Prophylactic oral/topical non-absorbed antifungal agents to prevent invasive fungal infection in very low birth weight infants

  • Review
  • Intervention

Authors


Abstract

Background

Invasive fungal infection is an important cause of mortality and morbidity in very preterm or very low birth weight infants. Uncertainty exists about the effect of prophylactic oral/topical non-absorbed antifungals to reduce mucocutaneous colonisation and so limit the risk of invasive fungal infection in this population.

Objectives

To assess the effect of prophylactic oral/topical non-absorbed antifungal therapy on the incidence of invasive fungal infection, mortality and morbidity in very preterm or very low birth weight infants.

Search methods

We used the standard search strategy of the Cochrane Neonatal Review Group. This included searches of the Cochrane Central Register of Controlled Trials (CENTRAL: The Cochrane Library, 2012, Issue 3), MEDLINE, EMBASE, and CINAHL (to August 2012), conference proceedings, and previous reviews.

Selection criteria

Randomised controlled trials or quasi-randomised controlled trials that compared the effect of prophylactic oral/topical non-absorbed antifungal therapy versus placebo or no drug or another antifungal agent or dose regimen in very preterm or very low birth weight infants.

Data collection and analysis

We extracted data using the standard methods of the Cochrane Neonatal Review Group with separate evaluation of trial quality and data extraction by two review authors.

Main results

Four trials, in which a total of 1800 infants participated, compared oral/topical non-absorbed antifungal prophylaxis (nystatin or miconazole) with placebo or no drug. These trials had various methodological weaknesses including quasi-randomisation, lack of allocation concealment, and lack of blinding of intervention and outcomes assessment. The incidence of invasive fungal infection was very high in the control groups of three of these trials. Meta-analysis found a statistically significant reduction in the incidence of invasive fungal infection [typical risk ratio 0.20 (95% confidence interval 0.14 to 0.27); risk difference -0.18 (-0.21 to -0.16)] but substantial statistical heterogeneity was present. We did not find a statistically significant effect on mortality [typical risk ratio 0.87 (0.72 to 1.05); risk difference -0.03 (-0.06 to 0.01)]. None of the trials assessed posthospital discharge outcomes.

Two trials (N = 265) assessed the effect of oral/topical non-absorbed versus systemic antifungal prophylaxis. Meta-analyses did not find any statistically significant differences in the incidences of invasive fungal infection or all-cause mortality.

Authors' conclusions

The finding of a reduction in risk of invasive fungal infection in very low birth weight infants treated with oral/topical non-absorbed antifungal prophylaxis should be interpreted cautiously because of methodological weaknesses in the included trials. Further large randomised controlled trials in current neonatal practice settings are needed to resolve this uncertainty. These trials might compare oral/topical non-absorbed antifungal agents with placebo, with each other, or with systemic antifungal agents and should include an assessment of effect on long-term neurodevelopmental outcomes.

Plain language summary

Prophylactic oral/topical non-absorbed antifungal agents to prevent invasive fungal infection in very low birth weight infants

Fungi such as candida (the organism that causes thrush) can cause severe infections in very low birth weight infants (birth weight < 1.5 kg). These infections are often difficult to diagnose and frequently cause death or disability. Therefore, it may be appropriate to attempt to prevent such infections by giving very low birth weight infants antifungal drugs as a routine part of their care. This review assessed specifically the effect of giving infants antifungal drugs that reduce skin and gut carriage of fungi to minimise the chances of a severe infection developing. The trials that were identified suggested that this treatment might reduce severe infection rates but there was no evidence that death rates were reduced. Larger and higher quality trials are needed to resolve this uncertainty.

Background

Description of the condition

Invasive fungal infection accounts for 10% of all cases of late-onset invasive infection in very preterm or very low birth weight (VLBW) infants (Stoll 2002). The reported mortality rates of greater than 25% are higher than those attributed to nosocomial bacterial infection in VLBW infants (Saiman 2000; Makhoul 2002; Stoll 2002; Benjamin 2003). Invasive fungal infection is also associated with short- and long-term morbidity, including adverse neurodevelopmental outcomes (Lee 1998; Friedman 2000; Saiman 2000; Benjamin 2006).

The overall incidence of invasive fungal infection in VLBW infants is between about 1% and 4%, but the risk of infection is inversely related to gestational age and birth weight. In extremely preterm or extremely low birth weight (ELBW) infants, reported incidences are between about 2% and 8%. Much higher incidences, up to 20%, have been reported for infants of birth weight less than 750 grams or gestational age at birth less than 26 weeks (Saiman 2000; Makhoul 2002; Horbar 2002; Karlowicz 2002; Clerihew 2006).

Observational studies suggest that mucocutaneous or tracheal fungal colonisation is a risk factor for invasive infection (Faix 1989; Pappu-Katikaneni 1990; Rowen 1994; Huang 1998). However, multivariate analyses that account for potential confounding variables, particularly ELBW, have not confirmed this association (Saiman 2000). Other putative risk factors for invasive fungal infection in VLBW infants include severity of illness at birth, the use of multiple courses of antibiotics (particularly third-generation cephalosporins), the use of parenteral nutrition, the presence of a central venous catheter, and exposure to histamine receptor subtype 2 antagonists (Rowen 1994; Benjamin 2006; Cotten 2006; Manzoni 2006).

The clinical presentation of invasive fungal infection in VLBW infants is similar to that of bacterial infection and this may cause delays in diagnosis and treatment. The diagnosis may be further delayed due to an inability to recover the organism from microbiological culture of blood, cerebrospinal fluid, or urine. A high index of suspicion and the use of additional laboratory and clinical tests may be needed to confirm the suspected diagnosis (Benjamin 2003).

