Fatal Lactobacillus endocarditis in a patient with transcatheter aortic valve replacement

Abstract

Lactobacillus species may translocate from the gastrointestinal tract into systemic circulation from ingested probiotics or commensal flora. Their pathogenic potential is still debated. Lactobacillus endocarditis is a rare entity with only around 120 cases reported in the literature. Here, we report the first case of fatal Lactobacillus endocarditis with involvement of a transcatheter aortic valve replacement with the following goals: to reaffirm the pathological significance of Lactobacillus spp, to demonstrate the potential limitations of the modified Duke criteria in diagnosing infective endocarditis of transcatheter aortic valve replacement, and to urge clinicians to aggressively search for and consider empiric treatment for endocarditis in patients with prosthetic valves who develop Lactobacillus bacteraemia.

Background

Lactobacilli are Gram-positive, microaerophilic or facultatively anaerobic, non-spore forming rods that are part of the normal flora of the gastrointestinal tract and the vagina.¹ Early case series defined Lactobacillus endocarditis as a clinical entity, often occurring in patients with structural heart disease.^2–4 Risk factors for Lactobacillus bacteraemia include diabetes, immunosuppression, central venous devices, recent antibiotic treatment and insults to the gastrointestinal mucosa.^5–9 Despite their rarity, Lactobacillus infections have received growing interest due to increased probiotic use, a growing immunocompromised population and the widespread use of screening colonoscopy.

Transcatheter aortic valve replacement (TAVR) is an increasingly used procedure to replace a diseased aortic valve. Infective endocarditis of TAVR (TAVR-IE) has an incidence of 0.2%–3.1% at 1-year postimplant, and carries an in-hospital mortality rate of over 30%.¹⁰ As the indication for TAVR continues to expand into lower risk profile patients, the incidence of TAVR-IE is expected to increase.

Herein, we detail what is to our knowledge the first case report of fatal Lactobacillus TAVR-IE with the following goals: to reaffirm the pathological significance of Lactobacillus spp, to demonstrate the potential limitations of the modified Duke criteria in diagnosing TAVR-IE, and to urge clinicians to aggressively search for and consider empiric treatment for endocarditis in patients with prosthetic valves who develop Lactobacillus bacteraemia.

Case presentation

An 83-year-old man with medical history including hepatitis C cirrhosis, IgG monoclonal gammopathy, chronic pancytopenia, Crohn’s disease with ileal resection, prostate cancer in remission, and severe aortic stenosis treated with TAVR 26 months prior was admitted to the hospital in August after surveillance blood cultures returned positive for Lactobacillus species.

The patient had undergone surveillance upper endoscopy with oesophageal variceal banding in April and had a routine dental cleaning in May for which he took periprocedural amoxicillin. In June, the patient had been admitted for 2 weeks of cough and 1 day of fever, chills, rigours and fatigue (figure 1, June admission). Blood cultures from three different venipuncture sites grew Lactobacillus species. MRI of the spine revealed L3–L4 intervertebral discitis and two micro-abscesses in the left psoas muscle. Chest X-ray, CT scan of the abdomen and pelvis, and urinalysis did not demonstrate any findings consistent with infection. Transoesophageal echocardiogram (TEE) showed no evidence of infective endocarditis.

Figure 1.

Timeline showing blood culture results (red), antibiotic courses (blue) and significant diagnostic studies (green). Each date is shifted by the same random number between 1 and 364. (+) refers to blood cultures from at least two venipuncture sites showing growth of Lactobacillus spp, while (−) refers to no bacterial growth after 120 hours. A/P, abdomen/pelvis; AMP, ampicillin; CLI, clindamycin; CTX, ceftriaxone; GEN, gentamicin; TEE, transoesophageal echocardiogram; TTE, transthoracic echocardiogram; VAN, vancomycin.

The patient’s symptoms resolved within 48 hours with empiric vancomycin and ceftriaxone which were subsequently narrowed to ampicillin based on susceptibility studies (table 1, June admission). The patient’s blood cultures became persistently negative after 6 days. He was discharged with a diagnosis of Lactobacillus bacteraemia without a clear focus of infection and prescribed a 42-day course of ampicillin. Blood cultures drawn on 6 August (12 days after completing the ampicillin course) were negative. Repeat surveillance cultures drawn on 12 August were once again positive for Lactobacillus species, thus prompting the current admission (figure 1, August admission).

