Comparison of fluconazole and nystatin oral suspensions for treatment of oral candidiasis in infants

Oral thrush is a common condition in young infants. Nystatin treatment is associated with frequent recurrences and difficulty in administration. Fluconazole was compared with nystatin for the treatment of oral candidiasis in infants. Thirty-four infants were randomized to either nystatin oral suspension four times a day for 10 days or fluconazole suspension 3 mg/kg in a single daily dose for 7 days. Clinical cures for nystatin were 6 of 19 (32%), and those for fluconazole were 15 of 15 (100%), P < 0.0001. In this small pilot study fluconazole was shown to be superior to nystatin suspension for the treatment of oral thrush in otherwise healthy infants. Oral candidiasis or thrush occurs in 1 to 37% of infants. 1, 2 For decades nystatin oral suspension has been the standard therapy for thrush. Although some very early studies suggested efficacy rates for nystatin as high as 90%, 1 these rates were not confirmed by a more recent study from Germany, which showed a clinical cure rate for nystatin of only 54% in otherwise healthy infants. 3 Several newer agents, primarily from the imidazole class of antifungals, have been evaluated for the treatment of thrush. Fluconazole has been effective, safe and easy to use for the treatment of thrush in immunocompromised children. 4 No study to date, however, has evaluated fluconazole in the treatment of oral candidiasis in otherwise healthy infants, and it is not Food and Drug Administration-approved for this indication. We therefore compared fluconazole suspension and nystatin oral suspension for treatment of thrush in infants with regard to clinical and microbiologic efficacy, recurrence rate and complication rate. Methods. Study design. This study was designed to be a prospective, unblinded, randomized clinical trial. We conducted this office-based study at two military medical facilities: Wilford Hall USAF Medical Center in San Antonio, TX; and Balboa US Naval Medical Center in San Diego, CA. The study was approved by the institutional review boards at both institutions. Written informed consent was obtained from the parent(s) of each infant enrolled in the study. Study population. Infants 1 to 12 months of age with clinical signs of oral thrush, with or without signs of candidal diaper dermatitis, were eligible for enrollment in the study. Exclusion criteria included: current use of any antimicrobial; an underlying chronic medical condition; known allergy to either drug used in the study; and known immunodeficiency. A study subject was considered evaluable for clinical and microbiologic cure if follow-up occurred at the end of drug therapy and evaluable for clinical and microbiologic recurrence if follow-up occurred at the end of the study protocol. Study drugs. Patients enrolled in the control group of the study received nystatin suspension containing 100 000 IU/ml. Parents were instructed to directly apply 1 ml of the suspension to the affected oral mucosa four times a day for 10 days. Application involved repeated rubbing of the candidal plaques with either soaked cotton swabs or a washcloth soaked in the nystatin suspension. Instruction regarding nystatin application was given either by the investigator enrolling the patient into the study or by trained nurses. Patients enrolled in the study group received fluconazole suspension (Diflucan, 10 mg/ml; Pfizer, New York, NY), 3 mg/kg administered orally once daily for 7 days. Study protocol. Initial enrollment of eligible patients occurred after preliminary confirmation of oral candidiasis with a positive potassium hydroxide preparation of oral plaques. The extent of oral candidiasis and the presence of any candidal diaper rash were noted at the time of enrollment (Day 0). The type of feeding (breast, bottle or both) was also noted, as was the presence of risk factors for development of thrush (pacifier use and the use of antibiotics within the previous 45 days). An initial culture of the oral plaques was obtained on Day 0 using a Venturi Transystem transport culture swab. Semiquantitative fungal cultures were done by standard methods. A patient with a negative initial fungal culture was dropped from the study. Patients were assigned to either the control (nystatin) group or study (fluconazole) group with the use of a standard randomization scheme. Mothers of breast-fed infants were prescribed nystatin cream to be applied to their nipples twice daily for the duration of the infant’s treatment. Infants assigned to the control group with a candidal diaper rash were prescribed nystatin cream, whereas infants assigned to the study group with a candidal diaper rash were prescribed only oral fluconazole. Patients in both groups were followed up on Day 3 of treatment (±1 day), at the end of therapy on Day 10 (±1 or 2 days) for patients given nystatin or Day 7 (±1 or 2 days) for patients given fluconazole and on Day 28 (±2 days). At each follow-up visit oral and skin examinations were performed. Possible complications of therapy by history or physical examination were recorded. Repeat fungal cultures were obtained at each follow-up visit. The absence of any oral plaques at the end of therapy (Day 7 or 10) was considered a clinical cure. A negative fungal culture obtained at that time was considered a microbiologic cure. Any patient who experienced a clinical failure on Day 7 (fluconazole group) or Days 10 to 14 (nystatin group) was offered the alternative medication. A clinical recurrence was determined by the presence of oral plaques confirmed to be fungal by potassium hydroxide preparation on Day 28. Similarly a microbiologic recurrence was determined by a positive fungal culture on Day 28. Statistical methods. Exploratory data analysis was performed with percentage use where appropriate. Means (with sd) were used for normally distributed continuous variables. Statistical significance of associations was assessed with either the chi square test of homogeneity or Fisher’s exact test (the latter used when