Background. Candidal infections are an important cause of morbidity and mortality in the very low birth weight (VLBW) infant. Current intervention focuses on treatment once candidal septicemia is either suspected or diagnosed. Studies have suggested that colonization with candidal species is an important risk factor for subsequent infection.
Objective. To determine whether prophylactic fluconazole for the first 28 days of life results in a decreased incidence of candidal colonization in the VLBW infant.
Research Design. Prospective, randomized, controlled, intention-to-treat design comparing prophylaxis with fluconazole versus placebo for the first 28 days of life.
Setting. A tertiary level intensive care nursery in a major teaching hospital in Charleston, South Carolina.
Patients. One hundred three infants with a birth weight of <1500 g, either inborn or outborn, who were admitted to the intensive care nursery between January 1998 and February 1999.
Methods. Infants were enrolled within 72 hours of life with rectal cultures performed on the day of randomization (DOR), as well as day of life (DOL) 7, 14, and 28. Those infants with a birth weight of <1250 g had additional cultures on DOL 35, 49, and 56. Cultures were plated on selective media for isolation of candidal organisms. Infants were randomized to receive either fluconazole (6 mg/kg) or placebo on the DOR. Subsequent doses were given every 72 hours until DOL 7 and then every 24 hours until DOL 28. Medication was given either intravenously or by feeding tube once the infant had been gavage feeding for a 48-hour period without feeding intolerance. Aspartate aminotransferase and alanine aminotransferase levels were obtained on DOR and DOL 7, 14, and 28 to assess for fluconazole toxicity. The minimal inhibitory concentration to fluconazole was determined for all positive cultures to assess the development of resistance.
Results. The infants who received fluconazole (n = 53) and placebo (n = 50) had no statistical difference in the major risk factors known to increase the chances of candidal septicemia in the VLBW infant. Rectal colonization by candidal species was detected in 8 of the 53 fluconazole-treated patients (15.1%) and in 23 of the 50 infants treated with placebo (46%). Fluconazole significantly reduced rectal colonization from DOL 14 through DOL 56 in all infants with a birth weight of <1250 g, and from DOL 14 through DOL 56 in all infants with a birth weight of 1250 to 1500 g. Alanine aminotransferase levels were higher in the fluconazole versus the placebo-treated group on DOL 14 (18.1 IU/L vs 15 IU/L), but no clinically significant abnormalities were observed. Candida albicans was the most common species isolated. There was no increase in species ofCandida noted for their intrinsic resistance to fluconazole, and there was no statistically significant difference in the minimal inhibitory concentrations to fluconazole for all C albicans isolates in either group at any period.
Conclusion. Prophylactic administration of fluconazole to the VLBW infant for the first 28 days of life is safe and results in a decreased risk of rectal colonization by candidal species. Larger studies to determine the effect of prophylaxis on candidal septicemia and development of resistance in such a selective high-risk group are warranted before initiation of routine prophylaxis. fluconazole, very low birth weight infant, prophylaxis, candidal sepsis, sensitivities to fluconazole.
