Corynebacterium diphtheriae-infective endocarditis in a patient with an atrial septal defect closure device

Abstract

An 18-year-old woman presented to our institution with fever, bilateral flank pain, headache and photophobia. She had a previous atrial septal defect (ASD) closure device inserted at the age of 9 years. Blood cultures on admission were positive for Corynebacterium diphtheriae, and transoesophageal echocardiogram (TOE) revealed an echodensity associated with the ASD closure device, most consistent with a vegetation. She was treated for infective endocarditis with 6 weeks of intravenous benzylpenicillin, and follow-up TOE showed resolution of the echodensity. To our knowledge, no cases of C. diphtheriae endocarditis of an ASD closure device have previously been reported.

Background

Non-toxigenic Corynebacterium diphtheriae is a rare cause of infective endocarditis, and is most commonly reported in those with unrepaired cyanotic heart disease, valvular disease, prosthetic valves, intravenous drug users or other predisposition to infective endocarditis.^{1 2} No cases to date have described infective endocarditis due to this organism associated with an atrial septal defect (ASD) closure device.

This case warrants attention because it brings together two uniquely rare conditions. We present the first case of non-toxigenic C. diphtheriae-infective endocarditis of an ASD closure device in an immunocompetent individual.

Case presentation

An 18 year-old, Australian born, Caucasian woman presented to our institution with acute onset of fever, severe bilateral flank pain, headache and photophobia. There were no respiratory symptoms including sore throat, and no urinary symptoms. Her past medical history was significant for a secundum ASD for which she underwent percutaneous insertion of a closure device at 9 years of age with no device-related complications. She also had a history of migraine, took no regular medications, and vaccinations were up to date in accordance with the Australian immunisation schedule.³ She was studying nursing in Australia, had no recent overseas travel and no history of intravenous drug use.

On examination, she was febrile at 38.8°C, tachycardic at 130 bpm with a blood pressure of 102/55 mm Hg. She had a witnessed episode of rigours and confusion in the emergency department. There was no rash and no peripheral stigmata of infective endocarditis. On auscultation, there was a grade II systolic murmur over the left upper sternal border, and bilateral vesicular breath sounds. There was bilateral flank tenderness on abdominal examination. Neurological examination was normal, and Kernig’s sign was negative. Examination of the oropharynx revealed prominent, slightly pitted, but non-inflamed tonsils, and nasoendoscopy demonstrated no other abnormalities of the upper respiratory tract.

Investigations

Blood tests revealed a white cell count of 10.7×10⁹/L (normal value 4.0–11.0), neutrophils of 10.1×10⁹/L (normal value 2.0–7.5), lymphocytes of 0.3×10⁹/L (normal value 1.5–4.0) and C-reactive protein was 71.9 mg/L (normal value 0.0–5.0). Haemoglobin, platelets, lipase, liver function tests, urea, electrolytes and creatinine, calcium, magnesium, phosphate were all within normal parameters.

ECG showed sinus tachycardia. Neck and chest radiographs, and renal tract ultrasound did not identify any abnormality. PCR for influenza viruses on nasopharyngeal swab was negative, and there was no growth on urine culture.

Cerebrospinal fluid had protein 0.37 g/L (normal value 0.15–0.45), glucose 4.3 mmol/L (normal value 2.7–4.2), leucocytes 9×10⁶/L and erythrocytes 15×10⁶/L and no growth after 48 hours. PCR of the cerebrospinal fluid was negative for herpes simplex virus 1 and 2, varicella zoster virus, enterovirus and parechovirus.

