Child’s Nerv Syst (2000) 16:4–7 © Springer-Verlag 2000

K. Murphy J. Bradley H.E. James

Received: 22 February 1999 Revised: 15 July 1999 K. Murphy School of Medicine of the University of California, San Diego, California, USA J. Bradley Division of Infectious Diseases, Department of Pediatrics, Children’s Hospital, San Diego, California, USA H.E. James Division of Neurosurgery, Children’s Hospital, San Diego, California, USA H.E. James (✉) 7930 Frost Street, Suite 103, San Diego, CA 92123, USA Tel.: +1-619-5604791 Fax: +1-858-5609679

B R I E F C O M M U N I C AT I O N

The treatment of Candida albicans shunt infections

Abstract Cerebrospinal fluid shunting procedures are performed for the treatment of hydrocephalus. Infection of ventriculoperitoneal shunts may create significant clinical management issues in these patients. The majority of these infections are bacterial, but occasionally a Candida albicans shunt infection may occur. We report two patients who acquired Candida albicans shunt infection and discuss their clinical presentation, management, and successful outcome. The treatment with or without removal of the shunt and the correct dosage and route of administration of the antifungal agents is not well documented. The dilemma of treatment of Candida albicans shunt

Introduction Infection is a common complication of ventricular shunting. Early infection rates, within 8 weeks of surgical placement of the shunt, range from 3% to 20% and late infection rates, from 2.7% to 31% [11, 18]. The average rate of infection is generally considered to be 6–7% for both early and late infection [14]. The majority of infections are bacterial in etiology, and their diagnosis and treatment are well described [10, 14, 15, 18]. However, a rare cause of shunt infection, with only 30 cases reported in the English literature, is Candida. Fungal shunt infection was first reported by Utz et al., in 1958 [21]. These infections are distinct from candida meningitis in that at the time of diagnosis there is often no detectable inflammation of the meninges. The patient gives the impression of being only mildly ill, often being slightly febrile with no nuchal rigidity, with a normal cerebrospinal fluid

infections in these patients and review of the limited literature on this subject are the subjects of this report. Key words CSF Shunts · Infections · Candida albicans

(CSF) cell count and blood chemistry values. In addition, the symptoms of shunt infection are often insidious, frequently waxing and waning [1]. While central nervous system candida infections related to systemic candida disease are often manifested by widespread organ involvement [22], this is not the case with candidal shunt infections, which makes them less likely to be considered. We present two cases of candida shunt infection, followed by a review of the literature on this topic.

Case reports Patient 1 A 26-week premature female whose neonatal period was complicated by intraventricular hemorrhage required a ventriculoperitoneal (VP) shunt, which was revised once during her initial hospitalization. The patient had no difficulties with the shunt until she presented to the emergency room at 2 years, 7 months of age with:

5

increasing malaise, nonlocalizing abdominal pain, mild headache, several episodes of emesis without diarrhea, and a temperature of 101.3°F. Laboratory results showed a peripheral white blood count (WBC) of 17,000 cells/mm3 with 15% polymorphonuclear leukocytes (PMN), 3% band forms, 34% lymphocytes, and 5% monocytes. The ventricular CSF from the shunt aspirate revealed only a few red blood cells and 11 white blood cells/mm3. The morning after admission, the child was asymptomatic and afebrile. A repeat peripheral WBC was 11,900 cells/mm3 with 40% PMNs, no band forms, 46% lymphocytes, 13% monocytes and 1% eosinophils. The blood culture was sterile, but the CSF culture was positive for Candida albicans on sheep blood agar, chocolate agar, and CMG broth. This was considered to be a contaminant given her clinical improvement without antifungal therapy. The shunt was not retapped. Six months later, the patient was again admitted to the hospital for similar nonspecific symptomology, but now had a new complaint of right lower quadrant pain. The abdominal examination revealed guarding and generalized tenderness. Poor dental hygiene with multiple severe dental caries and areas of erythematous infected gums were noted. Abdominal ultrasound showed a cystic nonechogenic structure adjacent to the distal end of the shunt, confirmed by computed tomography. Ventricular CSF aspirated from the shunt was remarkable for 53 WBC/mm3, with normal concentrations of glucose and protein, and a negative Gram’s stain. The abdominal cyst was tapped under fluoroscopy, and the culture of the fluid yielded Candida albicans. The ventricular fluid grew Candida albicans at a later date. The patient was then treated with intravenous amphotericin B, at 1.0 mg/kg per day. Intraventricular amphotericin B was injected into the shunt at a dose of 1 mg daily for 11 days. The distal end of the shunt was initially externalized until ventricular CSF cultures from the proximal shunt remained negative for 72 h, which occurred 7 days into therapy. At that time a new VP shunt was placed on the opposite side. She tolerated the treatment well with no complications and no recurrent fungal shunt infection, in 12 years of follow-up. She has mild motor and developmental delay, but is independent and attending school. Patient 2 This 2-year-old female had undergone a radical resection of a IV ventricle choroid plexus papilloma with VP shunt placement and 6 months later presented with anorexia, weight loss, occasional diarrhea, emesis and recurrent fever. Analysis of aspirated ventricular CSF revealed 64 WBC/mm3, a glucose level of 34 mg/dl, protein of 74 mg/dl and a negative Gram’s stain. Cultures were positive for Candida albicans. The patient was treated with intravenous amphotericin B at 0.5 mg/kg for the initial dose and then increased to 1 mg/kg per dose for 14 days. Simultaneously, she was given 1 mg of amphotericin B intraventricularly through the shunt for 7 successive days. The distal end of the shunt became obstructed on the 4th day of therapy and was externalized. The child became afebrile 3 days into amphotericin therapy, and following the completion of 14 days of treatment antifungal therapy was stopped for 72 h. Ventricular CSF cultures from the shunt were then obtained and were sterile. The distal end of the shunt was reinserted at that time. The patient was discharged the following day in a clinically stable condition. Subsequently, her VP shunt was converted to a ventriculojugular shunt owing to lack of absorption of CSF in the peritoneal cavity. Cultures at the time of the shunt revision were sterile. However, approximately 4 months later she experienced recurrence of anorexia and lethargy, with irregular respirations and low blood pressure. Her shunt was externalized as the distal end of the shunt was felt to be obstructed. Cultures of the distal shunt yielded Candida albicans, which was sensitive to amphotericin B but resistant to 5-fluorocytosine. The patient was treated for a total of 3 weeks with amphotericin B intravenously at 1 mg/kg per dose. In addition, she received 2 weeks of daily intra-shunt amphotericin B at 1 mg/dose with the first 10 days of therapy given

with the shunt externalized, followed by the placement of a new VP shunt. Ventricular CSF was cultured 48 h after completing the 3-week course of systemic amphotericin B. The cultures at that time, and 2 weeks later, were sterile. She continues to do well 5 years later, and although she has had other shunt revisions, subsequent cultures of CSF have been sterile. She has received no further antifungal therapy.

Discussion Although, the morbidity associated with fungal shunt infections is not well defined, bacterial shunt infections in children are associated with increased mortality rates as high as 10–15%, increased seizure risk, and the possibility of decreased intellectual development [6, 12, 15]. For these reasons, a physician should be aware of the occurrence of and the potential pathogens that may cause shunt infections [10]. Systemic infections are typically encountered in patients with serious underlying medical illnesses, hyperalimentation, systemic corticosteroid therapy, immunosuppressive therapy, acquired immunodeficiency, and instrumentation which breaks cutaneous or mucosal barriers [19]. In the literature which reports the age at which patients became symptomatic, 22 of 28 were children, and 13 of these were premature infants [4, 19]. This may only reflect the increased rate of shunt placement in premature infants, or it may be that prematurity is an associated risk factor for the development of fungal shunt infection [4, 19]. Low-birthweight premature infants have been found to have increased rates of candida infections, with mortality rates following systemic fungal infection as high as 54% [23]. Prior antibiotic treatment was also associated with fungal shunt infections in 13 of 23 cases reported. The pathogenesis of fungal shunt infections is speculative. Those occurring shortly after placement are likely to be the result of contamination at the time of the procedure. Among cases in the literature reporting the interval between shunt procedure and infection, 15 of 24 cases of infection occurred within 8 weeks of a shunt procedure [4, 5, 8, 9, 19, 20]. The interval in the remaining 9 cases ranged from 3 months to 2 years. Thus, contamination at the time of placement is unlikely as the sole explanation for these infections. It is possible that transient episodes of fungemia may result in inoculation of the CSF and shunt, as coincidental candidiasis in the mouth, perineum or urine was noted in 6 of the patients in the literature, with 1 patient also having candida documented in the blood. Children may tolerate fungal shunt infections better than adults [2, 8, 9, 19]. The presentation of fungal shunt infection may at times be similar to that of bacterial infection with shunt obstruction, with symptoms and signs that are often nonspecific: nausea, vomiting, lethargy, anorexia, irritability. Fever is an inconsistent finding. In some instances, complications of shunt obstruction may constitute the only presenting signs and symptoms.