Nosocomial infections are an important cause of morbidity and mortality in premature infants. In a recent study by Fanaroff et al,1 16% of very low birth weight (VLBW) infants developed late-onset infections. Candidal organisms cause 3% to 9%2–4 of these late-onset infections and have a mortality rate of 25% to 60%.4,5 A recent review of outcomes in infants <1250 g found that infants with candidemia or candidal meningitis had a significantly higher incidence of mortality or neurodevelopmental disabilities than do controls (60% vs 28%, respectively; P < .05).5 Published risk factors for candidal sepsis in this group of infants are prolonged intubation,3,6 corticosteroid administration,7 parenteral nutrition,6central line utilization, antibiotics,3,6 and the use of theophylline8; however, colonization with candidal species seems to be the most important predictor of invasive disease.9–14
Various reports indicate that 4.8% to 10% of VLBW infants are colonized at birth and subsequent colonization in the nursery setting results in an incidence of up to 64% by 4 weeks of age.10–14 Of all sites cultured to determine the incidence of colonization in this group of infants, a positive rectal culture seems to be most predictive. Candidal colonization can also occur by horizontal transmission, with the incidence peaking at 1 to 4 weeks of age.10,12–14 It has further been noted that the incidence of colonization correlates inversely with birth weight and gestational age.10–12,14
Fluconazole prophylaxis is effective for prevention of candidal colonization and sepsis in immunosuppressed children and adults.15,16 Neonatal candidal prophylaxis has been attempted with both oral nystatin17 and oral miconazole18 with little success. Several case reports have noted that fluconazole, a bis-triazole antifungal, has been effective for the treatment of VLBW infants with systemic candidal infections.19,21–22 With essentially equivalent oral and intravenous bioavailability, administration of fluconazole is more convenient than conventional antifungal agents such as amphotericin B. Furthermore, adverse effects with fluconazole seem minimal. Transient elevation of liver enzymes, which return to baseline after discontinuation of the drug, has been the only side effect noted with fluconazole in the treatment of VLBW infants for systemic disease with fluconazole.20–23
The purpose of this study was to evaluate the use of fluconazole to reduce candidal colonization, as determined by rectal culture, in the VLBW infant. Our hypothesis was that prophylactic treatment with fluconazole in the first 28 days of life in VLBW infants would result in a decreased incidence of candidal colonization. Reports of long-term prophylaxis with fluconazole in the adult literature have prompted concerns about development of, or selection for, more resistant species of Candida.24–26 Therefore, the effect of prophylactic treatment with fluconazole on the minimal inhibitory concentrations (MICs) of all candidal species isolated from the time of prophylaxis up to 4 weeks after the treatment period in those infants <1250 g at birth was also evaluated.
All infants with a birth weight <1500 g admitted to the neonatal intensive care unit at the Medical University of South Carolina (MUSC) between January 1, 1998 and February 20, 1999 were eligible for participation. Exclusion criteria were parental refusal, admission after 72 hours of life, cyanotic congenital heart disease, congenital defects requiring surgical intervention, fatal chromosomal defects, hepatic insufficiency as demonstrated by aspartate aminotransferase (AST), or alanine aminotransferase (ALT) elevation >4 times normal values for gestational age (>244 IU/L or >152 IU/L, respectively), or an infant so critically ill that he or she was not expected to live for >24 to 48 hours after birth. Patients were enrolled within the first 72 hours of life.
The average length of stay for infants with a birth weight of 1250 to 1500 g is 4 to 5 weeks and no cultures were collected from these infants after day 28 to avoid selecting for infants with an unusually complicated neonatal course. Rectal cultures were performed on the day of randomization (DOR), as well as on day of life (DOL) 7, 14, and 28 in all infants. Infants with a birth weight of <1250 g also had rectal cultures on DOL 35, 49, and 56. In all infants, AST and ALT levels were measured on the DOR, DOL 7, DOL 14, and DOL 28 to assess for hepatotoxicity while on drug or placebo therapy.
Infants were randomized to either fluconazole (6 mg/kg) or placebo (normal saline) intravenously. The investigational drug center (IDC) at MUSC performed randomization by random number assignment27with all others blinded to the assignment group. The drug was given every 72 hours until DOL 7 and then every 24 hours on DOL 8 to 28.23 The drug was given by feeding tube once the infant had tolerated gavage feeding for a 48-hour period. Maalox (CIBA Self-Medication Inc, Woodbridge, NJ) was used as the placebo once infants were changed to oral administration. Pfizer Inc supplied both oral and intravenous fluconazole (Diflucan, Roerig Division of Pfizer, Inc, New York, NY). All doses were distributed by the IDC to be given by the patient's nurse, with all investigators and caregivers blinded to the infant's randomization. After 1 week of life, dosing was adjusted to 3 mg/kg if there was evidence of renal insufficiency as indicated by serum creatinine >1.0 mg/dL. To adjust dosing for weight gain or loss, the IDC was notified of the infant's weight on a weekly basis with dosage adjusted accordingly. All patients were routinely screened for potential drug interactions with fluconazole and other concomitantly prescribed medications.
