Challenges in the diagnosis and management of Granulicatella elegans endocarditis in a 9-year-old child

Abstract

A 9-year-old child, with a background of repaired pulmonary atresia and Ebstein’s anomaly, presented with fever, night sweats and lethargy. Blood cultures grew Granulicatella elegans, a nutritionally variant Streptococcus and known cause of infective endocarditis (IE). Echocardiogram revealed no clear vegetation, but increased stenosis of the right ventricle to pulmonary artery conduit. The child was successfully managed with high-dose benzylpenicillin, completing 2 weeks in the hospital, and was discharged to complete the final 4 weeks of therapy with ceftriaxone in the community, as per European Society of Cardiology guidance. IE caused by any Granulicatella species is rare, with infection due to G. elegans rarer still. It is a Gram-positive bacteria that presents a diagnostic challenge due to non-specific symptoms at presentation and difficulty in growing the organism on culture medium. We present a case of G. elegans endocarditis in a young child, which illustrates the challenges in managing this condition and the importance of considering atypical organism endocarditis in children presenting with fever of unknown origin, in particular those with a background of congenital cardiac disease. We review the literature on Granulicatella endocarditis, and briefly discuss the challenges of managing this condition in a child with an autism spectrum disorder and learning difficulties.

Background

Granulicatella elegans is a rare cause of infective endocarditis (IE), with less than 10 previous known case reports, and no known previous reports of this organism causing IE in children. Granulicatella endocarditis has been shown to have an indolent and subacute course, with non-specific symptoms and only mildly raised inflammatory markers. Yet it is associated with significant morbidity, and mortality in up to 20% of cases. Here, we describe a case that demonstrates the diagnostic and therapeutic challenge in children and reminds clinicians to be aware of these bacteria and the advised antimicrobial management.

Case presentation

A 9-year-old child with an extensive cardiac history was admitted with fevers and night sweats for over a week.

The child had been referred to the local acute paediatric service by the regional paediatric cardiology team, whom the child’s mother had contacted due to her concerns about the fever. On review in the emergency department, the history was of intermittent fever for the past 8 days, mostly low grade, with the last recorded high temperature being 38.7°C 3 days previously. On further questioning, the mother reported that the child had become lethargic 8 days prior to admission, with night sweats and slightly reduced appetite. There was also a very mild cough, slight breathlessness, muscle aches and mild abdominal discomfort.

The child was drinking good amounts of fluids and there were no urinary symptoms. There were no unwell contacts, and the whole family had been isolating during the COVID-19 pandemic, due to the child’s increased vulnerability status. There was no recent travel history.

The cardiac history consisted of a diagnosis of pulmonary atresia with intact interventricular septum and mild Ebstein’s malformation. Initial surgical intervention included right ventricle to pulmonary artery (RV-PA) conduit and Blalock-Taussig shunt in the neonatal period. Balloon dilation of the pulmonary arteries was undertaken on a number of occasions over the subsequent 3 years, along with transcatheter occlusion of the shunt. Surgical replacement of the RV-PA conduit with a 21 mm bioprosthetic conduit, right pulmonary artery augmentation and tricuspid valve repair was then performed at 5 years of age. They had remained well from a cardiac point of view since that stage, under regular follow-up with the regional paediatric cardiology service.

The child was otherwise fit and well, although also known to the community paediatric team with an autism spectrum disorder and mild learning difficulties. They were on no regular medication, with no known allergies and routine immunisations were up to date.

On examination, the child appeared generally well, with observations within the normal range. There were no peripheral stigmata of endocarditis, such as splinter haemorrhages, Osler’s nodes and Janeway lesions, nor was there any rash or features of Kawasaki’s disease. Oral examination did not reveal any evidence of cavities or poor oral hygiene, and there was no history of recent dental procedure; however, formal dental examination was not performed. Pulses were of good volume and there was a previously documented grade 3/6 ejection systolic murmur at the left sternal edge. There was no hepatosplenomegaly, and the remainder of the examination was unremarkable.

Investigations

Baseline blood and an initial blood culture were taken on admission. A viral swab for Sars-CoV-2 was negative.

