Strategies for the Prevention of Neonatal Candidiasis

Invasive fungal infections represent the third-leading cause of late-onset sepsis in very-lowbirth-weight infants (VLBWI) and have a high rate of infection-associated mortality. The infantsat high risk for fungal sepsis are VLBWI with presence of additional risk factors that contributeto increased colonization and concentration of fungal organisms. Colonization withCandidaspp. in neonates is secondary to either maternal vertical transmission or nosocomial acquisition in the nursery. Multiple sites may become colonized and a direct correlation betweenfungal colonization and subsequent progression to invasive candidemia was determined.Randomized, single and multiple-center, placebo-controlled trials found intravenous fluconazole prophylaxis to be effective in decreasing fungal colonization and sepsis for at-risk preterminfants<1500 g birth weight. The prophylactic use of fluconazole was found to be safe with nosignificant development of fungal resistance. Fluconazole prophylaxis administered to pretermneonates with birth weight<1000 g and/or 27 weeks’ gestation or less has the potential ofreducing and potentially eliminating invasive fungal infections andCandida-related mortality.

Copyright©2012, Taiwan Pediatric Association. Published by Elsevier Taiwan LLC. All rightsreserved.

1. Introduction

Invasive fungal infections (IFI) represent a leading cause ofsepsis in very-low-birth-weight (VLBW) infants and result inhigh rates of morbidity and mortality.1e4Among many risk factors responsible for development of IFI, previousmucosal and skin colonization are of primary importance.Colonization withCandidain neonates is secondary toeither maternal transmission or to nosocomial acquisition inthe neonatal intensive care unit (NICU).

Because current therapies for systemic fungal diseasesare not universally successful and morbidity remains high,efforts may be better focused on preventing invasivedisease by interrupting the process of colonization andsubsequently preventing the development of serious fungal infections. In-depth knowledge of theCandidaspp. colonization patterns, timing, anatomic sites involved andspecific species distribution is important for both theclinician and scientist. The preventive strategy using fluconazole for the prevention of colonization and systemicinfections withCandidaspp. in both extremely low-birthweight (ELBW) and VLBW infants was found to be effective in several studies.

We present in this review an up-to-date summary of theinformation accumulated during recent years regarding theepidemiology and risk factors for IFI in the neonatal population, concentrating on the presently available strategiesfor decreasing fungal colonization, IFI and mortality in atrisk preterm infants.

2. Neonatal Invasive Fungal Infections

Invasive fungal infections (IFI) represent the third-leading(12.2%) cause of late-onset sepsis in VLBW (birth weight<1500 g) infants. Their estimated incidence is 1.6e3% inVLBW infants and up to 15e20% in ELBW neonates withaCandida-attributable mortality of 25e55% and presenceof associated neurodevelopmental impairment at 18 yearsold in 57% of survivors.1,2,11,12The incidence of IFI variesconsiderably between reporting NICUs, being related,mainly, to the number of ELBW infants<24e25 weeks ofage taken care at a specific NICU.

Candida albicansis the most prevalent fungal pathogenin neonatal disease, but the incidence of infection causedby otherCandidaspecies, particularlyCandida parapsilosisandCandida glabratahas also increased dramaticallyduring the past two decades.3,4,13,14Furthermore,C.glabrataandCandida krusei, another increasingly recovered representative ofCandidaspp., harbor innate resistance to the azole class of antifungal agents and mayrepresent therapeutic challenges in IFI.

The highest incidence of IFI in VLBW infants is recordedbetween the second and sixth weeks after birth. NeonatalIFI are not confined to hematogenous sepsis, but mayinvolve the central nervous system (meningitis and brainabscesses), urinary tract, soft and deep tissues, endocarditis, endophtalmitis, hepatitis and pneumonitis.15,16While the clinical picture is similar to that of late-onsetGram-negative bacterial infections, the blood cultures ofCandidaspp. may be negative even in cases of deep-organinvolvement and confirmedCandidameningitis.

