First Reported Case of Candida dubliniensis Endocarditis Related to Implantable Cardioverter-Defibrillator

A 36-year-old male presented to the ED with acute chronic hyponatremia found on routine weekly lab work with one-week history of generalized weakness, confusion, nausea/vomiting, and diarrhea. The patient has nonischemic cardiomyopathy of unknown etiology diagnosed in his teens with an AICD device placed 8 years ago and receiving milrinone infusion 3 years ago via peripherally inserted central catheter (PICC) line. Two sets of blood cultures grew Candida dubliniensis. The patient was started on micafungin and the PICC line was removed and replaced with a central line. A transthoracic echocardiogram (TEE) showed findings consistent with AICD lead involvement. The patient was continued on treatment for fungal infective endocarditis and transferred to another hospital where he had successful AICD lead extraction. Blood cultures upon transfer back to our facility were positive for methicillin-sensitive Staphylococcus aureus (MSSA). This bacteremia was thought to be secondary to right-sided internal jugular (IJ) central line and resolved with line removal and initiation of intravenous (IV) cefazolin. The patient was discharged on IV cefazolin and IV micafungin. He had a LifeVest® until completion of his antibiotic course and a new AICD was placed.


Introduction
Automatic implantable cardioverter-defibrillator (AICD) devices are used for a wide variety of cardiac conditions including secondary prevention of sudden cardiac death for patients with a previous cardiac arrest who have sustained ventricular tachycardia, coronary artery disease, nonischemic dilated cardiomyopathy, hypertrophic cardiomyopathy, and genetic arrhythmia syndromes. AICDs can also be used for the primary prevention of sudden cardiac death. [1].
Complications of AICD implantation include infections that may include vegetations on the leads or valves [2,3]. Estimates of the rate of infection postimplantation range from 0.13% [4] to 19.9% [5]. Pocket infections related to AICD devices usually involve staphylococci and occur within weeks of implantation.
Systemic AICD device infections were less common, presented later, and typically involve a wider range of microbes compared to pocket infections. Current guidelines call for AICD removal with TEE demonstrating valve or lead vegetation, presence of pocket infection, or positive blood cultures with S. aureus, coagulase-negative staphylococci, Cutibacterium, Candida species, or other high-grade bacteremia without alternative source [6]. Antimicrobial treatment without device removal has a high reinfection and failure rate [7]. In most cases, the long-term harm to the patient of repeated infections and mortality outweigh the risks of immediate extraction with other approaches like device retention and antibiotic therapy having a higher risk of mortality [8,9].
Candida-related ICD infection is extremely rare and there are only 23 Candida device-associated endocarditis cases published in the literature (Table 1) [10][11][12][13]. The majority of patients are over age 60 and had an infection with a time postimplantation ranging from less than a month to 16 years [14].
We report on a unique case of AICD-related fungemia complicated by simultaneous use of a chronic milrinone infusion. We describe the rare infection, the path to diagnosis, treatment, and postoperative complications observed prior to discharge.
We report the first case of Candida dubliniensis as a source of ICD fungemia in the English language literature. Candida dubliniensis is an opportunistic pathogen that was originally isolated from AIDS patients, but since then, it has been isolated from both immunocompetent and immunocompromised individuals [15,16]. Furthermore, invasive infections by C. dubliniensis have increased considerably in recent years in immunocompromised/immunosuppressed individuals [16].