6

Table 1 Intravenous therapy

Drug Intravenous treatment Amphotericin B

Fluconazole

a Patient expired from nonshunt-related causes

Ketoconazole Flucytosine Gentian violet

Blood cultures will usually be sterile, and the peripheral white blood cell count can be normal or elevated [4]. Cerebrospinal fluid examination is often unremarkable, with protein concentration frequently elevated and the glucose concentration usually within normal limits [4]. A mild pleocytosis may be noted [4]. A high index of suspicion for candida infection is the best assurance for an accurate diagnosis. Although previous investigators have suggested that candida found in shunt fluid is merely a contaminant and not a cause of infection [13], our review of the literature suggests otherwise [4, 19, 20]. Review of the literature does not indicate a well-investigated treatment of fungal shunt infections [8, 9]. Skepticism about the efficacy of intravenously administered amphotericin B derives from the poor penetration into the CSF [7, 19]. Doses of amphotericin B previously reported are generally 1 mg/kg per day for patients older than 4 months, and 0.5–0.6 mg/kg per day for infants [4, 19, 20]. Flucytosine was formerly used parenterally for several years in the therapy of candida infections, usually in combination with amphotericin B, achieving CSF concentrations approximately 50% of those in serum. Flucytosine was used in therapy of shunt infections by Rodgers et al. [16]. However, compassionate use of intravenous flucytosine was discontinued in 1988. Two reports describe the use of fluconazole, one at 6 mg/kg per day in a premature infant yielding excellent results [5], and another at 50 mg/kg per day in an 18-month-old with poor results [4]. Cruciani et al. [5] report high and prolonged CSF concentrations of fluconazole and attribute the speedy recovery of their patient to the efficacy of this treatment. There are no previous reports of morbidity related specifically to treatment of fungal shunt infections (Table 1). Extraventricular drainage is often used in treating fungal shunt infections and provides a route for intrathecal administration of antifungals [9, 19]. As with intravenous antifungals, the efficacy of intrathecal administration has not been well studied (Table 2). Some investigators cite concerns regarding reports of Parkinsonism [7], and myelopathy [3], and arachnoiditis, following intrathecal amphotericin B therapy. The intrathecal dosages

Dosage

Cured

Died

0.5 mg/kg (19) 0.6 mg/kg (4) 1.0 mg/kg (4, 19) 30 mg/day (68 years) (20) Not reported (8, 21) 50 mg/kg (4) Not reported (5) Not reported (8) Not reported (9, 16) Not reported (2)

3 1 10 0 1 1 1 1 0 1

0 0 0 1a 1 0 0 0 1 0

Table 2 Intra-shunt treatment Drug

Dosage

Cured

Died

Amphotericin B

0.05 mg/kg (4) 0.10 mg/kg (19) 0.30 mg/kg (19) 0.40 mg/kg (present study) 0.50 mg/kg (19) 4 mg total (20) Not reported (21) Not reported (8) Not reported (9)

2 1 1 1 2 0 0 1 0

0 0 0 0 0 1a 1 0 1

Ketoconazole Flucytosine a Patient

expired from non-shunt-related causes

Table 3 Outcome and treatment methods

Intravenous therapy only [2, 4, 5, 8, 19] Intravenous and intraventricular therapy [4, 9, 10, 19, 20] a One

Cured

Died

10

1

7

3a

patient expired from non-shunt-related causes

used range from 0.05 mg/day [4] to 0.5 mg/day [19]. Some investigators report that intrathecal antifungal therapy may help cure the infection [4, 19] (Table 3). There are some reports that indicate that shunt removal alone cures the infection. Gower et al. [9] report scanning electron microscopy of shunt lumen showing a mat of fungal hyphae and white blood cells adherent to the silicone elastomer tubing. Additionally, Rotrosen [17] reported adherence of Candida species to intravenous catheters, with greater adherence to polyvinyl chloride (PVC) than to Teflon. However, the literature documents that two of the three patients who died as a result of their fungal shunt infection did not have their shunts removed (see Table 4). In our limited experience, it appears that microbiological and clinical outcomes in fungal shunt infections are similar to those in bacterial infections, with removal of the infected shunt and administration of appropriate sys-

7

Table 4 Outcome with and without shunt removal

Shunt removeda [2, 4, 5, 8, 13, 16, 19, 20] Shunt not removed [9, 19, 21] a Shunt completely removed b One patient expired from non-shunt-related

Cured

Died

22

2b

1

2

causes

temic antifungal agents supplemented in certain cases with intraventricular injection of amphotericin B. It is important for the physician to recognize the subtle presentation of candida shunt infections in children, particularly since they are associated with significant risk of morbidity and mortality [24]. Acknowledgements This study has been sponsored in part by the Foundation for Pediatric and Laser Neurosurgery, Inc., San Diego, California, a non-profit corporation.