Rectal cultures were obtained by gentle insertion of a calcium alginate swab that was then streaked at the patient's bedside onto brain–heart infusion agar supplemented with 0.05 g/L of chloramphenicol and 0.05 of g/L gentamicin. Cultures were incubated aerobically at 35°C with examination for growth at 2, 5, and 7 days after plating. Those cultures with growth were further analyzed using standard microbiologic techniques by the department of microbiology at MUSC for identification of candidal species. The MIC to fluconazole for all Candidaisolates was determined at the University of Iowa according to the National Committee for Clinical Laboratory Standards (NCCLS) for susceptibility testing.28
An infant was considered colonized if candidal organisms were isolated from the rectal culture. Charts were reviewed on a weekly basis for known risk factors for sepsis. Duration of antibiotics, intubation, parenteral nutrition, total central catheter use (umbilical venous line, umbilical arterial catheter, peripherally inserted central catheter, Broviac), corticosteroids and theophylline were noted and the duration defined in days (with any portion of a day constituting 1 day of treatment). Infants were removed from the clinical study if they developed AST or ALT levels >3 times reported normal baseline values for neonates or culture positive fungemia (blood, urine, or cerebrospinal fluid).
Power analysis predicted an enrollment of 50 patients in each group would be required to show a reduction in the incidence of colonization from 34% to 14% with a P ≤ .05 and a power of 90%. All data were entered, and statistical analyses for group comparisons were performed, using a standardized data collection tool (Statview for Windows, Abacus Concepts Inc, Berkeley, CA). Descriptive statistics, χ2 test with Yates correction, the Fisher's exact test, and Student's t test were used as appropriate. The study protocol was approved by the institutional review board at MUSC, and informed written consent was obtained from the mother (or maternal grandmother if the mother was considered a minor) before enrollment.
One hundred ninety-one infants were born during the period of enrollment. Eighty-eight infants were excluded from the study. Twelve (14%) were admitted after 72 hours of age, 69 (78%) were excluded because of lack of parental consent, 6 infants (7%) were not expected to survive >48 hours, and 1 infant (1%) had congenital heart disease. Of the 103 infants enrolled in the study, 53 (51.5%) were randomized to the fluconazole treatment group and 50 (48.5%) were enrolled in the placebo treatment group. There were no significant differences between the treatment groups in any of the characteristics examined either at the time of randomization (Table 1) or during the subsequent hospital course (Table 2). Thirty-five of the 53 infants enrolled in the fluconazole-treated group (66%) and 33 of the 50 infants enrolled in the placebo-treated group (66%) were alive and remained hospitalized on DOL 56, the end of the observation period (P = .508). Overall, there were 5 deaths (9.4%) in the fluconazole treatment group and 10 (20%) in the placebo treatment group (P = .131; Table 2).
Effect of Fluconazole Prophylaxis on CandidaColonization
Candida colonization was detected over the course of the study in 8/53 of the fluconazole-treated patients (15.1%) and in 23/50 of the infants treated with placebo (46%; P = .0005). No infant demonstrated initial colonization after DOL 28. Fluconazole significantly reduced rectal colonization from DOL 14 through DOL 28 in all weight groups, which was persistent through DOL 56 in the <1250-g birth weight group (Fig 1). The highest incidence of colonization in the fluconazole group was on DOL 7, when 8 of 52 cultures (15.4%) were positive for a candidal species. In contrast, the highest incidence of colonization in the placebo group was on DOL 28, when 17 of 42 cultures (40.5%) performed were positive. MultipleCandida species were isolated in 1 culture from a fluconazole-treated patient and in 4 cultures from placebo-treated patients. The breakdown of Candida isolates is as follows:Candida albicans in 7 (13%) and 12 (24%),Candida parapsilosis in 1 (2%) and 10 (20%), andCandida glabrata in 1 (2%) and 1 (2%) of fluconazole versus placebo-treated infants. Furthermore, Candida lusitaniae was isolated in 3 (6%) and Candida tropicalis was isolated in 1 (2%) of those infants in the placebo-treated group.