Risk factors for IFI include prematurity, VLBW, use ofcentral venous lines, intubation, parenteral nutrition,broad-spectrum antibiotics administration (in particularthird-generation cephalosporins), prolonged hospitalization, abdominal surgery, exposure to an H2blocker, andcolonization withCandidaspp.12,17e19Of these manyfactors, mucosal and skin colonization are of primaryimportance. In a single-center study enrolling 146 VLBWinfants, fungal colonization occurred in 26.7% and onethird of those colonized developed IFI.5In 1991e1992,El-Mohandes20studied 82 (61 infants<1500 g) consecutiveNICU admissions and found a colonization rate withCandidaspp. of 19% with stools more frequently culturepositive than skin or gastric aspirates. Five of the studypatients developed fungal sepsis; the authors found that gastrointestinal colonization was strongly associated withdevelopment ofC. parapsilosissepsis. Rowen21reportedthat endotracheal colonization withCandidaenhances riskof systemic candidiasis in VLBW infants. Huang22founda colonization rate of 22% among 116 2-week-old VLBWinfants (rectal colonization present in 76% of colonizedinfants) with fungemia developing in three infants;however, in only one-third of infants could an associationbetween colonization and fungemia be found). Manzoni19found colonization and IFI rates of 32.1% and 8.1%,respectively, among the 689 VLBW infants admitted during1998e2005 at an NICU in Torino, Italy. After logisticregression, colonization of central venous catheters andcolonization in multiple sites were found to be independentrisk factors and predictors of progression to fungal sepsis.

Regarding the antifungal treatment of IFI, amphotericinB deoxycholate, amphotericin B lipid preparations andfluconazole are the most commonly used antifungal inneonates with invasive infections, as well as in olderpediatric patients; however, the development of newergeneration azoles (voriconazole, posaconazole) and echinocandins (mainly caspofungin), still undergoing extensivestudy on dosing and safety, may ultimately expand thetherapeutic options for single-agent and combinationtherapy for neonatal candidiasis.

3. Neonatal Colonization withCandidaspp.

Colonization withCandidaspp. in neonates may besecondary to either vertical transmission from the motheror horizontal (nosocomial acquisition) transmission in thenursery or NICU. Studies evaluating gastrointestinal colonization have shown that 4.8-10% of neonates will harbora strain ofCandidaon admission to the nursery.At NICUs,around 50% of the admitted infants may be colonized by theend of the first week of life and up to 64% could becolonized by 4 weeks old.5,29The majority of infants arecolonized withC. albicans, but a large percentage carryC.parapsilosis6,29(a pattern different from that encounteredin adults). An increase in the incidence ofC. parapsilosisneonatal infections has been reported during the past twodecades in parallel to a decrease in the representation ofC.albicansstrains among the fungal isolates recovered atvarious NICUs.

The Candidastrains colonizing the infant are mostfrequently acquired by transmission from the maternalgenitourinary or gastrointestinal tract during the perinatalperiod. Using molecular typing techniques, vertical transmission ofC. albicanswas documented in 33% of a groupof premature infants.29While vertical transmission iscommon, horizontal acquisition ofCandida(from careproviders or contaminated infusions) does occur in the NICUand is probably the most common mode of transmission forC. parapsilosis.6,29Candidaspp. were isolated from thehands of 29% (859 of 2989 sampled) of health workers ina multicenter study.6,18Candida parapsilosiswas isolatedfrom 19% andC. albicansfrom 5% of the hand cultures ofthe healthcare personnel sampled.