Description of the intervention

Given the difficulty in establishing an early diagnosis and the high level of associated morbidity and mortality, there is a need to assess the effect of strategies to prevent invasive fungal infection in VLBW infants (Brecht 2009). In addition to generic infection control practices and avoidance of modifiable risk factors, two broad chemoprophylactic strategies are employed in current clinical practice (Burwell 2006; Clerihew 2008; Ganesan 2008; O'Grady 2008; Howell 2009; Kaguelidou 2012):

  • Prophylaxis using systemically-absorbed antifungal drugs that achieve fungicidal concentrations in tissue, blood, cerebrospinal fluid, and urine. Evidence exists that systemic antifungal prophylaxis using fluconazole reduces the incidence of invasive fungal infection, but there is concern about toxicity (Frattarelli 2004) as well as the effect that its widespread use may have on the emergence of antifungal resistance (Brion 2007; Austin 2012)

  • Prophylaxis using oral/topical non-absorbed agents such as nystatin or miconazole. Observational studies have suggested that oral/topical non-absorbed antifungal prophylaxis reduces mucocutaneous fungal colonisation and the risk of invasive infection in VLBW infants (Ganesan 2008; Howell 2009). However, the specific effect of antifungal prophylaxis independently of other confounding interventions and variables is unable to be determined from these studies. Another concern is that hyper-osmolar nystatin preparations may increase the risk of adverse gastrointestinal events in VLBW infants (Ernst 1983;Radmacher 2012).

Why it is important to do this review

This review focuses on randomised comparisons of oral/topical non-absorbed antifungal prophylaxis compared with no antifungal prophylaxis or compared with systemic antifungal prophylaxis. The effect of systemic antifungal prophylaxis compared with no prophylaxis is addressed in another Cochrane review (Austin 2012).

Objectives

To assess the effect of prophylactic oral/topical non-absorbed antifungal therapy on the incidence of invasive fungal infection, mortality and adverse neurodevelopmental outcomes in very preterm or VLBW infants.

We examined the following interventions:

  1. oral/topical antifungal prophylaxis versus placebo or no drug;

  2. oral/topical antifungal prophylaxis versus systemic antifungal prophylaxis;

  3. one oral/topical antifungal regimen versus another oral/topical antifungal regimen.

We pre-specified these subgroup analyses:

  1. extremely preterm (< 28 weeks) or ELBW infants (< 1000 grams);

  2. trials in which participants were infants with fungal colonisation.

Methods

Criteria for considering studies for this review

Types of studies

Randomised or quasi-randomised controlled trials, including cluster randomised trials.

Types of participants

VLBW infants (< 1500 grams) or very preterm infants (< 32 weeks at birth).

Types of interventions

Antifungal prophylaxis with oral/topical non-absorbed drugs versus placebo or nothing or another antifungal drug regimen.

Types of outcome measures

Primary outcomes
  1. Confirmed invasive fungal infection as determined by:

    1. culture of fungus from a normally sterile site: cerebrospinal fluid, blood, urine, bone or joint, peritoneum, pleural space. Samples should have been collected using methods to minimise contamination with surface colonising organisms;

    2. findings on autopsy examination consistent with invasive fungal infection;

    3. findings on ophthalmological examination consistent with fungal ophthalmitis or retinitis;

    4. pathognomonic findings on renal ultrasound examination such as 'renal fungal balls'.

  2. Death prior to hospital discharge.

  3. Neurodevelopmental outcomes assessed beyond infancy (neurological evaluations, developmental scores, and classifications of disability, including auditory and visual disability, non-ambulant cerebral palsy, developmental delay) and cognitive and educational outcomes at five years or older (intelligence quotient and/or indices of educational achievement measured using a validated tool including school examination results).

Secondary outcomes
  1. Incidence of bronchopulmonary dysplasia (oxygen supplementation at 36 weeks postmenstrual age);

  2. Incidence of necrotising enterocolitis (Bell stage 2 or 3);

  3. Incidence of retinopathy of prematurity: a) any stage; b) requiring treatment;

  4. Duration of intensive care unit or hospital admission (days);

  5. Adverse events attributed to drug reactions or toxicity sufficient to cease drug administration.

Search methods for identification of studies

We used the standard search strategy of the Cochrane Neonatal Review Group.

Electronic searches

We searched the Cochrane Central Register of Controlled Trials (CENTRAL, The Cochrane Library, 2012, Issue 1), MEDLINE (1966 to August 2012), EMBASE (1980 to August 2012), and CINAHL (1982 to August 2012), using a combination of the following text words and MeSH terms: [Infant, Newborn OR Infant, Premature OR Infant, Low Birth Weight OR LBW OR infan* OR neonat*] AND [Mycoses/ OR fung* OR candid* OR Candida albicans OR Antifungal Agents/ OR Triazoles/ OR fluconazole OR azole OR amphotericin B OR nystatin OR nystan OR mycostatin OR nilstat OR nystex OR miconazole OR daktarin OR ketoconazole OR clotrimazole]. The search outputs were limited with the relevant search filters for clinical trials as recommended in the Cochrane Handbook. We did not apply any language restriction.

We searched ClinicalTrials.gov and Current Controlled Trials for completed or ongoing trials.

Searching other resources

We examined the references in studies identified as potentially relevant. We also searched the abstracts from the annual meetings of the Pediatric Academic Societies (1993 to 2012), the European Society for Pediatric Research (1995 to 2011), the UK Royal College of Paediatrics and Child Health (2000 to 2012), and the Perinatal Society of Australia and New Zealand (2000 to 2012). We considered trials reported only as abstracts to be eligible if sufficient information was available from the report, or from contact with the authors, to fulfil the inclusion criteria.

