|
|
Background Invasive candida infections are a major cause of morbidity and mortality in preterm infants. We performed a multicenter, randomized, double-blind, placebo-controlled trial of fluconazole for the prevention of fungal colonization and infection in very-low-birth-weight neonates.
Methods During a 15-month period, all neonates weighing less than 1500 g at birth from eight tertiary Italian neonatal intensive care units (322 infants) were randomly assigned to receive either fluconazole (at a dose of either 6 mg or 3 mg per kilogram of body weight) or placebo from birth until day 30 of life (day 45 for neonates weighing <1000 g at birth). We performed weekly surveillance cultures and systematic fungal susceptibility testing.
Results Among infants receiving fluconazole, fungal colonization occurred in 9.8% in the 6-mg group and 7.7% in the 3-mg group, as compared with 29.2% in the placebo group (P<0.001 for both fluconazole groups vs. the placebo group). The incidence of invasive fungal infection was 2.7% in the 6-mg group and 3.8% in the 3-mg group, as compared with 13.2% in the placebo group (P=0.005 for the 6-mg group and P=0.02 for the 3-mg group vs. the placebo group). The use of fluconazole did not modify the relationship between colonization and the subsequent development of invasive fungal infection. Overall mortality was similar among groups, as was the incidence of cholestasis. No evidence for the emergence of resistant candida species was observed, but the study did not have substantial power to detect such an effect.
Conclusions Prophylactic fluconazole reduces the incidence of colonization and invasive candida infection in neonates weighing less than 1500 g at birth. The benefit of treating candida colonization is unclear. (Current Controlled Trials number, ISRCTN85753869
[controlled-trials.com]
).
Very-low-birth-weight neonates are at risk for invasive fungal infection because of their immature immune system and the invasive supportive care they require.9,10,11,12,13,14 Prophylaxis with antifungal drugs is established in certain select high-risk patients, such as adult and pediatric patients with hematologic cancers and immunity defects.15,16,17,18,19 Among such patients, prophylactic fluconazole has decreased candida-related complications.20,21 After a single-center, randomized trial suggested similar efficacy in high-risk, extremely-low-birth-weight neonates (those with birth weights below 1000 g),22 some NICUs began using fluconazole for routine prophylaxis.23,24,25,26 Even so, expert opinion27,28,29,30 and Cochrane reviews31,32 do not recommend this practice because of the paucity of data regarding safety and resistance and the absence of adequately powered, multicenter trials. This study presents the results of such a multicenter, prospective, randomized, double-blind, placebo-controlled evaluation of fluconazole prophylaxis in very-low-birth-weight infants in the NICU.
Methods
Patients
From May 1, 2004, to July 31, 2005, we enrolled neonates at eight tertiary Italian NICUs in the study, which was approved by the ethics committee at each hospital. Parents or guardians provided written informed consent. Pfizer Italia supported the study with a grant and supplied both fluconazole and placebo; the sponsor was not involved in the study design, in the enrollment of patients, or in the collection, analysis, interpretation of the data or preparation of the manuscript. All authors vouch for the completeness and accuracy of the data presented.
The primary objective was to evaluate the effectiveness of fluconazole at doses of 3 mg and 6 mg per kilogram of body weight in the prevention of candida colonization and infection. Secondary objectives were assessment of the incidence of gram-positive and gram-negative sepsis, necrotizing enterocolitis, ligation of patent ductus arteriosus, threshold retinopathy of prematurity requiring surgery, severe intraventricular hemorrhage, bronchopulmonary dysplasia, and alteration of liver function at baseline and at the end of prophylaxis.
Very-low-birth-weight neonates who were admitted to the NICU before day 3 of life were eligible for enrollment. Exclusion criteria were a lack of parental consent and liver failure (levels of aspartate aminotransferase and alanine aminotransferase that were three times the upper limit of the range of normal values). The pharmacy at each center used computer-generated randomization lists to form three groups in a 1:1:1 ratio and prepared the daily drug doses. Infants received either 6 mg or 3 mg of fluconazole (Diflucan, Pfizer Italia) per kilogram of body weight every third day for the first 2 weeks and then every other day; the placebo group received 1 ml of normal saline on the same schedule.20,21 Extremely-low-birth-weight infants received prophylaxis for 6 weeks; neonates weighing 1000 to 1500 g received prophylaxis for 4 weeks,20,21,25 unless they were discharged earlier or required systemic antifungal therapy for proven or presumed invasive fungal infection. Administration of the study drug began on day 3 with one daily dose intravenously, if a catheter was present, or through an orogastric tube.
