability

The data is available upon reasonal request.