Data collection and analysis

We used the standard methods of the Cochrane Neonatal Review Group.

Selection of studies

Two review authors screened the title and abstract of all studies identified by the above search strategy. We reassessed the full text of any potentially eligible reports and excluded those studies that did not meet all of the inclusion criteria. We discussed any disagreements until we achieved consensus.

Data extraction and management

We used a data collection form to aid extraction of relevant information from each included study. Two review authors extracted the data separately. We discussed any disagreements until we achieved consensus. We asked the investigators for further information if data from the trial reports were insufficient.

Assessment of risk of bias in included studies

We used the criteria and standard methods of the Cochrane Neonatal Review Group to assess the methodological quality of any included trials. We requested additional information from the trial authors to clarify methodology and results as necessary. We evaluated and reported the following issues in the Risk of Bias tables:

  1. Sequence generation: We categorised the method used to generate the allocation sequence as:

    1. low risk: any random process e.g. random number table; computer random number generator;

    2. high risk: any nonrandom process e.g. odd or even date of birth; patient case-record number;

    3. unclear.

  2. Allocation concealment: We categorised the method used to conceal the allocation sequence as:

    1. low risk: e.g. telephone or central randomisation; consecutively numbered sealed opaque envelopes;

    2. high risk: open random allocation; unsealed or non-opaque envelopes, alternation; date of birth;

    3. unclear.

  3. Blinding: We assessed blinding of participants, clinicians and caregivers, and outcome assessors separately for different outcomes and categorised the methods as:

    1. low risk;

    2. high risk;

    3. unclear.

  4. Incomplete outcome data: We described the completeness of data including attrition and exclusions from the analysis for each outcome and any reasons for attrition or exclusion where reported. We assessed whether missing data were balanced across groups or were related to outcomes. Where sufficient information was reported or supplied by the trial authors, we reinstated missing data in the analyses. We categorised completeness as:

    1. low risk: < 20% missing data;

    2. high risk: > 20% missing data;

    3. unclear.

Measures of treatment effect

We calculated risk ratio (RR) and risk difference (RD) for dichotomous data and weighted mean difference (WMD) for continuous data, with respective 95% confidence intervals (CI). We determined the number needed to treat for benefit (NNTB) or harm (NNTH) for analyses with a statistically significant difference in the RD.

Unit of analysis issues

The unit of analysis is the participating infant in individually randomised trials, and the neonatal unit for cluster randomised trials.

Assessment of heterogeneity

We examined the treatment effects of individual trials and heterogeneity between trial results by inspecting the forest plots. We calculated the I² statistic for each RR analysis to quantify inconsistency across studies and describe the percentage of variability in effect estimates that may be due to heterogeneity rather than to sampling error. If we detected substantial heterogeneity (I² > 50%), we explored the possible causes (for example, differences in study design, participants, interventions, or completeness of outcome assessments).

Assessment of reporting biases

If more than five trials were included in a meta-analysis, we examined a funnel plot for asymmetry.

Data synthesis

We used the fixed-effect model in Review Manager 5 (RevMan 2011) for meta-analysis.

Subgroup analysis and investigation of heterogeneity

We prespecified the following subgroup analyses:

  1. extremely preterm (< 28 weeks) or ELBW infants (< 1000 grams);

  2. infants with fungal colonisation at trial entry.

Results

Description of studies

See: Characteristics of included studies; Characteristics of excluded studies; Characteristics of ongoing studies.

We included six eligible trials (reported in five primary publications): Sims 1988; Wainer 1992: Ozturk 2006; Violaris 2010; Aydemir 2011a; Aydemir 2011b; see Characteristics of included studies. We found one ongoing trial: Yekta 2012; see Characteristics of ongoing studies.

Included studies

ORAL/TOPICAL NON-ABSORBED ANTIFUNGAL PROPHYLAXIS VERSUS PLACEBO OR NO DRUG (COMPARISON 1):

Four trials compared oral/topical non-absorbed antifungal prophylaxis with placebo or no drug:

Sims 1988 quasi-randomly allocated 67 infants of birth weight < 1250 grams to receive either oral nystatin or no treatment until one week after endotracheal extubation (average five weeks).

Wainer 1992 recruited 600 infants of birth weight < 1750 grams. We made a consensus decision to include the trial because most participating infants were < 1500 grams. Participants were randomised to receive either oral miconazole or placebo until discharge. The study was undertaken in the late 1980s in South Africa. Due to limited resources mechanical ventilation was not offered to ELBW infants (12% of the participants).

Ozturk 2006 randomly allocated 938 VLBW infants to receive either prophylactic oral nystatin (100000 IU three times daily) or no treatment. Infants in the control group who had oral fungal colonisation detected at trial entry or on surveillance cultures were treated with nystatin (100,000 IU three times daily).

Aydemir 2011a randomly allocated 185 VLBW infants to receive either oral nystatin 100000 IU three times daily or "equal volumes of intravenous or oral normal saline" placebo every third day until the 30th day after birth (or 45th day in ELBW infants).

The primary outcomes of all studies were fungal colonisation and invasive fungal infection. All provided data on in-hospital mortality but none assessed any postdischarge outcomes.

ORAL/TOPICAL NON-ABSORBED VERSUS SYSTEMIC ANTIFUNGAL PROPHYLAXIS (COMPARISON 2):

Two trials compared oral/topical antifungal prophylaxis with systemic antifungal prophylaxis:

Violaris 2010 randomised 80 VLBW infants to receive either oral nystatin or fluconazole beginning between days five to seven after birth. Outcome data on invasive fungal infection and mortality were reported.