Weekly surveillance of liver function (levels of serum aspartate aminotransferase, alanine aminotransferase, 
-glutamyltransferase, and bilirubin) was performed for the duration of administration of the study drug. Drug interactions with fluconazole were screened. Invasive fungal infection was defined as a positive culture from blood (a peripheral site), urine (collected by sterile suprapubic puncture or bladder catheterization, with growth of 
10,000 organisms per milliliter), or cerebrospinal fluid. Diagnosis relied on guidelines from international consensus documents33,34 and recommendations of the Italian Neonatology Society's Fungal Infections Task Force.35 Invasive fungal infections were treated with intravenous liposomal amphotericin B at the recommended doses. Investigators remained unaware of study-group assignments during such treatment.
Fungal Isolation and Identification
The following cultures were obtained: at baseline and then weekly, we performed surveillance cultures of the ear canal (at birth only), stool, gastric aspirate, and nasopharynx or endotracheal secretions while infants were receiving fluconazole or placebo; cultures were obtained from surgical devices after removal; and clinical cultures were obtained from sites indicated by the physician (e.g., skin and respiratory secretions). Baseline colonization was defined as the isolation of fungi from the ear canal at birth or from any site during days 1 and 2 of life.
Specimens were incubated on chromogen culture plates (Albicans ID, BioMérieux) to identify Candida albicans colonies as blue stains after 48 hours at 37°C. Colonies were speciated through a miniaturized system of biochemical tests (Vitec Yeast, BioMérieux). Isolates were tested for sensitivity to fluconazole with standardized microbroth dilution assays (ATB-Fungus-2-Int, BioMérieux) according to recommendations of the National Committee for Clinical Laboratory Standards (NCCLS).36 The interpretative breakpoint of fluconazole resistance was defined as at least 64 µg per milliliter. A breakpoint of 16 to 32 µg per milliliter was considered to be indicative of dose-responsive susceptibility.36
Statistical Analysis
The following variables were analyzed: incidence of colonization (at least one site), invasive fungal infection, death from all causes before discharge from the hospital, candida-related deaths (death within 3 days after the last positive culture from any site in the absence of other causes or isolation of candida species at autopsy), the presence of natively fluconazole-resistant species, secondary outcomes (including those potentially related to fluconazole), the rate of progression of colonization to invasive fungal infection, and patterns of the sensitivity of isolates to fluconazole. Infants receiving either 6 mg or 3 mg of fluconazole per kilogram were compared separately with those receiving placebo. Since there were no significant differences between the 6-mg group and the 3-mg group, a post hoc analysis was performed comparing the placebo group with both fluconazole groups combined. Proportions and continuous variables were compared with the use of Fisher's exact two-tailed test and the t-test, respectively. Risk ratios and 95% confidence intervals were calculated to compare cumulative between-group incidences with the use of Stata software. A multivariate logistic-regression analysis was performed with adjustment for important risk factors possibly associated with invasive fungal infection. The Wald test was used to assess the significance of the estimated coefficients.
The number of patients needed for each group was estimated in 82 infants for colonization on the basis of a two-sided type I error rate of 0.05 or less and a power of 90% to detect an absolute difference of at least 66% (a decrease from 30% to 10%) in the cumulative incidence of fungal colonization between infants in the fluconazole groups and those in the placebo group, given a pretrial incidence of 30%. The number needed for each group was also estimated in 89 infants for invasive fungal infection on the basis of a two-sided type I error rate of 0.05 or less and a power of 80% to detect an absolute difference of at least 80% (a decrease from 15% to 3%) in the cumulative incidence of invasive fungal infection between infants in the fluconazole groups and those in the placebo group, given a pretrial incidence of 15%. A total of 118 infants would have been needed to reach a power of 90%. Given the low incidence of invasive fungal infection in the two fluconazole groups, the study was underpowered to detect significant differences for this outcome between these groups. Assuming an incidence of 4% in the 3-mg group, 1141 infants would have been needed to reach a power of 80% to detect an absolute difference of 50%, and 424 infants would have been needed for a difference of 75%. Power calculations were performed with the use of S-plus software, version 2000 (MathSoft).
Results
Patients
A total of 363 very-low-birth-weight neonates survived for at least 3 days and were potentially eligible for the study. Of these infants, 27 were ineligible owing to liver failure at baseline (3 patients), a lack of consent (12 patients), and missing or incorrect randomization (12 patients). Another 14 infants were removed from the analysis owing to incomplete data (3 in the 6-mg group, 3 in the 3-mg group, and 4 in the placebo group), incorrect drug administration (2 in the 6-mg group and 1 in the 3-mg group), or accidental administration of five doses of fluconazole instead of placebo (1 in the placebo group). A total of 322 neonates were randomly assigned, 112 to the 6-mg group, 104 to the 3-mg group, and 106 to the placebo group (Figure 1).