The first set of blood cultures grew Gram-positive bacilli in clusters after 24 hours in the anaerobic bottle, and after 48 hours in the aerobic bottle. The organism was later identified as C. diphtheriae using the VITEK 2 gram-positive identification card and confirmed with VITEK MS. Corynebacterium API identified the biotype as gravis. The toxin gene was not detected, and minimum inhibitory concentration (MIC) by Etest showed penicillin MIC 0.38 mg/L (Clinical and Laboratory Standards Institute breakpoints for Corynebacterium species, <0.12 s, 0.25–2 I, >4 R), ceftriaxone MIC 1.50 mg/L, vancomycin MIC 0.38 mg/L.⁴

Transthoracic echocardiogram showed an independently mobile echodensity on the left atrial surface of the ASD closure device (figure 1). There was normal appearance and function of all chambers and valves. Transoesophageal echocardiogram (TOE) confirmed the presence of a 3-mm linear mobile echodensity associated with the left atrial surface of the device (figure 2). The echodensity was most consistent with a vegetation.

Figure 1.

Transthoracic echocardiogram apical four chamber image showing mobile echodensity (red arrow) on left atrial side of atrial septic defect closure device.

Figure 2.

Transoesophageal echocardiogram image showing 3-mm mobile echodensity (red arrow) on left atrial side of atrial septal defect closure device (yellow arrow).

Differential diagnosis

The initial differential diagnosis for this patient’s febrile illness was broad, and included urinary tract infection, meningitis or a viral illness. The differential diagnosis narrowed when the only positive microbiology result was the positive blood culture. With no clear source for the C. diphtheriae bacteraemia, attention turned to possible infective endocarditis, particularly in the setting of a cardiac murmur and cardiac device in situ.

Our case met one major (echocardiogram evidence of infective endocarditis) and three minor (predisposition, fever >38°C, microbiologic evidence with positive blood culture) clinical criteria to meet Dukes modified criteria for diagnosis of infective endocarditis.²

Treatment

The patient was initially treated with ceftriaxone and aciclovir empirically for possible meningitis. Benzylpenicillin 2.4 g 4-hourly was added when Gram-positive bacilli were identified in blood culture and vancomycin added when it was identified to be C. diphtheriae.

Once sensitivities returned confirming susceptibility to penicillin, and viral PCR of the cerebrospinal fluid returned negative, aciclovir was ceased and antibiotics were rationalised to benzylpenicillin.

Our patient’s symptoms resolved after 24 hours of antibiotics and she remained clinically well with no ongoing microbiological evidence of bacteraemia. Her case was discussed at our institution’s multidisciplinary cardiology and cardiothoracic case meeting, and it was determined to treat medically, rather than surgically, in the first instance. She was discharged on day 16 and completed a total of 6 weeks of benzylpenicillin 14.4 g per 24 hours via a continuous infusion as an outpatient. Due to the possibility of thrombus being associated with the left-sided vegetation, she was commenced on warfarin with a target International Normalised Ratio of 2–3.

Outcome and follow-up

Our patient was treated for C. diphtheriae infective endocarditis of the ASD closure device. She completed 6 weeks of intravenous benzylpenicillin therapy. Repeat TOE at 6 weeks was normal with resolution of the previously visualised echodensity (figure 3). She has remained well through to 6-month follow-up with normal inflammatory markers. Anticoagulation was ceased after 6 months. Follow-up transthoracic echocardiography has shown stable appearance of the ASD closure device.

Figure 3.

Transoesophageal echocardiogram showing resolution of the mobile echodensity on left atrial side of atrial septal defect closure device (yellow arrow).

Discussion

There are few cases of C. diphtheriae-infective endocarditis reported in the literature; a review by Muttaiyah et al ¹ of the literature up to 2007 found 76 cases and our further literature search found an additional 6 reported cases.^5–10 Interestingly, C. diphtheriae endocarditis cases predominantly occurred in persons <30 years of age.¹ Even rarer are cases of ASD closure device-related endocarditis, with a literature review by Amedro et al,¹¹ finding only 21 reported cases, of which 18 underwent surgical removal of the device, and 13 cases were considered ‘late infective endocarditis’ (6 months or more after device implantation). Examining both these groups, no cases of C. diphtheriae endocarditis of an ASD closure device have been reported.