Aydemir 2011b randomly allocated 187 VLBW infants to receive either oral nystatin 100,000 U/ml eight hourly or intravenous fluconazole 3 mg/kg every third day versus until 30 days after birth (or 45 days after birth in ELBW infants).

ONE ORAL/TOPICAL NON-ABSORBED ANTIFUNGAL REGIMEN VERSUS ANOTHER ORAL/TOPICAL NON-ABSORBED ANTIFUNGAL REGIMEN (COMPARISON 3):

We did not find any trials that compared different dose regimens of oral/topical non-absorbed antifungal prophylaxis.

Excluded studies

We excluded three studies (Harris 1960; Damjanovic 1993; Herruzo-Cabrera 1994; see Characteristics of excluded studies).

Risk of bias in included studies

Quality assessments are described in the table Characteristics of included studies and displayed in Figure 1. Only one small trial was quasi-randomised and lacked allocation concealment (Sims 1988). The most common methodological weakness was lack of blinding of caregivers and investigators and assessors to the nature of the intervention. Only one trial is likely to have been truly placebo-controlled (Wainer 1992). All of the trials reported complete or near-complete assessment for primary outcomes.

Figure 1.

Risk of bias graph: review authors' judgements about each risk of bias item presented as percentages across all included studies.

Effects of interventions

ORAL/TOPICAL NON-ABSORBED ANTIFUNGAL PROPHYLAXIS VERSUS PLACEBO OR NO DRUG (COMPARISON 1):

Primary outcomes:

Incidence of invasive fungal infection (Outcome 1.1): Meta-analysis of data from four trials found a statistically significant reduction in the intervention group but with significant and substantial statistical heterogeneity (Figure 2): typical RR 0.20 (95% CI 0.14 to 0.27), I² = 80%; typical RD -0.18 (95% CI -0.21 to -0.16); NNTB five infants.

Figure 2.

Forest plot of comparison: 1 Oral/topical non-absorbed antifungal prophylaxis vs placebo or nothing, outcome: 1.1 Incidence of invasive fungal infection.

Forest plot inspection suggested that the direction of the effect size estimate from Wainer 1992 was inconsistent with those of the other three trials. Meta-analysis omitting data from Wainer 1992 removed heterogeneity, but did not change the pooled effect size estimate [revised typical RR 0.16 (95% CI 0.11 to 0.23), I² = 0%].

Death prior to hospital discharge (Outcome 1.2): None of the individual trials or a meta-analysis of data from all four trials found a statistically significant effect. We found no evidence of statistical heterogeneity (Figure 3): typical RR 0.87 (95% CI 0.72 to 1.05), I² = 0%; typical RD -0.03 (95% CI -0.06 to 0.01).

Figure 3.

Forest plot of comparison: 1 Oral/topical non-absorbed antifungal prophylaxis vs placebo or nothing, outcome: 1.2 Mortality.

Neurodevelopmental outcomes: Not reported by any trials.

Secondary outcomes:

Incidence of bronchopulmonary dysplasia (Outcome 1.3): Aydemir 2011a did not find a statistically significant difference: RR 1.29 (95% CI 0.67 to 2.49). Outcome not reported in the other trials.

Incidence of necrotising enterocolitis (Outcome 1.4): Aydemir 2011a did not find a statistically significant difference: RR 0.97 (95% CI 0.40 to 2.33). Outcome not reported in the other trials.

Incidence of retinopathy of prematurity (Outcome 1.5): Aydemir 2011a did not find a statistically significant difference in the incidence of retinopathy requiring surgery: RR 0.62 (95% CI 0.30 to 1.28). Outcome not reported in the other trials.

Duration of intensive care unit stay (Outcome 1.6): Three trials reported length of stay in intensive care (Sims 1988; Wainer 1992; Aydemir 2011a). None of the trials individually, or a meta-analysis of data from all trials, found a statistically significant difference: WMD -0.52 (95% CI -4.34 to 3.29) days (I² = 0%) (Figure 4).

Figure 4.

Forest plot of comparison: 1 Oral/topical non-absorbed antifungal prophylaxis vs placebo or nothing, outcome: 1.6 Length of stay in NICU (days).

Adverse events attributed to drug reactions or toxicity sufficient to cease drug administration: Not reported by any trials.

Subgroup analyses

  1. Extremely preterm or ELBW infants: Data were available from one trial (Ozturk 2006).

    1. invasive fungal infection: Ozturk 2006 found a statistically significant reduction RR 0.12 (95% CI 0.06 to 0.26); RD -0.27 (95% CI -0.35 to -0.20).

    2. mortality: Ozturk 2006 did not find a statistically significant effect: RR 1.07 (95% CI 0.51 to 2.25); RD 0.01 (95% CI -0.05 to 0.06).

  2. Trials in which participants were infants with fungal colonisation: None of the trials restricted participation to infants with fungal colonisation at trial entry.

ORAL/TOPICAL NON-ABSORBED VERSUS SYSTEMIC ANTIFUNGAL PROPHYLAXIS (COMPARISON 2)

Primary outcomes:

Incidence of invasive fungal infection (Outcome 2.1): Meta-analysis of data from two trials did not detect a statistically significant difference (Figure 5): typical RR 1.89 (95% CI 0.66 to 5.39), I² = 0%; typical RD 0.03 (95% CI -0.02 to 0.09).

Figure 5.