In a multicenter study of neonatal bloodstream infections withCandidaspp., molecular typing demonstratedthat the same fungal strains were present in the bloodstream and the gastrointestinal tract of the sick infant, illustratingthe relationship between gastrointestinal tract colonizationand development of IFI.17In this study, infants>1000 gwere more likely to be colonized withC. albicansthanC.parapsilosis, while infants<1000 g showed no differencesbetween the rate of colonization ofC. albicansversusC.parapsilosis. The same study found by univariate analysisthat rectal colonization by fungi was a risk factor for candidemia. Recently, in a single-center study from Virginia,USA, in which 50 high-risk VLBW infants were enrolled,Candidawas isolated from surveillance cultures in 31 (62%)infants and the colonization was inversely proportional withthe gestational age. Fungal colonization of the skin andgastrointestinal tract was reported to occur before that ofthe respiratory tract andC. albicanswas found more likelythanC. parapsilosisto colonize multiple sites.30Vendetuolli et al31obtainedCandidasurveillance cultures from51 NICU patients during 2005e2007 and found that 16infants (15 of them<27 weeks gestational age) with positive cultures developed subsequent IFI compared with the34 others who did not have fungal colonization and did notdevelop IFI. The positive surveillance cultures led to anearly initiation of antifungal treatment and only one infantdied ofCandidainfection.

4. Prevention of Neonatal Fungal Colonizationand Invasive Disease

4.1. Fluconazole prophylaxis

Fluconazole is an azole with a long half-life, very goodcerebrospinal penetration, low protein binding and goodtissue, body fluids and mucocutaneous areas levels (higherthan plasma ones), characteristics that allow long dosingintervals, excellent tissue penetration and easy elimination.32,33At the present time, fluconazole is considered tobe the main alternative to amphotericin B in the treatmentof neonatal candidiasis and has been fairly well characterized in this population.

In VLBW infants, prevention of fungal colonization andIFI by fluconazole prophylaxis has been shown to beeffective in 13 studies (nine retrospective and fourprospective randomized) during the past decade.36e48Thefour randomized, placebo-controlled studies completedduring 2000-2007 are presented in detail inTable 1.36e38,44In 2001, Kicklighter et al and Kaufman et al reported thefirst two randomized, double-blind, placebo-controlledtrials of administration of fluconazole prophylaxis inextremely VLBW infants in order to preventCandidacolonization and IFI.36,37They used protocols which treated(starting in the first 72 hours of life) VLBW infants witha birth weight of<1500 g and 1000 g, respectively, with anintravenous/oral or intravenous-only fluconazole dose of 3or 6 mg/kg in different daily schedules till day 28e42.36,37The authors reported significant decreases inCandidacolonization and IFI rates, without changes in the sensitivities of the fungal isolates to fluconazole during thestudy period and without documentation of adverse effectsof the fluconazole therapy. In 2005, Kaufman et al38extended their previous study and administered fluconazole twice weekly (instead of daily treatment) at a dose of 3 mg/kg/day and demonstrated equivalence between thedaily and twice-weekly dosages in respect of colonizationand IFI occurrence reductions between the two schedules.A recent multicenter, prospective, randomized clinicaltrial has added important data on the safety and efficacyof this practice.44Manzoni et al enrolled during 2004e2005in Italy all neonates<1500 g at birth from eight tertiaryNICUs (322 infants) and randomly assigned them to receiveeither fluconazole (intravenously or through an orogastrictube, 3 or 6 mg/kg/day every third day for the first2 weeks and then every day) or placebo from birth untilDay 30 of life (Day 45 for those<1000 g at birth). Theauthors reported significant reductions in fungal colonization and IFI for both treatment groups, but could not provean effect of the therapy on the relationship betweencolonization and the subsequent development of IFI. Theoverall mortality and the incidence of cholestasis were notdifferent between the treatment and placebo groups andno evidence for emergence of resistantCandidaspp. wasobserved.

Overall, 400 and 1522 evaluable VLBW neonates weretreated with various fluconazole prophylaxis regimes in thefour prospective randomized and the nine retrospectivestudies mentioned, respectively. Summarizing all theavailable data, fluconazole prophylaxis was found to reduceIFI by>85% in all infants<1500 g and by 91% in high-riskinfants<1000 g. The all-cause mortality was 11% for thefluconazole-treated group compared with 16.3% forcontrols while nearly eliminating (96% reduction)Candidaerelated mortality, with no documentation of significantadverse effects.8e10,23In a recent study completed onpatients who were enrolled in the initial randomized,controlled study on fluconazole prophylaxis of IFI performed by Kaufmann et al, there were no long-termadverse effects on neurodevelopmental parameters andquality of life at 8e10 years old.