Table 1.

Susceptibilities of Lactobacillus spp grown during each of the two hospital admissions

	June admission Minimum Inhibitory Concentration (MIC) (µg/mL)	August admission MIC (µg/mL)
Ampicillin/sulbactam	1	1
Clindamycin	0.75	1
Metronidazole	256	≥256
Daptomycin	–	2
Penicillin	–	0.5

Each date is shifted by the same random number between 1 and 364.

During the current admission, the patient reported mild fatigue, but was otherwise asymptomatic. He denied abdominal pain and respiratory symptoms. Vital signs were within normal limits. Physical examination demonstrated a chronically ill-appearing male, awake and conversant, with a stable 2/6 systolic murmur. The abdomen was soft, non-tender and mildly distended. Neurological examination was normal. No stigmata of infective endocarditis were present.

Investigations

Admission labs showed leucopenia (2.6 K/UL; reference range 3.5–10.5 K/UL), thrombocytopaenia (105 K/UL; reference range 140–390 K/UL) and anaemia (10.4 g/dL; reference range 13.0–17.5 g/dL). His blood count differential revealed mild neutropaenia (absolute neutrophil count=1.5 K/UL; reference range 1.5–8.0 K/UL). Chart review revealed pancytopenia dating back two decades, and similar laboratory values months before the June admission. Erythrocyte sedimentation rate was 36 mm/hour (reference range 3–10 mm/hour) and C reactive protein was 1.1 mg/dL (reference range 0.0–0.5 mg/dL). Alanine aminotransferase was 29 unit/L (reference range 0–52 unit/L) and aspartate aminotransferase was 32 unit/L (reference range 0–39 unit/L).

ECG showed sinus rhythm with a left bundle branch block, which had been present since the patient’s TAVR. The PR interval was within normal limits.

CT scan of the chest, abdomen and pelvis demonstrated a moderate-sized splenic infarction and a small wedge infarction in the right kidney (figure 2). MRI of the spine was stable compared with the MRI done in June. Paracentesis showed no evidence of spontaneous bacterial peritonitis.

Figure 2.

CT scan of the abdomen and pelvis showing a moderate-sized splenic infarction (left) and a small renal infarction (right) marked by red arrows.

Repeat TEE revealed a highly mobile, linear echodensity superior to the leaflets of the TAVR prosthesis with involvement of the TAVR cage (figure 3). There was no evidence of paravalvular abscess or valve dysfunction. Blood cultures grew Lactobacillus spp, presumed to be recurrence of the patient’s previous infection. A diagnosis of Lactobacillus TAVR-IE was made.

Figure 3.

Transoesophageal echocardiogram (mid-oesophageal aortic valve long-axis view) showing a vegetation with involvement of the TAVR prosthesis. TAVR, transcatheter aortic valve replacement.

Treatment

Due to the patient’s significant comorbidities, our institution’s cardiac surgeons deemed the patient a high-risk candidate for surgical intervention, and a plan for prolonged antibiotic therapy was pursued. Intravenous ampicillin was reinitiated, however given the prior recurrence of bacteraemia on monotherapy despite initial blood culture clearance, gentamicin was added (figure 1, August admission).

One day after starting gentamicin, the patient began to experience episodes of lightheadedness and dizziness. Due to concern for possible ototoxicity, gentamicin was discontinued and replaced with oral clindamycin, per susceptibility studies (table 1, August admission). A final antibiotic regimen of intravenous ampicillin 2 g every 4 hours and oral clindamycin 600 mg three times a day for 6 weeks was established with a plan for indefinite amoxicillin prophylaxis following the completion of this course.

Outcome and follow-up

On the morning of planned discharge, the patient became acutely encephalopathic. In addition to altered mental status, neurological examination was significant for right–left confusion, acalculia and finger agnosia.