Forest plot of comparison: 2 Oral/topical non-absorbed prophylaxis vs. systemic antifungal prophylaxis, outcome: 2.1 Incidence of invasive fungal infection.

Death prior to hospital discharge (Outcome 2.2): Meta-analysis of data from two trials did not detect a statistically significant difference (Figure 6): typical RR 0.04 (95% CI -0.02 to 0.11), I² = 66%; typical RD 0.03 (95% CI -0.02 to 0.09).

Figure 6.

Forest plot of comparison: 2 Oral/topical non-absorbed prophylaxis vs. systemic antifungal prophylaxis, outcome: 2.2 Mortality.

Neurodevelopmental outcomes: Neither of the trials reported any neurodevelopmental outcomes.

Secondary outcomes:

Incidence of bronchopulmonary dysplasia in surviving infants: Aydemir 2011b did not find a statistically significant difference: RR 1.29 [95% CI 0.67 to 2.49] RD 0.04 [95% CI -0.07 to 0.16]. Not reported by Violaris 2010.

Incidence of necrotising enterocolitis: Meta-analysis of data from both trials did not detect a statistically significant difference (Figure 7): typical RR 1.22 (95% CI 0.58 to 2.60), I² = 0%; RD 0.02 (95% CI -0.05 to 0.09).

Figure 7.

Forest plot of comparison: 2 Oral/topical non-absorbed prophylaxis vs. systemic antifungal prophylaxis, outcome: 2.4 Necrotising enterocolitis.

Incidence of retinopathy of prematurity: Aydemir 2011b did not find a statistically significant difference in the incidence of retinopathy requiring surgery in surviving infants: RR 1.24 [95% CI 0.51 to 2.98], RD 0.02 [95% CI -0.07 to 0.11]. Not reported by Violaris 2010.

Duration of intensive care unit stay: Aydemir 2011b did not find a statistically significant difference: MD -1.00 (95% CI -7.63 to 5.63) days. Not reported by Violaris 2010.

Adverse events attributed to drug reactions or toxicity sufficient to cease drug administration: Not reported by either of the trials.

Subgroup analyses

  1. Extremely preterm or ELBW infants: Neither of the trials provided subgroup data.

  2. Trials in which participants were infants with fungal colonisation: Neither of the trials restricted participation to infants with fungal colonisation at trial entry.

Discussion

Summary of main results

Meta-analysis of data from four trials suggests that oral/topical non-absorbed prophylaxis reduces the risk of invasive fungal infection in VLBW infants significantly and substantially. None of the trials or a meta-analysis of their data found a statistically significant effect on mortality. Meta-analysis of data from three trials did not detect an effect on the duration of intensive care. The trials reported only limited data on other neonatal morbidities that may be affected by invasive fungal infection. None of the trials assessed long-term neurodevelopmental outcomes.

Two trials assessed the effect of oral/topical non-absorbed antifungal prophylaxis (nystatin) versus systemic antifungal (fluconazole) prophylaxis. Meta-analyses did not find any statistically significant effects on the incidence of invasive fungal infection or all-cause mortality but much larger studies would be needed to exclude more modest but important effect sizes. Meta-analysis did not detect an effect on the incidence of necrotising enterocolitis but only limited data on other neonatal morbidities were reported.

Overall completeness and applicability of evidence

The finding that oral/topical non-absorbed antifungal prophylaxis reduces the risk of invasive fungal infection in VLBW infants should be interpreted and applied with caution. The existence of substantial statistical heterogeneity in the meta-analysis raises concern that the estimate of effect is not robust. The applicability of the finding is also limited by the very high incidence of invasive fungal infection in the control populations in the three trials that found a statistically significant effect on the incidence of invasive fungal infection (Sims 1988; Ozturk 2006; Aydemir 2011a). About one-sixth to one-third of infants in the control groups developed invasive fungal infection, much higher than the < 5% incidence estimated in large cohort studies (Saiman 2000; Makhoul 2002; Horbar 2002; Karlowicz 2002; Clerihew 2006; Howell 2009). This limits the applicability of the NNTB estimate (five infants), since in clinical settings with lower incidences of invasive fungal infection a much larger number of infants would need treatment to prevent a single extra case of invasive fungal infection.

Quality of the evidence

The largest trial (N = 948) contributed 84% of the weighted estimate of risk ratio effect on invasive fungal infection (Ozturk 2006). This trial of nystatin prophylaxis was undertaken in Turkey within the past decade. More than one-third of participants were ELBW infants receiving intensive care interventions. The criteria for diagnosing invasive fungal infection appear to be have been robust. Efforts to limit contamination of microbiological cultures by surface colonising organisms were made, for example, fungal urinary tract infection was based on culture of organisms from two separate supra-pubic bladder aspirates. However, caregivers or assessors were not blinded to the intervention and this may have caused surveillance and ascertainment bias if thresholds for investigation and diagnosis of suspected invasive fungal infection were adjusted according to treatment status. Although 25% of control VLBW infants received nystatin to treat oral fungal colonisation detected at trial entry or during the trial period, this is likely to have reduced the effect size of the primary intervention.

The second largest trial (N = 600) did not detect a statistically significant effect of miconazole prophylaxis on the incidence of invasive fungal infection (Wainer 1992). This trial was placebo-controlled and therefore less prone to surveillance bias. The trial was undertaken in South Africa 25 years ago in a settling with few intensive care resources. Twelve per cent of participants were ELBW and the overall incidence of invasive fungal infection was 2% in the control group. This lower incidence may be related to the fact that because of resource limitations ELBW infants did not receive intensive care interventions. Two-thirds of ELBW infants died. The applicability of the trial's findings to modern neonatal intensive care settings in high-income countries is therefore likely to be limited.

