Introduction

Bacterial co-infection in patients with coronavirus disease 2019 (COVID-19) can complicate diagnosis and management because of overlapping clinical manifestations, laboratory abnormalities, and chest imaging findings among pathogens [1]. Although pneumonia associated with COVID-19 and multiple bacterial pathogens, including Legionella, has rarely been reported, such cases may follow a severe clinical course [2, 3]. Cases with more than one pathogen may have an atypical presentation, and failure to recognize a treatable bacterial infection may delay appropriate therapy.

We report a case of suspected triple co-infection with severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), Legionella pneumophila (L. pneumophila), and Streptococcus pneumoniae (S. pneumoniae) that developed after domestic travel to a hot spring in Japan in a patient with a history of splenectomy. This case underscores the importance of considering L. pneumophila involvement and other treatable bacterial co-infections when COVID-19 pneumonia is accompanied by recent hot-spring exposure, host risk factors, and atypical imaging findings. Careful exposure history taking and appropriate pathogen-directed testing are essential for identifying treatable co-infections, guiding timely empirical therapy, and facilitating subsequent antimicrobial adjustment or de-escalation [4, 5].

Timeline

The patient was in his usual state of health before a 2-day hot-spring trip. After returning home, his wife developed fever first, followed by the patient the next day. Both were diagnosed with COVID-19 at a local clinic, and the patient was transported to our hospital after a worsening productive cough and altered sensorium. Subsequent hospital course, antimicrobial treatments, and laboratory results are summarized in Figure 1.

Display full size

Figure 1. Clinical course and chest radiographs of a patient with pneumonia. (A) Clinical course showing changes in peripheral oxygen saturation, white blood cell count, and C-reactive protein level after initiation of remdesivir, ceftriaxone, and levofloxacin. (B) Serial chest radiographs showing gradual improvement of pulmonary infiltrates. CRP: C-reactive protein; CTRX: ceftriaxone; LVFX: levofloxacin; SpO₂: peripheral oxygen saturation; WBC: white blood cell count.

Narrative

A man in his 70s was admitted to our hospital with fever, worsening productive cough, and altered sensorium. His medical history was notable for esophageal cancer treated with subtotal esophagectomy, followed by reconstruction and concomitant splenectomy 10 years prior to presentation. He had no history of smoking, consumed alcohol daily, and had no known drug allergies. He was not currently receiving corticosteroids, immunosuppressive agents, or anticancer therapy. He was independent in activities of daily living, had previously received pneumococcal vaccination, but had not been vaccinated against COVID-19.

Several days before admission, he traveled to a hot spring for 2 days with a group of nine people, including his wife. After returning home, his wife developed fever, and he developed fever the following day. After being diagnosed with COVID-19 at a local clinic, both were managed conservatively with antipyretics. His productive cough subsequently worsened, and he developed altered sensorium that warranted emergency transport by ambulance to our hospital.

On admission, his body temperature was 38.3°C, blood pressure was 130/60 mmHg, pulse rate was 72 beats/min, respiratory rate was 18 breaths/min, and oxygen saturation was 90% on room air. The patient was mildly confused, with a Glasgow Coma Scale score of 14 (E4V4M6). Physical examination revealed coarse crackles in the right lung. He had cough, sputum production, anorexia, and malaise, but no headache, chills, sore throat, abdominal pain, diarrhea, or myalgia.

Diagnostics

Laboratory diagnosis

Laboratory tests showed leukocytosis (white blood cell count, 13,900/μL) with neutrophil predominance, an elevated C-reactive protein level (23.6 mg/dL), hyponatremia (133 mEq/L), and hypoalbuminemia (2.5 g/dL). Total bilirubin and liver enzyme levels were within normal limits, as was renal function. Urinalysis and urine culture were unremarkable. Urinary antigen tests for L. pneumophila serogroup 1 and S. pneumoniae were positive. These tests were performed using the ImmunoCatch^® Legionella (product code E-ET02) and ImmunoCatch^® Pneumococcus kits (product code E-ET04), respectively (Eiken Chemical Co., Ltd., Tokyo, Japan), both of which are immunochromatographic assays.

SARS-CoV-2 infection was confirmed by nucleic acid amplification testing on admission. Additional in-house real-time reverse transcription polymerase chain reaction testing performed on hospital day 3 using a QuantStudio 5 system (model A28134, 96-well, 0.2 mL; Thermo Fisher Scientific) with TaqMan™ Fast Virus 1-Step Master Mix (product No. 4444434; Thermo Fisher Scientific), Primer/Probe N2 (2019-nCoV) (product No. XD0008; Takara Bio Inc.), and a custom S-gene primer/probe set (Thermo Fisher Scientific) yielded cycle threshold values of 21.52 for the N2 gene and 22.71 for the S gene, consistent with a high SARS-CoV-2 RNA burden. Mutation screening using TaqPath™ 1-Step RT-qPCR Master Mix, CG (product No. A15299; Thermo Fisher Scientific) and the TaqMan™ SARS-CoV-2 Mutation Panel (product No. A49785; Thermo Fisher Scientific) on the same platform detected the G339D and L452R mutations. According to the institutional testing algorithm in use at that time, the G339D-positive pattern was interpreted as being consistent with an Omicron variant.

Gram staining of purulent sputum showed abundant Gram-positive cocci (4+), and sputum culture yielded S. pneumoniae, including both mucoid and non-mucoid morphotypes (2+ each). Species identification and antimicrobial susceptibility testing of the sputum isolate were performed using the RAISUS ANY automated microbiology system (model RAISUS ANY; Nissui Pharmaceutical Co., Ltd., Tokyo, Japan) with an RSCP2 identification and susceptibility plate (product No. 04135; Nissui Pharmaceutical Co., Ltd., Tokyo, Japan). Blood cultures were negative.