Case Presentation
A 36-year old male presented to the ED with generalized weakness, confusion, and fatigue starting one week prior. The symptoms were associated with some episodes of nausea, vomiting, and loose semisolid stools. The patient reported shortness of breath that had progressed to rest over the past weeks. He denied any fever but reported some chills. He denied any chest pain, palpitations, lower extremity swelling, abdominal pain, or other complaints. The patient has routine weekly labs which showed hyponatremia (sodium = 122 mmol/dL, baseline 130 mmol/dL) so he was sent for evaluation to the emergency room by his cardiologist. The patient has a past medical history of nonischemic cardiomyopathy diagnosed at age 16. A transthoracic echocardiogram done about 9 months prior showed a left ventricular ejection fraction of less than 20% and severe concentric left ventricular hypertrophy. The patient had an automatic implantable cardioverter-defibrillator (AICD) placement 8 years prior for primary prevention of sudden cardiac death. The patient had been on a chronic milrinone infusion delivered through a peripherally inserted central catheter (PICC) line for the past 3 years. This was initiated as a bridge to transplant. However, during transplant evaluation, he was noted to have secondary pulmonary hypertension and would need a com-bined heart and lung transplant, and no transplant center in the state would accept his insurance for a combined transplant. Besides cardiomyopathy, he also has a history of chronic atrial fibrillation, congenital hydrocephalus with ventriculoperitoneal shunt (since the age of 2), and spinal stenosis. Of note, the patient was taking apixaban, bumetanide, magnesium oxide, allopurinol, digoxin, metolazone, eplerenone, and carvedilol.
On exam, the patient had a red, swollen left upper extremity at the site of his PICC and white nail beds on the left hand and cool extremities, but no lower extremity edema. The patient was afebrile, had an elevated heart rate of 109 beats per minute (bpm), an elevated respiratory rate of 26 breaths per minute, and SpO2 was 96% on 2 L nasal cannula. All other vitals were within normal limits.
Labs on admission are as follows: WBC: 10:9 × 10 3 /mcL; sodium: 123 mmol/L; potassium: 4.7 mmol/L; chloride: 90 mmol/L; bicarbonate: 22; BUN: 27 mg/dL; creatinine: 0.9 mg/dL; glucose: 102 mg/dL; INR: 1.56; total bilirubin: 4.12 mg/dL; alkaline phosphatase: 166 units/L; AST: 44 units/L; and ALT: 27 units/L. The patient was initially diagnosed with viral gastroenteritis and his symptoms of nausea, vomiting, and diarrhea resolved within 2 days. As the team was preparing to discharge the patient, 2 sets of blood cultures taken at admission came back positive for Candida dubliniensis. The patient's WBC count increased from 11.0 to 15.2 with continued stability in his vital signs other than an increase in his heart rate to 120 s bpm. A chest X-ray on admission showed no acute abnormalities, stable cardiomegaly, and an AICD in place. An electrocardiogram (EKG) showed atrial flutter. CT scan of the head showed a left VP shunt with mild ventriculomegaly of the lateral and third ventricles unchanged since 2017. No evidence of intracranial hemorrhage was found. A lumbar puncture obtained clear CSF that was sent to microbiology. The CSF cultures had no growth. The patient was started on empiric micafungin 100 mg IV daily and vancomycin.
A blood culture of the PICC line and a culture of the catheter tip was positive for Candida dubliniensis as determined by matrix-assisted laser desorption/ionization time-of-flight (MALDI-TOF) mass spectrometry. Minimum inhibitory concentrations (MICs) for antifungals are listed in Table 1.
The PICC line was removed and a right-sided internal jugular vein (IJV) central line was placed. Vancomycin was discontinued after one day once the culture grew yeast. A negative blood culture of C. dubliniensis was drawn 4 weeks after the first positive culture. Micafungin was continued for a total of 6 weeks after the initial positive culture.
Transesophageal echocardiogram (TEE) ( Figure 1) showed a new finding of echogenicity of one of the three leads consistent with lead vegetation. Consistent with a TEE done 8 months ago, there was minimal thickening of the aortic valve, mild thickening of the mitral valve with moderate to severe mitral regurgitation, and moderate to severe tricuspid regurgitation. One of the leads had a mobile echodensity measuring 0:4 × 1:3 cm, consistent with lead vegetation. No valvular vegetations were identified on TEE. The findings of the TEE were consistent with AICD fungal septicemia. An AICD lead extraction was done one week after admission at