A subgroup analysis of outcomes for infants colonised with fungi at trial entry was not possible. None of the trials prespecified fungal colonisation as an entry criterion. Between 25% and 45% of participating infants had fungal colonisation, but subgroup data for these infants were not available in the published reports of the included trials. Even if these data become available for analysis, those from the largest trial would be of limited value since infants in the control group received antifungal treatment if oral fungal colonisation was detected (Ozturk 2006).

Potential biases in the review process

The existence of substantial statistical heterogeneity in the meta-analysis of the effect of oral/topical non-absorbed antifungal prophylaxis versus placebo or no drug on the incidence of invasive fungal infection raises concern that the estimate is not robust (Figure 2). The heterogeneity may be due to differences between the trials including population characteristics (proportion of ELBW infants), nature of the intervention (miconazole in one trial, nystatin in the others), methodological quality issues (particularly unblinded allocation and intervention) and the effect of other co-interventions (availability of intensive care for ELBW infants). Forest plot inspection suggested that the direction of the effect size estimate from Wainer 1992 was inconsistent with those of the other three trials. In a post hoc sensitivity analysis, removal of this trial from the meta-analysis removed statistical heterogeneity from the RR estimate and but did not change the direction or size of the estimate.

Authors' conclusions

Implications for practice

The available trial data remain insufficient to guide clinical practice. Although meta-analysis suggests that oral/topical non-absorbed antifungal agents (nystatin or miconazole) reduce the risk of invasive fungal infection, methodological weaknesses limit the validity and applicability of this finding.

Implications for research

Further randomised controlled trials of oral/topical non-absorbed antifungal prophylaxis are needed to provide more valid and precise estimates of effect size. Because most neonatologists who currently use antifungal prophylaxis target infants thought to be at greatest risk, mainly ELBW or extremely preterm infants with additional risk factors, a trial restricted to this population of infants or even smaller or lower gestation infants may be appropriate and acceptable (Burwell 2006; Clerihew 2008; Howell 2009). Oral/topical non-absorbed antifungal prophylaxis may be compared with placebo or with systemic prophylaxis (Isaacs 2008; Austin 2012). Any trial should aim to assess long-term outcomes, particularly disability-free survival, as well as the effect on invasive fungal infection.

Acknowledgements

David Henderson-Smart for his guidance.

Data and analyses

Download statistical data

Comparison 1. Oral/topical non-absorbed antifungal prophylaxis vs placebo or nothing
Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size
1 Incidence of invasive fungal infection4 Risk Ratio (M-H, Fixed, 95% CI)Subtotals only
1.1 All VLBW infants41800Risk Ratio (M-H, Fixed, 95% CI)0.20 [0.14, 0.27]
1.2 Only ELBW infants1349Risk Ratio (M-H, Fixed, 95% CI)0.12 [0.06, 0.26]
2 Mortality4 Risk Ratio (M-H, Fixed, 95% CI)Subtotals only
2.1 All VLBW infants41800Risk Ratio (M-H, Fixed, 95% CI)0.87 [0.72, 1.05]
2.2 Only ELBW infants1349Risk Ratio (M-H, Fixed, 95% CI)1.07 [0.51, 2.25]
3 Bronchopulmonary dysplasia1171Risk Ratio (M-H, Fixed, 95% CI)1.29 [0.67, 2.49]
4 Necrotising enterocolitis1185Risk Ratio (M-H, Fixed, 95% CI)0.97 [0.40, 2.33]
5 Retinopathy of prematurity1177Risk Ratio (M-H, Fixed, 95% CI)0.62 [0.30, 1.28]
6 Length of stay in NICU (days)3833Mean Difference (IV, Fixed, 95% CI)-0.52 [-4.34, 3.29]
Analysis 1.1.

Comparison 1 Oral/topical non-absorbed antifungal prophylaxis vs placebo or nothing, Outcome 1 Incidence of invasive fungal infection.

Analysis 1.2.

Comparison 1 Oral/topical non-absorbed antifungal prophylaxis vs placebo or nothing, Outcome 2 Mortality.

Analysis 1.3.

Comparison 1 Oral/topical non-absorbed antifungal prophylaxis vs placebo or nothing, Outcome 3 Bronchopulmonary dysplasia.

Analysis 1.4.

Comparison 1 Oral/topical non-absorbed antifungal prophylaxis vs placebo or nothing, Outcome 4 Necrotising enterocolitis.

Analysis 1.5.

Comparison 1 Oral/topical non-absorbed antifungal prophylaxis vs placebo or nothing, Outcome 5 Retinopathy of prematurity.

Analysis 1.6.

Comparison 1 Oral/topical non-absorbed antifungal prophylaxis vs placebo or nothing, Outcome 6 Length of stay in NICU (days).

Comparison 2. Oral/topical non-absorbed prophylaxis vs. systemic antifungal prophylaxis
Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size
1 Incidence of invasive fungal infection2267Risk Ratio (M-H, Fixed, 95% CI)1.89 [0.66, 5.39]
2 Mortality2267Risk Ratio (M-H, Fixed, 95% CI)1.65 [0.72, 3.77]
3 Bronchopulmonary dysplasia1171Risk Ratio (M-H, Fixed, 95% CI)1.29 [0.67, 2.49]
4 Necrotising enterocolitis2267Risk Ratio (M-H, Fixed, 95% CI)1.22 [0.58, 2.60]
5 Retinopathy of prematurity1171Risk Ratio (M-H, Fixed, 95% CI)1.24 [0.51, 2.98]
6 Length of stay in NICU (days)1171Mean Difference (IV, Fixed, 95% CI)-1.0 [-7.63, 5.63]
Analysis 2.1.