|
|
|
Data regarding fungal colonization, invasive fungal infection, and distribution of fungal species and colonization sites and infection are shown in Table 3 and Table 4. Colonization occurred less frequently in the 6-mg group (9.8%) and the 3-mg group (7.7%) than in the placebo group (29.2%; P<0.001 for both comparisons). Also less frequent was invasive fungal infection (2.7% in the 6-mg group and 3.8% in the 3-mg group, as compared with 13.2% in the placebo group; P=0.005 and P=0.02, respectively). The 21 episodes of invasive fungal infection were caused by C. albicans (16 patients), C. parapsilosis (2), C. glabrata (2), C. tropicalis (1), and C. guilliermondii (1). One neonate was infected with both C. parapsilosis and C. glabrata. Invasive fungal infections occurred in 2 of 7 neonates with baseline colonization (28.6%) in the fluconazole groups combined and in 4 of 14 neonates (28.6%) in the placebo group.
|
|
Mortality
Overall mortality was similar in the three groups (8.0% in the 6-mg group [P=0.81] and 8.7% in the 3-mg group [P=1.0], as compared with 9.4% in the placebo group). In the fluconazole groups, no deaths were attributable to candida infection, as compared with two deaths (1.9%) in the placebo group (P=0.23 for the 6-mg group and P=0.50 for the 3-mg group) (Table 3).
Isolates Natively Resistant to Fluconazole
There were no significant between-group differences in the incidence of colonization and infection with C. krusei, C. glabrata, and C. guilliermondii. Overall, three isolates from these species were seen in both of the fluconazole groups (C. krusei, C. glabrata, and C. guilliermondii) and two in the placebo group (both C. glabrata). C. glabrata caused two nonfatal invasive fungal infections, one each in the 6-mg group and the placebo group.
Minimal Inhibitory Concentration
Sensitivity to fluconazole as measured by the minimal inhibitory concentration (MIC) required to inhibit the growth of 90% of the isolates (MIC90) did not vary during the study period, and all isolates remained sensitive to fluconazole (C. albicans, 0.125 to 2.0; and C. parapsilosis, 0.25 to 2.0). MICs of fluconazole in the isolated strains did not increase in any of the neonates at any center during the study period. Satisfactory sensitivity (MIC90: 1 [colonizing isolates], 2.0 and 8 [infecting isolates]) continued for C. glabrata. All infecting and colonizing isolates were sensitive to amphotericin B and flucytosine (MIC90, 0.125 to 1.0 and 0.125 to 0.25, respectively).
Secondary Outcomes
There were no significant differences in secondary outcomes (Table 5). No serious adverse events or fluconazole-related toxic effects were recorded. Drug administration was not discontinued because of presumed adverse events, intolerance, or potentially dangerous interactions with other drugs. At 4 weeks of age, infants who received fluconazole had increased levels of aspartate aminotransferase and alanine aminotransferase. For aspartate aminotransferase, the mean (±SD) levels were 16.8±11.0 U per liter in the fluconazole groups combined and 13.1±10.0 U per liter in the placebo group (P=0.004); for alanine aminotransferase, the mean levels were 22.8±16.0 U per liter in the fluconazole groups combined and 19.5±11.0 U per liter in the placebo group (P=0.06). These modifications were not observed at 6 weeks or at hospital discharge. Elevations in levels of more than two times the range of normal in serum aspartate aminotransferase and alanine aminotransferase were recorded in 4 neonates who received fluconazole and in no neonates in the placebo group (P=0.31 for both comparisons); such increases in -glutamyltransferase levels occurred in 13 neonates who received fluconazole and in 6 who received placebo (P=1.0). At 6 weeks of age, three infants in the fluconazole groups (none with previous alterations in levels) maintained serum levels that were slightly above normal (one for aspartate aminotransferase and two for -glutamyltransferase), as compared with two infants (both for -glutamyltransferase) in the placebo group. These abnormalities were transient and absent at discharge. No infants reached levels of more than three times the normal range of values or had clinical signs of hepatotoxicity or cholestasis, and none required treatment for cholestasis. None of the neonates required phototherapy to treat hyperbilirubinemia.
|
This prospective, randomized, multicenter study showed that prophylactic fluconazole prevents colonization and infection by candida species in very-low-birth-weight infants. Our findings are similar to those from single-NICU studies21,22,23,24,25,26 and to those observed in immunocompromised adults and children.15,16,17,18,19 Colonization is considered a major risk factor for invasive fungal infection in the preterm neonate.10,11,14,37 In our study, fluconazole was effective in preventing rather than treating colonization.
Exposure of newborns to fungi is a complex problem and results from both horizontal and vertical transmission. Treatment of maternal vaginal candidiasis and good hand hygiene of health care workers should reduce the risk of "new entry" candida in the NICU. Breaking the horizontal NICU transmission cycle of hand colonization from caregivers to neonates and vice versa by fluconazole prophylaxis should reduce the risk that preterm infants pose to one another.