Discussion
With the increasing use of AICD devices for the secondary and primary prevention of cardiac death, clinicians should be vigilant of the possibility of infection caused by AICD placement. Risk factors that can increase infection rates include heart failure, diabetes, renal disease, immunosuppressed state, oral anticoagulation, chronic lung disease, and recent device modifications [10,17]. Our patient had several comorbidities that increased his risk of an infection including use of oral anticoagulation, a history of heart failure, and pulmonary hypertension. His primary risk factor, however, was chronic PICC line for milrinone infusion, despite PICC being exchanged about every 6 months and the patient practicing adequate PICC care at home. Our patient's scenario was unique because, to our knowledge, it is the first reported case of C. dubliniensis-ICD fungemia with sepsis, in the English-language literature. He denied fever which was not uncommon in candidemia. Additionally,     6 Case Reports in Cardiology his age was younger than the average of the other case reports which were related to AICD-Candida infection (see Table 2); he has nonischemic cardiomyopathy diagnosed at age 16 and has multiple comorbidities that complicate diagnosis and treatment.

Case Reports in Cardiology
Candida dubliniensis was first recovered postmortem from a lung specimen of a patient who expired in the United Kingdom in 1957. This organism was originally misidentified as C. stellatoidea. This species was discovered to be unique from other Candida species in the city of Dublin, resulting in the name Candida dubliniensis. This organism was found in mainly oral cavities, primarily those of HIV-infected individuals. Though this species has been found in oral cavities of healthy individuals and has been implicated infections in healthy individuals, it is rare for this organism to produce infection in immunocompetent adults. It appears that diminished T-cell activity, such as that seen in HIV, could allow overgrowth of C. dubliniensis [15,18].
This species only accounts for a small percentage of Candida infections [15]. However, infections associated with this species have been reported (Table 3). While the table represents many of the cases in which this organism was implicated, it is by no means an exhaustive list. In reviewing these cases, it seems C. dubliniensis is most commonly found in those with some degree of immunocompromise. In particular, many reported cases have occurred in patients with liver disease and/or substance abuse. Understanding the population at risk for this infection can help clinicians identify this organism in a timely manner.
C. dubliniensis has demonstrated resistance to antifungal agents, specifically fluconazole. While these resistant strains do not account for the majority of strains, caution should be exercised when prescribing empiric therapy. In addition, fluconazole resistance is easily developed in vitro. This further highlights the need for caution and utilization of other antifungal agents if C. dubliniensis is found [15,17].
Although our patient had some signs of sepsis including an elevated heart rate and respiratory rate, he was afebrile and appeared to have viral gastroenteritis. After a WBC count spike and Candida growth 2 days later, our team suspected sepsis, but did not know the source. The possibilities included the PICC line for milrinone infusion, an AICD infection, and an infection from the left VP shunt. The TEE identified a clear AICD lead infection and a CT head without contrast combined with negative CSF cultures ruled out a VP shunt infection. Additionally, the TEE detected no valvular vegetations which lowered the possibility of a PICC linerelated infective endocarditis. The TEE and microbiology results made a strong case for Candida dublinesis infective endocarditis related to the AICD leads. The AICD device was removed at an outside facility due to patient preference with implantation of a new AICD device 5 days later. Finally, an MSSA infection related to the right-side IJV line placed in the hospital for antibiotic infusion delayed the patient's discharge further.
The patient was on IV micafungin for treatment of Candida dubliniensis. Recommendations from the Infectious Disease Society of America recommend the use of echinocandin (such as micafungin), amphotericin B, or amphotericin B with 5-flucytosine for treatment of Candida endocarditis with or without an ICD device [32,33]. Regardless of the medication chosen, resolution of the fungemia  [30] 59, M COPD, diabetes mellitus type 2, ulcerative colitis Pneumonia [31] 45, F None reported Keratitis 7 Case Reports in Cardiology usually occurs with ICD device removal, although one case report describes resolution of Candida endocarditis with medical management and no surgical intervention [34]. In our patient, treatment was continued for 4-6 weeks after ICD device removal.

Learning Points/Take Home Messages
(i) Candida dubliniensis is an opportunistic pathogen originally isolated from AIDS patients but can be isolated from immunocompetent individuals (ii) We report the first case in literature of Candida dubliniensis-ICD sepsis with fungemia and multiple possible sources of infection including a VP shunt and an indwelling PICC for milrinone infusion (iii) Clinicians should do a full sepsis workup for all patients admitted to the hospital with an ICD device, tachycardia, and tachypnea even in absence of fever and leukocytosis