Initial blood results were as follows: haemoglobin 137 g/L (115–140 g/L), mean cell volume 83 (77–91 fL), white cell count 8.72×10⁹/L (3.0–10.0×10⁹/L), platelet count 165×10⁹/L (150–450×10⁹/L), neutrophil count 7.34×10⁹/L (2.0–6.0×10⁹/L), lymphocyte count 0.69×10⁹/L (1.5–4.0×10⁹/L), sodium 136 mmol/L (135–146 mmol/L), potassium 4.5 mmol/L (3.5–5.5 mmol/L), urea 3.4mmol/L (2.5–6.5 mmol/L), creatinine 36 µmol/L (29–53 μmol/L), albumin 42 g/L (35–50 g/L), alanine aminotransferase 15 IU/L (0–36 IU/L), alkaline phosphatase 161 IU/L (76–308 IU/L), C reactive protein (CRP) 58 mg/L (<5 mg/L).

By 12 hours into her admission, Gram-positive cocci in chains were isolated from the blood cultures. At this point, two further blood cultures were taken, 1 hour apart and from separate sites, and intravenous antibiotics started immediately following this.

Scanty growth was observed on blood and chocolate agar after 48 hours at 37°C. The organism was identified as G. elegans—a nutritionally variant Streptococcus (NVS), by Bruker matrix-assisted laser desorption/ionization-time of flight mass spectrometry (Bruker, Massachusetts, USA). The second and third blood cultures were both positive for the same organism. The minimum inhibitory concentrations (MICs) reported were penicillin 0.04 mg/L (<0.125 mg/L), ceftriaxone 0.06 mg/L, vancomycin 0.50 mg/L and teicoplanin 0.032 mg/L by Etest.

Transthoracic echocardiogram did not demonstrate any vegetations, although acoustic windows of the conduit had always been limited and it remained difficult to assess this area in any detail. However, Doppler of the proximal conduit demonstrated a significant increase in velocity (3.5 m/s) compared with the previous echocardiogram a year before. There was moderate tricuspid regurgitation, which was also of higher velocity (4.5 m/s) than before, confirming an increase in the degree of conduit obstruction and RV pressures. There was moderate right atrial and ventricular dilation with preserved systolic ventricular function.

Given the elevated risk status of the patient for IE and evidence of persistent bacteraemia by an organism known to cause IE with new evidence of conduit stenosis on echocardiogram, IE was the most likely diagnosis to explain the clinical presentation of the patient. While transoesophageal echocardiogram may have given more detailed imaging of the conduit, given the high clinical suspicion, it was felt unlikely this would change management strategy and would have required general anaesthetic to undertake, given the child’s age.

Treatment

The child was initially treated empirically with benzylpenicillin and cefotaxime pending organism identification, after discussion with the regional cardiology centre. However, this was amended the following day to intravenous amoxicillin, based on guidance from microbiology, to provide empirical cover for Enterococci as well as non-enterococcal Streptococci pending organism identification.

Following organism identification and demonstrated penicillin susceptibility, benzylpenicillin 50 mg/kg every 4 hours was restarted. Synergistic gentamicin 3 mg/kg was added, once a day for the first 2 weeks as per European Society of Cardiology (ESC) guidelines, due to the risk of penicillin tolerance from monotherapy. The patient, unfortunately, developed an immediate urticarial reaction following the first dose of gentamicin, and a more severe urticarial reaction on repeated exposure, necessitating immediate antihistamine treatment and stoppage of infusion during the second dose.

A decision to continue with benzylpenicillin monotherapy was made based on: the NVS isolate being G. elegans, with penicillin MICs consistently less than 0.125 mg/L on repeated testing by Etest, the absence of large valve vegetation on echocardiography or any associated complications, such as perivalvular abscess, the absence of metastatic seeding of disease at presentation, and the risk of developing more severe type I hypersensitivity reaction on repeated exposure to gentamicin. However, it was opted to carefully monitor the clinical response to treatment while on benzylpenicillin monotherapy, and later on ceftriaxone, given the risk of treatment failure from potential penicillin tolerance with monotherapy, despite penicillin susceptibility in vitro. In the event of treatment failure, the plan was to escalate to amoxicillin plus ceftriaxone, as per the Paediatric American Heart Association guidelines, to treat like enterococcal IE despite limited evidence.

A peripherally inserted central catheter (PICC) was inserted under general anaesthetic in the regional cardiology centre, to facilitate antibiotic delivery and blood sampling. The antibiotic treatment duration was for a total of 6 weeks followed by a clinical review.