CT of the brain revealed a left parietal intraparenchymal haematoma (figure 4). Cerebral angiogram identified a ruptured mycotic pseudoaneurysm involving a distal parietal branch of the left middle cerebral artery. The patient underwent technically successful external ventricular drain placement and surgical resection of the mycotic pseudoaneurysm through a left parietal craniotomy. Over the subsequent 3 weeks, the patient developed multiple ventilator-associated pneumonias, small bowel obstruction and required multiple large-volume paracenteses for decompensated cirrhosis. Unfortunately, he developed circulatory shock refractory to vasopressors and worsening respiratory failure. His family subsequently elected a transition to comfort care. The patient ultimately passed away with his family at his bedside.

Figure 4.

CT of the brain (top) obtained shortly after our patient became acutely encephalopathic showed a left parietal intraparenchymal haematoma. Angiogram of left internal carotid artery (bottom) revealed a ruptured bilobed mycotic pseudoaneurysm in a distal branch (red arrows).

Discussion

Lactobacillus spp remains an extremely rare cause of infective endocarditis and bacteraemia, thought to represent 0.05%–0.4% of cases.^{8 11} While there have been around 120 cases of Lactobacillus endocarditis described in the literature, to the best of our knowledge, this case is the first reported fatal Lactobacillus endocarditis with involvement of a TAVR prosthesis and the second case overall.¹²

Given its increasing prevalence, probiotic consumption as an aetiology for Lactobacillus bacteraemia and endocarditis is a topic of interest. Our patient endorsed consumption of yoghurt products containing probiotic strains. In a review of 11 cases of Lactobacillus endocarditis that were ‘apparently linked to previous use of probiotics,’ nine cases involved patients with an identified alteration in gut mucosal integrity, recent endoscopy or tooth extraction.¹³ Alternatively, an observational study of 400 inpatients with a high prevalence of immunosuppression, impaired intestinal integrity and abnormal heart valves who were prescribed Lactobacillus-based probiotics found no cases of Lactobacillus bacteraemia that were identical to probiotic strains.¹⁴ Though these findings are reassuring, no universally accepted guidelines have been established for the administration of probiotics in these high-risk patients and thus caution is advised.¹⁵

Commensal gut Lactobacilli may also serve as a source of Lactobacillus bacteraemia for patients with predisposing factors. In an observational study of 193 patients undergoing upper endoscopy, 16 developed bacteraemia after the procedure with three developing Lactobacillus bacteraemia.¹⁶ In addition, a case of Lactobacillus endocarditis following colonoscopy has been reported.¹⁷ Hence, the aetiology of our patient’s Lactobacillus bacteraemia can be conceptualised as follows: alterations in gut mucosal integrity (cirrhosis with oesophageal varices and recurrent ascites, history of Crohn’s disease) and recent upper endoscopy likely facilitated exogenous (probiotic-containing yoghurt) or endogenous (commensal flora) Lactobacillus spp translocation from the gastrointestinal tract into systemic circulation. Underlying pancytopenia and the presence of a prosthetic valve increased the risk for this bacteraemia to evolve into endocarditis.

Endocarditis was only definitively diagnosed in our patient during the August admission when embolic complications (figure 2) and a vegetation (figure 3) were found. Neither were present during the June admission. Applying the modified Duke criteria to the June admission, our patient had one major criterion (persistently positive blood cultures) and two minor criteria (TAVR, fever) categorising him as possible, but not definite, endocarditis. During the August admission, he met two major criteria (persistently positive blood cultures, echocardiographic evidence of vegetation) and two minor criteria (TAVR, vascular phenomena), establishing the diagnosis of definite prosthetic valve endocarditis (PVE).