4.2. Fluconazole resistance issues

The use of fluconazole prophylaxis in NICU raised concernabout the development of resistance, including possibleselection ofCandidaspp. with acquired resistance or ofnatively fluconazole-resistant strains likeC. glabrataandC.kruseireplacing fluconazole-susceptible species. Epidemiologic data from 127 adult institutions in the USA showedthat the incidence of fluconazole-resistantCandidaspp. isrising, prompting concerns that that such resistancepatterns may soon develop in the neonatal population.50,51In a pediatric study reporting on the antifungal susceptibility of 179 clinicalCandidaspp. isolated from four children’s hospitals in the USA (patient age not detailed),52C. albicans,C. parapsilosis,C. glabrata,Candida tropicalis,Candida lusitaniaeandC. kruseiwere identified from43%, 32%, 8%, 8%, 8% and 1%, respectively. Fluconazoleresistance was recorded amongC. glabrata,C. tropicalisandC. kruseiisolates (7%, 36% and 100%, respectively).AmongC. parapsilosisisolates, 18% were resistant toamphotericin B. Caspofungin was the most active agentin vitroagainst allCandidaspp. (100% for all of them, withthe exception ofC. parapsilosiswhere a 98% susceptibilityrate was reported).

No increase in the number of resistant organisms havebeen observed in five major neonatal prophylaxisstudies36e38,41,44where surveillance cultures for colonization (at baseline and weekly during the study period) wereperformed. There was no significant increase in minimuminhibitory concentration to fluconazole of the fungal isolatesduring the prophylaxis period (between 4 and 6 weeks) orthe entire study period (2-3 years). Furthermore, there wasno emergence or increase in the incidence of colonization orinfection fromC. glabrataorC. kruseireported in anystudies, including a recent single-center analysis of 10years.However, there is general agreement that eventargeted fluconazole prophylaxis in neonates will still drivethe emergence of azole resistance.54The follow-up periods(up to 30 months) are probably insufficient to detect clinically significant changes in the resistance profile of fungalisolates. Antifungal prophylaxis may take years after theintroduction of fluconazole prophylaxis to become established in NICUs.55Sarvikiki et al55recently describes theirexperience with fluconazole prophylaxis over a 12-year period; they reported noC. glabrataorC. krusei-IFI andonly two cases of resistantC. parapsilosis-IFI. It was interesting to remark that when fluconazole prophylaxis wasadministered daily at 6e12 mg/kg/kg in infants<1000 g and<30 weeks gestational age between 1991-2000, no fungalresistance was detected. However, when the use of fluconazole was broadened to the entire NICU in 2000, theemergence of resistance was noted in two patients whowere infected with an endemic strain that had previouslybeen susceptible.55In a case report from 2007, a prematureinfant with persistentC. albicanscandidemia was successfully treated with caspofungin after the isolate developedresistance to fluconazole and voriconazole during the initialtreatment with fluconazole.

4.3. Nystatin prophylaxis

Oral nystatin was the first antifungal drug studied for theprophylaxis of IFI in preterm infants. Previous small studiesdemonstrated a significant reduction in fungal bloodstreaminfection or meningitis when the prophylaxis was administered early after birth.One recent randomized,controlled trial compared oral fluconazole to nystatinprophylaxis initiated during the first 7 days of life in VLWBinfants<1500 g.60IFI were diagnosed in 2.35 (5.3%) of thefluconazole-treated premature infants compared with 6/42(14.3%) of the nystatin-treated infants. There was nomortality in the fluconazole-treated patients comparedwith 6/42 (7.5%) in the nystatin-treated infants (p=0.03),when the majority of fatalities were caused by bowelperforation and necrotizing enterocolitis. In fact, goodsafety data are lacking for the use of nystatin in neonates;in addition, the drug can be administered only enterally(and therefore not in patients with necrotizing enterocolitisor other gastrointestinal pathologies) and, subsequently,will affect only the gastrointestinal fungal colonization.