Outcome and follow-up

There was excellent clinical and microbiological response to treatment. All subsequent blood cultures were sterile and inflammatory markers normalised. There was no evidence of metastatic seeding. The child remained clinically well with no further fevers after 2 weeks of benzylpenicillin monotherapy and 20 days total antibiotic treatment. She was discharged home to complete a further 4 weeks of intravenous ceftriaxone 100 mg/kg/day treatment in the community, as per ESC guidance.

Normal inflammatory markers were sustained. Multiple surveillance blood cultures on and off antibiotic therapy remained negative, with no evidence of bacteraemia relapse, even when tested 2 months following completion of antibiotic regimen. There was no recrudescence of fever or any of the symptoms at presentation. Repeat echocardiogram confirmed stable appearances in the degree of conduit obstruction and she has subsequently undergone successful RV-PA conduit replacement with a 25 mm pulmonary homograft for ongoing severe conduit stenosis. She continues to remain clinically well and is closely followed up by both local and tertiary centre teams.

Discussion

We reviewed all cases of G. elegans endocarditis, of which, to the best of our knowledge, there are no previous paediatric case reports. We also reviewed the literature on IE in children due to other NVS.

In developed countries, IE is a rare, but life-threatening condition in children, affecting approximately 0.05–0.12/1000 each year.¹ This incidence is, however, significantly increased in the presence of congenital heart disease to 6.1/1000.² Congenital heart disease is now the prominent underlying condition in IE, accounting for over 75% of cases, having overtaken rheumatic heart disease in the last century.³ IE in children is most frequently caused by Staphylococcus aureus and viridans group of Streptococci; however, Granulicatella species count for a rare, but recognised cause.⁴

Known as NVS due to the need for additional agents (L-cysteine and pyridoxal hydrochloride) to be incorporated into standard media for successful isolation, Granulicatella are fastidious cell wall-deficient Streptococci which were first described in 1961 by Frenkel and Hirsch.⁵ They are catalase-negative, oxidase-negative, facultative anaerobic Gram-positive bacteria.⁶ Based on 16S rRNA gene sequencing studies in the 1990s, they were reclassified by Collins and Lawson into the separate genera Abiotrophia and Granulicatella, which was subdivided into Granulicatella adiacens, Granulicatella balaenopterae and G. elegans.⁶ They are a component of normal oral, upper respiratory and urogenital flora in humans, often found in dental plaque and dental abscesses, and a rare but documented cause of IE.⁷

As a genus, Granulicatella species are thought to responsible for 6% of cases of IE.^{7 8} Of these, G. adiacens is the most prevalent, with G. elegans accounting for just 8%, therefore, accounting for 0.48% of IE overall.^{8 9} However, these fastidious bacteria are both difficult and slow to grow, often leading to delays in obtaining positive blood culture results, which presents a significant diagnostic challenge. It is thought, therefore, that cases of Granulicatella endocarditis may commonly be under-reported as ‘culture-negative’ endocarditis,^{7 9} where no microbiological diagnosis is found. This ‘culture-negative’ endocarditis has been shown to account for 8%–36% of all IE cases in children.⁴

Previous case reports of G. elegans endocarditis in adults, or any NVS endocarditis in children, show that patients tend to present with prolonged low-grade fever alongside non-specific constitutional symptoms, listing fatigue, malaise, generalised weakness and arthralgia among the most common.^10–16 Classic peripheral stigmata of IE are now rare, especially in children. This vague presentation can lead to diagnostic uncertainty, such as in our patient’s case. Due to her vague and mild symptoms, endocarditis was not suspected in the first instance, and only one set of blood cultures was taken and antibiotics not started.

The extent of elevation of inflammatory markers in IE has been shown to vary depending on the causative organism. A 2005 study showed that white cell count is generally only mildly raised on admission, but CRP was raised to a mean of 193 mg/L in S. aureus endocarditis, and to 232 mg/L in disease caused by Streptococcus pneumoniae.¹⁴ The same study also revealed that CRP is higher in native valve disease (137 mg/L) than in prosthetic valve disease (86 mg/L).¹⁴ This study did not present any data on CRP rise in G. elegans endocarditis specifically; however, case reports have shown CRP levels of <10 mg/L.^{13 14} Our patient had a higher CRP of 58 mg/L. Relatively low inflammatory markers add to the challenge of recognising and diagnosing G. elegans endocarditis.