Comparison 2 Oral/topical non-absorbed prophylaxis vs. systemic antifungal prophylaxis, Outcome 1 Incidence of invasive fungal infection.

Analysis 2.2.

Comparison 2 Oral/topical non-absorbed prophylaxis vs. systemic antifungal prophylaxis, Outcome 2 Mortality.

Analysis 2.3.

Comparison 2 Oral/topical non-absorbed prophylaxis vs. systemic antifungal prophylaxis, Outcome 3 Bronchopulmonary dysplasia.

Analysis 2.4.

Comparison 2 Oral/topical non-absorbed prophylaxis vs. systemic antifungal prophylaxis, Outcome 4 Necrotising enterocolitis.

Analysis 2.5.

Comparison 2 Oral/topical non-absorbed prophylaxis vs. systemic antifungal prophylaxis, Outcome 5 Retinopathy of prematurity.

Analysis 2.6.

Comparison 2 Oral/topical non-absorbed prophylaxis vs. systemic antifungal prophylaxis, Outcome 6 Length of stay in NICU (days).

What's new

DateEventDescription
27 January 2013New search has been performedThis updates the review 'Prophylactic oral/topical non-absorbed antifungal agents to prevent invasive fungal infection in very low birth weight infants' (Austin 2009).
27 January 2013New citation required but conclusions have not changed

Updated search identified new trials for inclusion (Violaris 2010; Aydemir 2011a; Aydemir 2011b).

Conclusions unchanged.

History

Protocol first published: Issue 1, 2002
Review first published: Issue 1, 2004

DateEventDescription
4 June 2009New citation required and conclusions have changedSubstantive update.
14 October 2007New citation required and conclusions have changedSubstantive amendment.

Contributions of authors

William McGuire (WM) screened the title and abstract of all studies identified by the search strategy. Nicola Austin (NA) and WM screened the full text of the report of each study identified as of potential relevance. NA and WM extracted the data separately, compared data, and resolved differences by consensus, and with discussion with Brian Darlow (BD). NA, BD and WM completed the final review.

Declarations of interest

None.

Sources of support

Internal sources

  • Christchurch Women's Hospital, Christchurch, New Zealand.

  • Centre for Reviews and Dissemination, University of York, UK.

External sources

  • The Cochrane Neonatal Review Group has been funded in part with Federal funds from the Eunice Kennedy Shriver National Institute of Child Health and Human Development National Institutes of Health, Department of Health and Human Services, USA, under Contract No. HHSN267200603418C., USA.

    Eunice Kennedy Shriver National Institute of Child Health and Human Development National Institutes of Health, Department of Health and Human Services

Differences between protocol and review

None.

Characteristics of studies

Characteristics of included studies [ordered by study ID]

Aydemir 2011a

MethodsRandomised controlled trial
Participants185 VLBW infants
InterventionsOral nystatin 100,000 U/ml 8 hourly (N = 94) versus normal saline placebo* (N= 91) every third day versus until the 30th day after birth (or 45th day in ELBW infants)
Outcomes

Fungal colonisation and invasive infection

Death prior to hospital discharge

Emergence of fungi with native azole resistance
Adverse drug reactions

Notes

Setting: Zekai Tahir Burak Maternity Hospital, Ankara, Turkey; 2008-9

*Report states placebo controlled but unclear how this was achieved

The same infants form the oral nystatin group in both Aydemir 2011a and Aydemir 2011b

Risk of bias
BiasAuthors' judgementSupport for judgement
Allocation concealment (selection bias)Low riskComputer-generated allocation
Blinding (performance bias and detection bias)
All outcomes
Unclear riskReport states placebo controlled but unclear how this was achieved
Blinding of participants and personnel (performance bias)
All outcomes
Unclear riskReport states placebo controlled but unclear how this was achieved
Blinding of outcome assessment (detection bias)
All outcomes
Unclear riskReport states placebo controlled but unclear how this was achieved
Incomplete outcome data (attrition bias)
All outcomes
Low riskComplete follow-up

Aydemir 2011b

MethodsRandomised controlled trial
Participants187 VLBW infants.
InterventionsOral nystatin 100,000 U/ml 8 hourly (N = 94) versus fluconazole 3 mg/kg (N = 93) every third day versus until the 30th day after birth (or 45th day in ELBW infants)
Outcomes

Fungal colonisation and invasive infection

Death prior to hospital discharge

Emergence of fungi with native azole resistance
Adverse drug reactions

Notes

Setting: Zekai Tahir Burak Maternity Hospital, Ankara, Turkey; 2008-9

The same infants form the oral nystatin group in both Aydemir 2011a and Aydemir 2011b

Risk of bias
BiasAuthors' judgementSupport for judgement
Allocation concealment (selection bias)Low riskComputer-generated allocation
Blinding (performance bias and detection bias)
All outcomes
Unclear riskReport states placebo controlled but unclear how this was achieved
Blinding of participants and personnel (performance bias)
All outcomes
Unclear riskReport states placebo controlled but unclear how this was achieved
Blinding of outcome assessment (detection bias)
All outcomes
Unclear riskReport states placebo controlled but unclear how this was achieved
Incomplete outcome data (attrition bias)
All outcomes
Low riskComplete follow-up

Ozturk 2006

MethodsRandomised controlled trial
Participants938 VLBW infants with subgroup report for ELBW infants (N = 349)
InterventionsNystatin 100000 IU orally, 8 hourly (N = 475) versus no drug (N = 463)
OutcomesFungal colonisation and invasive fungal infection
Notes