Laboratory Side Effects of Fluconazole
Fluconazole had no clinically significant effect on hepatic enzymes. The ALT concentration in the fluconazole-treated patients (18.1 IU/mL) was significantly higher than in the control patients (15.0 IU/mL) on DOL 14 (P = .008). However, no significant difference in hepatic enzymes was seen on DOL 28 and no infant was removed from the study because of unacceptably high elevation in values or clinical signs of hepatic toxicity.
Effect of Fluconazole Prophylaxis on Development ofCandida Resistance
Analysis of the effect of fluconazole on development of resistance was limited to C albicans since 7 of the 9 (78%)Candida isolates in the treatment group werealbicans species. There were no significant differences in the MICs of isolates from the treated and control patients at any time during the treatment period or for the 4 weeks after discontinuation of the study drug (Fig 2). There was no increase in isolation of species of Candida noted for their intrinsic resistance to fluconazole, with 1 infant in each group colonized with C glabrata. One infant in the fluconazole group remained colonized with C albicans from DOR to DOL 56. Those isolates from this infant were all susceptible to fluconazole with an MIC of 0.12 to 0.25 μg/mL.
Risk Factors for Colonization
Previous reports have noted a greater proportion of colonized infants to have a lower gestational age,10,12 vaginal delivery,14 an increased length of antibiotics,10,12–14 steroids,14intubation,10,14 central line use,10 length of parenteral nutrition,13 Intralipid,13 and theophylline.10 Although the present study was not designed to determine factors associated with colonization, gestational age was found to be statistically associated with colonization in both groups (P < .05; Table 3). Furthermore, vaginal delivery, length of parenteral nutrition, and Intralipid in the fluconazole-treated and length of antibiotics in the placebo-treated group were factors associated with colonization.
Colonization and Infection
Although this study was not designed to assess prevention of candidal infection, there were 2 infants in each group who developed candidal infection. Both infants in the placebo-treated group were colonized before developing sepsis. One was a 1318-g infant who was colonized with C parapsilosis on DOL 28 and developed a urinary tract infection on DOL 42. The second infant, 561 g at birth, was colonized with C parapsilosis from DOL 7 to DOL 35 and C lusitaniae from DOL 14 to DOL 49. This infant developed C parapsilosis fungemia on DOL 99.
Colonization was not detected in either infant in the fluconazole group before the onset of fungemia. The first infant was a 720-g male infant born at 27½ weeks by cesarean section secondary to fetal distress whose clinical course was complicated by severe respiratory distress syndrome, a perforated urachus, and coagulase-negative staphylococcal bacteremia on DOL 12. On DOL 14, his blood culture showed growth of C parapsilosis. The infant was subsequently removed from support on DOL 20 because of intractable pulmonary disease and sepsis. The second infant was a 1200-g male infant born at 27½ weeks by vaginal delivery with maternal preterm labor whose clinical course was complicated by severe respiratory distress syndrome, nonsurgical necrotizing enterocolitis, chronic lung disease, and 4 episodes of nosocomial bacteremia. He then developed a C parapsilosis urinary tract infection on DOL 83 that was successfully treated with fluconazole.
We have performed the first randomized, placebo-controlled, intention-to-treat trial looking at prophylaxis of the VLBW infant with fluconazole to prevent horizontal acquisition of candidal colonization. Our results show that fluconazole prophylaxis reduced the incidence of candidal colonization during the first month of life for all infants with a birth weight of <1500 g from 46% to 15%, a statistically significant reduction. Furthermore, a statistically significant reduction in colonization rates was apparent during the second month of life in those infants with a birth weight of <1250 g. In this relatively small sample of infants studied, there was no evidence of hepatotoxicity, as evidenced by elevated AST or ALT levels, or development of resistant organisms.
The incidence of Candida colonization in our placebo-treated group is comparable to that in previous prospective studies of VLBW infants. Baley et al10 followed 146 VLBW infants and found the highest incidence of colonization on DOL 7 with an incidence 48.7%. Rowen et al14 followed 116 VLBW infants and found that 64% of colonized infants were culture-positive by DOL 14. The earlier peak in colonization incidence in previous studies may be explained by their use of multiple culture sites to assess colonization status.