Several case reports have described similar clinical time courses with early findings consistent with possible endocarditis and negative TEE, and a repeat TEE 1–3 months after presentation revealing a diagnosis-clinching vegetation.^18–22 In particular, three of these cases occurred in patients in whom the vegetation grew on a prosthetic component.^18–20 Indeed, the modified Duke criteria have been shown to have lower sensitivity and specificity for PVE compared with native valve endocarditis, partly due to more frequently equivocal echocardiographic findings.^{23 24} An even greater diagnostic challenge is present with TAVR-IE. In a recent review of 593 patients with TAVR-IE, Harding et al found that the combined sensitivity of TTE and TEE was 68% compared with 86% with conventional PVE.¹⁰ As Puls et al describe, ‘abscess,’ ‘new partial dehiscence of prosthetic valve,’ and ‘new valvular regurgitation’ are particularly difficult echocardiographic criteria to satisfy in TAVR-IE, making it difficult to achieve definite diagnosis by modified Duke criteria.²⁵

These limitations may be magnified in endocarditis due to Lactobacillus, which may progress on a subacute timescale.^{19 26} Our patient’s initially negative TEE could be a manifestation of shadowing artefact paired with an extremely small vegetation. Alternative imaging modalities such as ¹⁸F-FDG PET (18-fluorodeoxyglucose positron emission tomography)/CT, SPECT (single-photon emission computerized tomography)/CT and cardiac CT, as outlined by recent European Society of Cardiology guidelines, may be particularly valuable for these patients.²⁷ Clinicians should also consider treating all Lactobacillus bacteraemia as endocarditis in patients with prosthetic valves, especially TAVR, even if definite diagnosis cannot be achieved. Early aggressive medical management is especially prudent in this population because the cohort of patients who are eventually diagnosed with Lactobacillus TAVR-IE is one of extremely high surgical risk, owing to the risk factors that contributed to the development of Lactobacillus bacteraemia and the decision to undergo TAVR.

There is currently no guideline-directed antibiotic regimen for Lactobacillus endocarditis. While the diagnosis of TAVR-IE was not made during our patient’s June admission, he was treated with a 6-week course of ampicillin for Lactobacillus bacteraemia, which is an acceptable antibiotic regimen for PVE due to some other pathogens such as viridans group Streptococci, Streptococcus gallolyticus (bovis) and Enterococcus species.²⁸ Indeed, our patient’s blood cultures cleared on this regimen. Despite this, we hypothesise that endocarditis was present during the June admission and inadequately treated with ampicillin monotherapy as evidenced by the recurrent growth of Lactobacillus spp during the August admission, although we were unable to obtain confirmation at the species level.

In accordance with this, an early case series of Lactobacillus endocarditis showed that inadequate treatment is more common with penicillin monotherapy compared with combination therapy.⁴ A common combination regimen includes a penicillin combined with an aminoglycoside, which is based on early in vitro studies showing synergism towards Lactobacilli.^{2 4} Because our patient demonstrated signs of gentamicin toxicity, an alternative combination of ampicillin and clindamycin was considered appropriate as well. Ultimately, early initiation of susceptibility-guided combination antibiotic therapy is crucial for successful medical management.²⁹

In summary, we present the first case report of fatal Lactobacillus endocarditis with involvement of a TAVR prosthesis. This case highlights the limitations of the modified Duke criteria when applied to TAVR-IE of insidious onset, such as with Lactobacillus. As the number of TAVR procedures increases, clinicians must be aware of the pathological significance of Lactobacillus bacteraemia and aggressively search for endocarditis. Clinicians should appreciate the limited diagnostic value of TEE in this setting and consider using alternative imaging modalities to aid in diagnosis. Treatment for endocarditis with susceptibility-guided combination antibiotic therapy should also be considered even if a definite diagnosis cannot be made.

Learning points.

• Alterations in gut mucosal integrity, recent endoscopy and recent dental extraction, particularly in the setting of immunocompromise, may facilitate bacterial translocation of either probiotic Lactobacilli strains or endogenous commensal Lactobacilli strains into systemic circulation.

• Lactobacillus bacteraemia should be considered a true pathological entity and workup for a localising infection should be initiated.

• The modified Duke criteria have limited sensitivity in diagnosing definite infective endocarditis of transcatheter aortic valve replacement (TAVR), particularly with infections of insidious onset such as Lactobacillus. Clinicians should consider using non-echocardiographic imaging modalities to aid in diagnosis and lower their threshold to treat for endocarditis in patients with TAVR who present with Lactobacillus bacteraemia.