The literature suggests mechanisms of infection in early device-related infective endocarditis include bacteraemia from the puncture site at the time of insertion, or seeding of bacteria before or after device insertion.¹¹ The source of bacteraemia is unclear in this case, and it would be highly unlikely for there to have been an occult source of sepsis from the time of device insertion 9 years ago that has remained dormant until now; it is more likely that there has been a recent unidentified source of bacteraemia. In the Northern Territory (NT) of Australia, there are high rates of non-toxigenic C. diphtheriae cutaneous infection and colonisation. This organism has also been associated with respiratory infection/colonisation, and this may have been the source of bacteraemia in this case.^{12 13} It is also important to remember that toxin production is regulated by the diphtheria toxin repressor gene, which is present in many non-toxigenic strains and may then serve as a potential reservoir for re-emergence of toxigenic strains emphasising the importance of vaccination.¹⁴

This infection described in our case report was from a strain of non-toxigenic diphtheria. The diphtheria vaccine attenuates the local and systemic effects of the toxin only. This does not prevent carriage or colonisation. In a population with high rates of immunisation, toxin production confers no advantage and the metabolic cost puts toxigenic strains at a selective disadvantage. The NT of Australia has immunisation rates of >95%, and colonisation with non-toxigenic strains is not uncommon being a temperate climate.^{12 13} So, although our patient was fully vaccinated, this does not prevent infection with non-toxigenic strains, which has been associated with invasive disease.^15–17 Waning immunity in Australia is more of a concern for patients over the age of 50 without booster doses, who in a place with high rates of childhood immunisation, lose the opportunity for natural immunity and if exposed to toxigenic strains, may not have adequate protective antibodies in response.¹⁸

With respect to treatment of non-toxigenic C. diphtheriae-infective endocarditis, Muttaiyah et al¹ also reviewed the use of gentamicin as an adjunct to penicillin and suggests that there was no difference in the rate of surgical intervention or mortality with dual antibiotic therapy compared with monotherapy.¹ For this reason and the fact that our patient responded well to benzylpenicillin initially, single-agent benzylpenicillin was used, despite the intermediate MIC to penicillin for this isolate. Many early case reports and reviews of C. diphtheriae endocarditis in the literature described an aggressive disease course with cases having complications of cerebral embolic phenomena, pseudoaneurysms, septic arthritis and other septic emboli phenomenon and highlights the importance of early appropriate therapy.^{15 19}

The American Heart Association and European Society of Cardiology have evidence-based guidelines for which highest-risk populations will need future antibiotic prophylaxis for prevention of infective endocarditis.^{20 21} This population includes those with either prosthetic valves or prosthetic material used in a cardiac valve repair, those with previous infective endocarditis, those with congenital heart disease with prosthetic material in situ or untreated congenital heart disease, and those with valve disease after cardiac transplantation.^{20 21} Our patient falls into the second of these groups which places her at a higher risk of future infective endocarditis, therefore requiring prophylactic antibiotics. The procedures for which our patient may require prophylactic antibiotics are also outlined in the two guidelines, which include at-risk dental procedures, surgical procedures involving the skin or muscles, invasive respiratory tract procedures, gastrointestinal or genitourinary procedures where antibiotic therapy would usually be required to prevent or treat established infection, and any cardiac or vascular prosthesis implantation.^{20 21}

This case describes a unique clinical scenario not previously reported and adds to literature already published on treatment approaches in both C. diphtheriae-infective endocarditis and infective endocarditis of ASD closure devices. We have demonstrated a successful outcome with early initiation of treatment, and non-operative management with benzylpenicillin antibiotic monotherapy.

Learning points.

• Corynebacterium diphtheriae is a rare cause of infective endocarditis.

• Similar to prosthetic heart valves, an atrial septal defect (ASD) closure device should be considered as a possible focus of infection in the setting of a febrile illness.

• There are no current guidelines on the treatment of infective endocarditis involving an ASD closure device or involving C. diphtheriae. Optimal antibiotic regimens and indications for surgical removal of the foreign cardiac device are unclear. Multidisciplinary team discussion is therefore recommended.