References 1. Barone SR, Krilov LR (1995) Neonatal candidal meningitis in full term infants with congenital cutaneous candidiasis. Clinical Pediatr 34:217–219 2. Bayer AS, Edwards JE, Seidel JS, Guze LB (1976) Candida meningitis: report of seven cases and review of the English literature. Medicine (Baltimore) 55:477–486 3. Carnevale NT, Gagliani JN, Stevens DA, Herrick MK, Languin JW (1980) Amphotericin-B induced myelopathy. Arch Intern Med 140:1189–1192 4. Chiou CC, Wong TT, Lin HH, Hwang B, Tang RB, Wu KG, Lee BH (1994) Fungal infection of ventriculoperitoneal shunts in children. Clin Infect Dis 1049–1053 5. Cruciani M, et al (1992) Use of fluconazole in the treatment of Candida albicans hydrocephalus shunt infection (letter). Eur J Clin Microbiol Infect Dis 11:957 6. Dennis M, Fitz CR, Netley CT (1981) The intelligence of hydrocephalic children. Arch Neurol 38:607–615 7. Fisher JF, Dewald J (1981) Parkinsonism associated with intraventricular amphotericin-B. J Antimicrob Chemother 12:97–109 8. Gower DJ, Crone K, Alexander E Jr, Kelly DL Jr (1986) Candida albicans shunt infection: report of two cases. Neurosurgery 19:111–124

9. Hiruma M, Nitta M, Kagawa S (1986) A case of cerebrospinal fluid shunt colonization and obstruction by Candida albicans. Mykosen 29:255–260 10. James HE (1991) The spectrum of treatments for shunt infections. Concepts Pediatr Neurosurg 11:54–59 11. James HE, Walsh J (1982) Intraventricular cerebrospinal fluid antibiotic concentrations in patients with intraventricular infections. Neurosurgery 10:50–54 12. McCullough DC, Baylor-Martin LA (1982) Current prognosis in overt neonatal hydrocephalus. J Neurosurg 57:378–383 13. O’Brien M, Parent A, Davis B (1979) Management of ventricular shunt infections. Child’s Brain 5:304–309 14. Quigley MR, Reigel DH, Kortyna R (1989) Cerebrospinal fluid shunt infections: report of 41 cases and a critical review of the literature. Pediatr Neurosci 15:111–120 15. Raimondi AJ, Soare P (1974) The intellectual development in shunted hydrocephalus. Am J Dis Child 127:664–671 16. Rodgers BM, Vries JK, Talbert JL (1978) Laparoscopy in the diagnosis and treatment of malfunctioning ventriculoperitoneal shunts in children. J Pediatr Surg 13:247–253 17. Rotrosen D, Gibson TR, Edwards JE (1983) Adherence of Candida species to intravenous catheters. J Infect Dis 147:594

18. Schoenbaum SC, Gardner P, Shillito J (1975) Infections of cerebrospinal fluid shunts; epidemiology, clinical manifestations and therapy. J Infect Dis 131:543–552 19. Shapiro S, Javed D, Mealy J Jr (1989) Candida albicans shunt infection. Pediatr Neurosci 15:125–130 20. Sugarman B, Massanari RM (1980) Candida meningitis in patients with CSF shunts. Arch Neurol 180–181 21. Utz JP, Treger A, McCullough NB, Emmons CW (1958–1959) Amphotericin B. Intravenous use 21 patients with systemic fungal diseases. Antibiot Annu 1958–1959:628–634 22. Voice RA, Bradley SF, Sangeorzan JA, Kauffman CA (1994) Chronic candidal meningitis: an uncommon manifestation of candidiasis. Clin Infect Dis 19:60–66 23. Weese-Mayer DE, Fondriest DW, Brouillette RT (1987) Factors associated with candidemia in the neonatal intensive care unit: a case control study. Pediatr Infect Dis J 6:190–196 24. Yoger R, Davis AT (1980) Neurosurgical shunt infections. Child’s Brain 6:74–81