VLBW infants acquire candidal colonization through both vertical and horizontal means. Our population of infants had an initial colonization rate of 8/103 (7.8%), similar to the incidence in previous reports.9,10,13 Two of the 8 infants with initial colonization were born by cesarean delivery without previous rupture of membranes, were colonized with C albicans, and were in the fluconazole group. In these infants, colonization was either the result of vertical transmission across the placental membranes or of horizontal acquisition that occurred before randomization (both infants were randomized at 68 hours of life).
The use of antimicrobial prophylaxis is associated with the potential risk of selection of resistant organisms. This may take the form of either a subtle increase in the MICs of previously sensitive species or an increased incidence of inherently resistant species. Candida krusei and C glabrata are Candida species with intrinsic resistance to fluconazole, and the use of fluconazole prophylaxis in immunosuppressed adult patients has been associated with an increased incidence of infection with these species.24–26 No increase in either the incidence of isolation of resistant species (defined by the NCCLS as an MIC of 64 μg/mL or above28) or in the MIC of C albicanswas seen in this study. Our study, however, was too short in duration and involved too few patients to determine the risk of selection for resistance. Furthermore, it is important to note that susceptibility testing of fungal isolates is imperfect and MIC breakpoints may not accurately predict in vivo susceptibility.29
The only known side effect of fluconazole in the VLBW infant is transient elevation in liver enzyme values that resolves promptly after discontinuation of therapy.19,23 In our study, the 53 infants who received fluconazole had a statistically significant increase in the ALT value on DOL 14 that was considered clinically insignificant. No infant required removal from the present study because of elevated liver enzyme levels.
This study demonstrates that fluconazole prophylaxis is both safe and effective for reducing the incidence of candidal rectal colonization in VLBW infants. In this relatively small study, there were no clinically significant side effects of the medication; no evidence of development of resistance to fluconazole was detected. This study was not designed to assess the effectiveness of fluconazole prophylaxis for prevention of candidal infection. Evaluation of the effectiveness and the relative risk versus benefit of fluconazole prophylaxis will require a controlled trial in a much larger number of infants. Therefore, routine prophylaxis is not presently advocated in this group of high-risk infants.
The Pfizer Corporation contributed grant support and medication. Supplemental grant support was received from the Medical University of South Carolina Research Development Group.
Sam Messer, PhD, at the University of Iowa, performed the susceptibility testing on all Candida isolates.
- Received March 7, 2000.
- Accepted June 20, 2000.
Reprint requests to (S.D.K.) Wake Medical Center, 3000 New Bern Ave, Box 14465, Raleigh, NC 27620. E-mail:
- VLBW =
- very low birth weight •
- MIC =
- minimal inhibitory concentration •
- MUSC =
- Medical University of South Carolina •
- AST =
- aspartate aminotransferase •
- ALT =
- alanine aminotransferase •
- DOR =
- day of randomization •
- DOL =
- day of life •
- IDC =
- investigational drug center •
- NCCLS =
- National Committee for Clinical Laboratory Standards
- Faix RG,
- Kovarik SM,
- Shaw TR,
- Johnson RV
- Lee BE,
- Cheung P,
- Robinson JL,
- Evanochko C,
- Robertson CM
- Botas CM,
- Kurlat I,
- Young SM,
- Sola A
- Waggoner-Fountain, LA, Walker MW, Hollis, RJ, et al
- Baley JE,
- Kliegman RM,
- Boxerbaum B,
- Fanaroff AA
- Hageman JR,
- Stenske J,
- Keuler H,
- Randall E
- Price MF,
- LaRocco MT,
- Gentry LO
- Rex JH,
- Rinaldi MG,
- Pfaller MA
- Shott S. Statistics for Health Professionals. 1st ed. Philadelphia, PA: WB Saunders; 1990:57–67
- National Committee for Clinical Laboratory Standards. Reference Method for Broth Dilution of Susceptibility Testing for Yeasts: Proposed Strategy. Villanova, PA: National Committee for Clinical Laboratory Standards; 1992. M27-P
- Copyright © 2001 American Academy of Pediatrics