4.4. Present recommendations

According to the available findings, it is accepted todaythat the decisions regarding fluconazole prophylaxis should be made according to its effect on the colonization and IFIincidence, mortality and neurodevelopmental impairmentfollowing IFI. Individual NICUs have to examine their incidence, neurodevelopmental impairment outcomes andmortality with IFI by various gestational ages and decide ontheir standard prophylaxis practices.

Criticism and concern regarding the routine use of fluconazole in NICUs is related, mainly, to the lack of satisfactory data about the true burden of IFI in preterminfants, leading to a wide range of incidence rates, withmany NICUs reporting rates of 2e3% or less, which, theoretically, would not justify the use of fluconazole. In fact,most of the studies showing the efficacy of fluconazoleprophylaxis were published from centers having high ratesof IFI.36e38,41,44Many studies report only on the rates ofbloodstream infections and data on specific local antibioticuse and feeding practices, antifungal prophylaxis policieswith drugs other than fluconazole (mostly nystatin), infection control measures adopted, resuscitation practices ornumbers of patients admitted with complicated abdominalsurgery procedures are incomplete. Additional concerns arerelated, of course, to the effects of widespread use offluconazole particularly on specificCandidasubspecies likeC. glabrataandC. krusei(and evenC. parapsilosis) knownto have intrinsic resistance to this drug. Therefore, futurestudies will have to provide additional information on someunsolved issues like the effect of fluconazole on overallmortality, efficacy of prophylaxis for shorter periods oftime (like during the antibiotic therapy only), possible livertoxic effects, interaction with other drugs and identification of premature infant subpopulations with the mostbeneficial outcome, while continuously and carefullymonitoring the possible development of resistance with thebroad use of the drug.

Meanwhile, the most beneficial approach seems to be bytargeting fluconazole prophylaxis to all infants<1000 g or ≤27 weeks gestational age admitted at NICUs withmoderate-to-high rates of IFI (5e10%, for example) whilethey require intravenous access, starting on Day 1 of lifeand continuing till 6 weeks of life. This approach was found,in the USA, to prevent 2000-3000 IFI, 200-400Candidarelated deaths and 900e1200 infants developing neurodevelopmental impairment (of a total of 30,000 preterminfants<1000 g or≤27 weeks’ gestation born each year).The infants with a birth weight of 1000e1500 g may also becandidates for such an approach or a more targeted one(using prophylaxis only from those with central venouscatheters or those receiving antibiotics and as a function ofthe overall IFI rates at the specific NICU).

How are such recommendations accepted and implemented? In a European survey on the prophylactic use offluconazole in NICUs, 193 structured questionnaires on thepractice of fluconazole prophylaxis, completed by 223neonatologists from 28 countries, were analyzed.61Use ofprophylaxis was reported by 55% of the interviewedneonatologists, while fluconazole represented the mostcommonly used prophylactic agent (92%). The dose mostcommonly used was 3 mg/kg/day (88% of the physiciansusing the drug) with an administration interval of 72 hoursin 52% of them. The decision not to use prophylaxis wasmainly based on local data on the incidence of IFI, concernabout the possible increase in antifungal resistance and lack of specific guidelines by the pediatric and neonatologysocieties.

5. Conclusions

Fluconazole prophylaxis was shown, in carefully performed,randomized, placebo-controlled studies, to reduceCandidaspp. colonization and IFI rates without emergence of drugresistance and without adverse effects during extendedperiods of time. Fluconazole prophylaxis administered topreterm neonates with birth weight<1000 g and/or 27weeks’ gestation or less has the potential of reducing andpotentially eliminating IFI andCandida-related mortality.