Setting: Division of Neonatology, Erciyes University Hospital, Turkey, 2002-2005

25% of control VLBW infants received nystatin (100000 IU orally, 8 hourly) to treat oral fungal colonisation detected at trial entry or during the trial period

Risk of bias
BiasAuthors' judgementSupport for judgement
Allocation concealment (selection bias)Low riskRandomly assigned by "someone not directly involved in the study" using random number tables
Blinding (performance bias and detection bias)
All outcomes
High riskUnblinded
Blinding of participants and personnel (performance bias)
All outcomes
High riskUnblinded
Blinding of outcome assessment (detection bias)
All outcomes
High riskUnblinded
Incomplete outcome data (attrition bias)
All outcomes
Low riskComplete follow-up

Sims 1988

MethodsQuasi-randomised controlled trial; odd or even hospital number allocation
Participants67 infants of birth weight < 1250 grams
Interventions

Nystatin 1ml orally, 8 hourly (N = 33) versus no drug (N = 34)

Treatment from inclusion until one week after endotracheal extubation

Outcomes

Fungal colonisation and invasive fungal infection

Duration of mechanical ventilation and duration of intensive care admission

Notes

Setting: Los Angeles County-University of Southern California Medical Centre, 1985 to 1986

The study took place during a period of overcrowding in the intensive care unit; 222 infants with a birthweight < 1250 grams were born during a 12 month period; 55 died within 48 hours, 88 relatively healthy infants were transferred elsewhere and 67 of the remaining 88 infants were recruited to the study.

One infant in the control group had Candida albicans pneumonia supported by postmortem evidence. All the other affected infants had positive urine and blood cultures

Risk of bias
BiasAuthors' judgementSupport for judgement
Allocation concealment (selection bias)High riskQuasi-randomised
Blinding (performance bias and detection bias)
All outcomes
High riskUnblinded
Blinding of participants and personnel (performance bias)
All outcomes
High riskUnblinded
Blinding of outcome assessment (detection bias)
All outcomes
High riskUnblinded
Incomplete outcome data (attrition bias)
All outcomes
Low riskComplete follow-up

Violaris 2010

MethodsRandomised controlled trial
Participants

80 VLBW infants

Haemodynamically unstable infants and infants with severe congenital anomalies or abnormal liver function tests were not eligible to participate

Interventions

Nystatin (100,000 units/kg/day) in each side of the mouth (N = 42) versus fluconazole (4 mg/kg) orally (N = 38) beginning on day five after birth

Medications were continued until full oral feedings were attained or systemic fungal infection was diagnosed

OutcomesInvasive fungal infection, invasive bacterial infection, biochemical indices related to liver function, mortality
NotesSetting: Brooklyn Hospital Center, New York; 1997- 1998
Risk of bias
BiasAuthors' judgementSupport for judgement
Allocation concealment (selection bias)Low riskComputerised randomisation and allocation
Blinding (performance bias and detection bias)
All outcomes
High riskUnable to blind interventions
Blinding of participants and personnel (performance bias)
All outcomes
High riskUnable to blind interventions
Blinding of outcome assessment (detection bias)
All outcomes
High riskUnable to blind interventions
Incomplete outcome data (attrition bias)
All outcomes
Low riskComplete follow up

Wainer 1992

  1. a

    ELBW: extremely low birth weight
    VLBW: very low birth weight

MethodsRandomised controlled trial
Participants600 infants of birth weight < 1750 grams.
InterventionsMiconazole 0.75ml orally 3 times daily (N = 302) vs. placebo (N = 298)
OutcomesFungal colonisation and invasive fungal infection. Duration of mechanical ventilation and duration of intensive care admission
Notes

Setting: Baragwanath Hospital, South Africa from October 1989 to July 1990

Due to limited resources, mechanical ventilation was not offered to infants with birth weight < 1000 gm. This group made up 12% (73/600) of the infants and had a high mortality rate (67%)

Risk of bias
BiasAuthors' judgementSupport for judgement
Allocation concealment (selection bias)Unclear riskNot described
Blinding (performance bias and detection bias)
All outcomes
Low riskPlacebo-controlled
Blinding of participants and personnel (performance bias)
All outcomes
Low riskPlacebo-controlled
Blinding of outcome assessment (detection bias)
All outcomes
Low riskPlacebo-controlled
Incomplete outcome data (attrition bias)
All outcomes
Low riskComplete follow-up

Characteristics of excluded studies [ordered by study ID]

StudyReason for exclusion
Damjanovic 1993Not a randomised controlled trial
Harris 1960The gestational age or birth weight of participants was not reported - assumed to include term and preterm infants
Herruzo-Cabrera 1994A prospective cohort study but not a randomised controlled trial

Characteristics of ongoing studies [ordered by study ID]

Yekta 2012

Trial name or titleComparison of Lactobacillus reuteri and nystatin prophylaxis on Candida colonization and infection in very low birth weight Infants
MethodsRandomised controlled trial
ParticipantsVery preterm or VLBW infants
Interventions
  • Probiotic: Lactobacillus reuteri 100 million CFU/day

  • Nystatin 50000 unit/3 times a day, both for orally and by orogastric route

OutcomesRectal and skin swabs for Candida colonization taken weekly. Blood culture for Candida infection taken weekly
Starting dateFebruary 2012
Contact informationMehmet Yekta, Zekai Tahir Burak Maternity and Teaching Hospital, Turkey
NotesClinicalTrials.gov Identifier: NCT01531192