Chest computed tomography (CT) showed multifocal bilateral ground-glass opacities with a nonsegmental distribution, predominantly in the peripheral and subpleural lungs (Figure 2). Superimposed on these opacities, a relatively well-defined area of consolidation with air bronchograms was observed in the right middle lobe in a peribronchovascular distribution with an adjacent faint reticular opacity. Furthermore, trace bilateral pleural effusion was observed. No significant mediastinal or hilar lymphadenopathy was observed.

Display full size

Figure 2. Chest computed tomography on admission. (A–C) Multifocal bilateral ground-glass opacities are distributed predominantly in the peripheral and subpleural lung fields. (D, E) A relatively well-defined consolidation with air bronchograms is present in the right middle lobe in a peribronchovascular distribution, accompanied by adjacent faint ground-glass and reticular opacities. (F) A small subpleural band-like opacity is present in the left lower lobe, with trace bilateral pleural effusions.

Taken together, these microbiological and radiological findings supported the diagnosis of suspected triple co-infection involving SARS-CoV-2, L. pneumophila, and S. pneumoniae.

Patient perspective

At an outpatient follow-up visit after recovery, the patient was informed of the clinical course and favorable outcome. He understood that sharing the details of his case could be educational for clinicians, and provided written informed consent for publication.

Discussion

This case highlights the difficulty in diagnosing pneumonia in patients with COVID-19 when more than one pathogen is involved. Although bacterial co-infection at the time of COVID-19 presentation is relatively uncommon, treatable bacterial pneumonia may be overlooked if all clinicoradiological findings are attributed solely to SARS-CoV-2 [1]. In this case, neutrophilic leukocytosis, productive cough, focal consolidation, marked inflammation, positive pneumococcal and L. pneumophila urinary antigen tests, growth of S. pneumoniae in sputum culture, and recent hot-spring exposure supported a suspected triple co-infection rather than COVID-19 alone.

Several clinical features further supported suspected L. pneumophila infection. Current guidance recommends testing for Legionnaires’ disease in patients with pneumonia who have severe disease or overnight travel during the incubation period; moreover, the recommendations include combining urinary antigen testing with lower respiratory tract culture or molecular testing when Legionnaires’ disease is suspected [6, 7]. Hot-spring exposure is an important epidemiological clue because hot-spring bath water has been molecularly confirmed as the source of infection in a sporadic case of L. pneumophila pneumonia in Japan [8]. In addition, a relatively well-defined peribronchovascular consolidation with ground-glass opacities constitutes a characteristic CT pattern that is more frequent in L. pneumophila pneumonia than in pneumococcal pneumonia [9]. Urinary antigen testing for Legionella has high specificity; the reported pooled sensitivity and specificity were 79% and 100% overall and 86% and 100% for L. pneumophila serogroup 1, respectively [10]. Although respiratory culture or molecular testing for Legionella was unavailable, the positive urinary antigen result, when interpreted together with the exposure history, clinical features, and CT pattern, supported concomitant legionellosis. Persistent antigen excretion after a previous infection was unlikely because the patient had no known history of prior legionellosis [11]. Incidental false-positive results and pseudoepidemics have been reported [12, 13]. Pontiac fever was also considered but was less consistent with this patient’s radiographic pneumonia, hypoxemia, and focal consolidation [14].

The timing of symptom onset was compatible with the possible acquisition of all three pathogens around the time of hot-spring travel. The reported incubation periods are a mean of 2.6 days for SARS-CoV-2 Omicron infection, 2–14 days for Legionnaires’ disease, and 1–3 days for pneumococcal pneumonia; the low cycle threshold values on hospital day 3 were consistent with a high viral RNA burden and recent active SARS-CoV-2 infection [7, 15, 16].

The clinical presentation may have reflected overlapping but largely independent risk factors: lack of COVID-19 vaccination, recent hot-spring exposure, and prior splenectomy. SARS-CoV-2-related airway epithelial injury and impaired mucociliary clearance may have compromised local airway defenses and facilitated bacterial co-infection [17]. Concomitant L. pneumophila and S. pneumoniae infections occur rarely and have mainly been described in case reports [18, 19]. Thus, this case is best interpreted as involving overlapping risks and possible sequential infections rather than a proven pathogen–pathogen interaction.

The history of splenectomy supports the clinical significance of pneumococcal infection but does not directly explain susceptibility to SARS-CoV-2 or Legionella. This is because asplenia mainly predisposes patients to infections that are caused by encapsulated bacteria, especially S. pneumoniae, through impaired splenic clearance and polysaccharide antigen responses [20]. However, prior pneumococcal vaccination does not eliminate this residual risk [21]. Unlike the fatal case reported by Fukuda et al., in which L. pneumophila and S. pneumoniae co-infection developed during dexamethasone and baricitinib therapy for COVID-19 [3], our patient had a post-splenectomy host risk rather than pharmacological immunosuppression, and improved with early pathogen-directed therapy.

This case underscores the need for early microbiological evaluation when bacterial co-infection is suspected in patients with COVID-19 pneumonia. In such cases, multiplex molecular panels may complement exposure-directed testing by facilitating the simultaneous detection of additional respiratory pathogens [3, 22].

Conclusions

We report a case of suspected triple co-infection involving SARS-CoV-2, L. pneumophila, and S. pneumoniae after hot-spring travel in a post-splenectomy patient. Clinicians should consider Legionella infection when COVID-19 pneumonia is accompanied by recent hot-spring or aerosolized water exposure, atypical imaging findings, or host risk factors, and should promptly evaluate treatable bacterial pathogens.