Rates of candidemia in the NICU vary greatly among institutions. NICU-related factors (e.g., the average use of broad-spectrum antibiotic per infant5) may be implicated, in addition to the well-known preterm-related factors. Decreasing the incidence of infection would benefit a NICU's present and future preterm infants, a factor that may lead to the adoption of intermittent prophylactic regimens with the targeting of prophylaxis to ever-more-selected subgroups of high-risk neonates.9,37 Prophylaxis should be adapted to the infection rate in each NICU to optimize the number needed to treat, in line with guidelines recommending antifungal prophylaxis in "carefully selected patients in units with high rates of invasive candidiasis."38 In our study, the number needed to treat was eight (five among extremely-low-birth-weight infants) to prevent invasive fungal infection.
Two randomized, placebo-controlled studies of prophylactic fluconazole have been conducted among preterm infants, one in which 3 mg per kilogram of the drug was administered and the other in which 6 mg per kilogram was administered.21,22 Our study was underpowered to detect differences in the outcome of invasive fungal infection between the two groups, owing to the low incidence of such infection. Power analysis indicated that approximately 1100 patients would be required for this evaluation.
A post hoc analysis was performed to evaluate the effect of at least 3 mg per kilogram of fluconazole by comparing both the 6-mg group and the 3-mg group with the placebo group. This analysis requires cautious interpretation. The higher dose could be more effective against strains that are dose-susceptible to fluconazole but could also be more toxic because of increased drug exposure. In this study, at 4 weeks treated infants showed higher levels of alanine aminotransferase and significantly higher levels of aspartate aminotransferase, although the levels remained within the normal reference range. However, at 6 weeks and at discharge, these modifications were not observed, and all treated neonates were discharged without abnormalities possibly caused by fluconazole. Mild and transient increases of liver enzymes, without clinical implications, are described in infants receiving fluconazole.21,22,39
Another concern regarding fluconazole prophylaxis is the emergence of fungal resistance.18 In this small, brief study, sensitivity of susceptible strains remained unchanged. (MIC90 remained below the cutoff of NCCLS guidelines.) Another concern is the selection of natively resistant candida species (e.g., C. glabrata and C. krusei), which are occasionally associated with fluconazole exposure.40,41 Consistent with other, short-term studies,19,42 there was no significant change in candida ecology among the three groups, and MICs for C. glabrata remained in the susceptible range during the study period. Nevertheless, our study was underpowered to detect a change in fungal ecology because of its short duration. Thus, we were unable to detect shifts in candida species or the establishment of acquired resistance mutations. Long-term surveillance of fungal ecology will be important for NICUs that adopt a prophylaxis strategy.
In this multicenter study, fluconazole prophylaxis reduced fungal colonization and infection in preterm neonates. Future studies should further refine the identification of neonates who are at highest risk for infection for whom prophylaxis would be most optimally suited. Such trials should also help to define the lowest effective dose for such prophylaxis and the possible long-term emergence of fungal resistance.
Supported by Pfizer Italia.
Some data from this study were presented in preliminary form at the 46th annual Interscience Conference on Antimicrobial Agents and Chemotherapy, San Francisco, September 27–30, 2006.
No potential conflict of interest relevant to this article was reported.
Source Information
From Sant'Anna Hospital, Turin (P.M., C.G., D.F., G.G.); Policlinico Umberto I, Rome (I.S., R.P.); Mangiagalli Hospital IRCCS, University of Milan, Milan (L.P., F. Mosca); San Matteo Hospital, Pavia (L.D., M.S.); Arcispedale, Reggio Emilia (C.M.); Fatebenefratelli Hospital, Benevento (G.V.); University of Bologna, Bologna (E.T.); University of Messina, Messina (G.C.); and University of Turin, Turin (R.A., M. Maule, F. Merletti, M. Mostert) — all in Italy.
Address reprint requests to Dr. Manzoni at the Neonatology and Hospital Neonatal Intensive Care Unit, Sant'Anna Hospital, Corso Spezia 60, 10126 Turin, Italy, or at paolomanzoni{at}hotmail.com.
References
| |||||||||||||||||||||||||||||||||||||||||||||||||||||
Related Letters:
Prophylactic Fluconazole in Preterm Neonates
Vain N. E., Vazquez L. N., Procianoy R. S., Silveira R. C., Manzoni P., Mostert M., Farina D.
Extract |
Full Text |
PDF
N Engl J Med 2007;
357:1348-1349, Sep 27, 2007.
Correspondence
This article has been cited by other articles:
HOME | SUBSCRIBE | SEARCH | CURRENT ISSUE | PAST ISSUES | COLLECTIONS | PRIVACY | HELP | beta.nejm.org Comments and questions? Please contact us. The New England Journal of Medicine is owned, published, and copyrighted © 2008 Massachusetts Medical Society. All rights reserved. |
Previous