Granulicatella endocarditis has been associated with large vegetations,10–14 16–18 with one case series showing mean vegetation length of 16 mm with multivalvular involvement in 13% of cases.¹⁶ This is significant given that vegetations of greater than 15 mm are a strong predictor of embolic events and mortality.¹⁷ In the case of our patient, no clear vegetation was seen on echocardiogram; however, given the clinical picture, the strong risk factors, persistent bacteraemia caused by an organism known to cause IE and the significant new narrowing of the RV-PA conduit on echocardiogram, the decision was made to manage as endocarditis.

Scoring tools, such as the HANDOC score, have been used to evaluate the need for echocardiography in adult patients with streptococcal bacteraemia. These scoring systems look at various factors, including: presence of valve disease, number of positive blood cultures and duration of symptoms; with a score of over 3 being highly sensitivity and specific for endocarditis. Our patient would score highly, with a score of 5, however, the HANDOC score has not been specifically validated in the paediatric population or in patients with congenital heart disease.¹⁸

Granulicatella organisms exhibit variable penicillin sensitivities due to deficient cell walls, and treatment failures can occur despite penicillin sensitivity, due to inherent penicillin tolerance.¹⁹ Among NVS, G. elegans are more susceptible to penicillin compared with G. adiacens or Abiotrophia species (Abiotrophia defectiva) with MICs <0.125 mg/L.²⁰ The 2015 ESC guidelines state, therefore, that Granulicatella endocarditis should be treated with benzylpenicillin, ceftriaxone or vancomycin for 6 weeks, along with a synergistic aminoglycoside for the first 2 weeks.²¹ Accordingly, our patient was managed with gentamicin and benzylpenicillin once the organism was identified, although this regime required adaptation as the patient developed a severe urticarial rash with gentamicin. In our case, the child responded well to antimicrobial therapy with sterile subsequent blood cultures; however, treatment failure is common, in children as well as adults.¹³ Valve replacement surgery is often required in Granulicatella endocarditis, in up to 65% of adult cases according to a 2015 case series.¹⁶

The literature shows that Granulicatella endocarditis presents a higher risk of morbidity and mortality than endocarditis secondary to more common causative bacteria. The 2015 case series referenced above found that of 23 patients with Granulicatella endocarditis, 30% developed heart failure, 30% suffered embolic complications and perivalvular abscess occurred in 11%.¹⁶ There was a mortality rate of around 21%, compared with mortality of 0%–12% in Streptococcus viridans endocarditis.¹⁶ It is worth noting, however, that this data is limited in its relevance to our case by both its small sample size and that the average age of the patients was 66 years, with no patients under 18 years. It is not clear whether mortality is also significantly increased in the paediatric population.

Other complications, in cases specifically of G. elegans endocarditis, have included splenic embolisation, seizures,¹⁶ intracerebral aneurysms,²² discitis¹¹ and haemodynamic compromise.¹⁶ Fortunately, in the case of our patient, there was good response to antimicrobial therapy and the child remained afebrile with no clinical complications for the remainder of the treatment course.

Our case also illustrates the ‘real life’ challenges of managing a lengthy antibiotic course, or any prolonged hospital stay, in a child with autism. It required additional factors to be considered. For example, as far as possible minimising the number of times blood samples for gentamicin dosing had to be obtained. Samples were taken from the PICC line or by finger prick where possible. Initially, a plan had been made for ambulation in between antibiotic doses once stable; however, this was distressing for the child, so they remained on the ward until final discharge. Antibiotic administration could be stressful for the child, as was being in an unfamiliar environment, so the assistance of play therapists proved invaluable in this case. This child’s care required a true multidisciplinary approach, with involvement from acute paediatricians, community paediatricians, paediatric cardiologists, microbiologists, paediatric nurses and play therapists.

Learning points.

• Infective endocarditis (IE) should be suspected in children presenting with fever of unknown origin.

• Clinicians should be aware of Granulicatella bacteraemia and endocarditis, and consider this in cases of children with prolonged fever and non-specific symptoms, especially those with cardiac disease.

• Granulicatella endocarditis can present indolently, and with only mildly raised inflammatory markers, but morbidity and mortality can be significant.

• Clinicians should appreciate the guidance on antimicrobial therapy in Granulicatella endocarditis, where penicillin tolerance may be an issue.

• The management of IE in children in hospital requires a multidisciplinary approach, especially those with additional needs, such as autism.