an expected cell size was <5). Comparisons between treatment groups were stratified by potential confounders to control for their effect. Statistical significance of the associations was judged to occur at P < 0.05. Results. Demographics. Forty-seven patients were initially enrolled. Of these, 19 patients were randomized to the fluconazole group and 28 to the nystatin group. Four patients receiving fluconazole were not evaluable for clinical cure (2 were lost to follow-up; 2 did not meet initial enrollment age eligibility requirements). Seven patients from the nystatin group were not evaluable for clinical cure; all were lost to follow-up. Thirty-six patients were evaluable for clinical cure, 21 in the control group and 15 in the study group. All 15 patients in the study group were evaluable for microbiologic cure, and 18 patients in the control group were similarly evaluable (end of therapy cultures were lost in 3 patients receiving nystatin). No patients were dropped out of either group because of complication or poor compliance. No statistical differences were noted between groups for gender, age or the presence of risk factors (bottle-feeding, pacifier use and recent antibiotic use). Clinical efficacy. Six of 21 patients randomized to receive nystatin were clinically cured at 10 days representing a clinical cure rate of 28.6%. This compares with a clinical cure rate of 100% (15 of 15) at 7 days for the patients randomized to receive fluconazole (P < 0.0001). Microbiologic efficacy. All patients enrolled in the study had initial cultures positive for Candidaalbicans. One of the 18 patients evaluated for microbiologic efficacy in the nystatin group (5.6%) had microbiologic cure by Day 10 of therapy. Eleven of 15 patients in the fluconazole group (73.3%) were cured microbiologically by Day 7 of therapy, and 10 of these 11 patients (90.9%) were cured by Day 3 of therapy (P < 0.0001). All positive cultures at the end of therapy were for C. albicans except for one positive culture of Candidaparapsilosis on Day 7 in a patient receiving fluconazole who had no clinical evidence of disease at that time. Clinical and microbiologic recurrence. Thirteen patients were evaluated for clinical recurrence in the fluconazole group. Three (23%) of these patients had a clinical relapse of thrush by Day 28. The microbiologic recurrence rate in the fluconazole group was 56% (5 of 9). Of these 5 positive cultures 4 were C. albicans and one culture was Candida lusitaniae. The latter patient was disease-free at the time of this positive culture. Because 15 of the 21 patients in the nystatin group received fluconazole after initial treatment failure, little information was available to evaluate clinical and microbiologic recurrence rates in the control group. This precludes statistical comparison of the clinical and microbiologic recurrence rates. Complications of therapy and coexistence of candidal diaper rash. There were no complications of therapy in either the nystatin or the fluconazole group. This includes the patients prescribed fluconazole after initial treatment failure with nystatin. Nineteen percent of the patients in the nystatin group and 20% of the patients in the fluconazole group had a candidal diaper rash on Day 0. All of these patients were clinically cured of their rashes by the end of therapy. Two patients in the nystatin group who did not initially have a candidal diaper rash developed one during the course of therapy with oral nystatin suspension. Discussion. Since the 1950s the standard treatment for thrush has been nystatin suspension, a polyene antifungal that, although fungicidal at very high dosages, is fungistatic at typical therapeutic dosages. 1 Early studies in the 1950s evaluating the efficacy of nystatin demonstrated clinical cure rates of up to 80 to 90%. 1 Most of these early studies were hampered, however, by poor study design and the use of variable dosages of nystatin that in some cases were 4 to 10 times the standard dose of 100 000 units four times daily. 1 Those studies also noted poor mycologic cure rates and high relapse rates. 1 More recent studies have demonstrated clinical cure rates with nystatin of no more than 51% in immunocompromised patients and 54% in healthy infants and even worse mycologic cure rates of 11 and 15.2% in immunocompromised and healthy patients respectively. 3, 4 Hoppe 1 has noted that beyond its only moderate antifungal activity at normally prescribed doses, the primary problem with nystatin suspension in the treatment of thrush in infants is its poor adherence to the oral mucosa. Infants will readily swallow the suspension and thus greatly limit the duration of contact this essentially topical agent has with the candidal lesions. 1 The four times a day dosing schedule and the minimum duration of therapy of 10 to 14 days also create considerable compliance problems. Several medications have been evaluated in hopes of finding a more effective and convenient treatment for thrush in infants. In the 1950s and 1960s gentian violet and amphotericin B suspensions were compared with nystatin. Efficacy rates were no more than 75% for gentian violet and 80% for amphotericin B, and both antifungals present the same problems of application as nystatin. 1 The first generation imidazole antifungal, clotrimazole, as been tried as a topical cream applied directly to the oral mucosa, as an oral troche and as a suppository inserted into the tip of a pacifier. Other routes of administration of clotrimazole have never been adequately studied. 1 Another imidazole, the nonabsorbed agent miconazole gel, has been evaluated in a controlled study and was effective (clinical cure rate of 99% after 12 days of therapy) and safe. 3 In a study in 1989 the dibasic imidazole ketoconazole was used to treat thrush in healthy infants with a clinical cure rate of 100%. 5 The use of ketoconazole is limited by its many drug-drug interactions and significant side effects such as hepatotoxicity and inhibition of steroid synthesis. 6 Fluconazole, an orally absorbed tribasic imidazole, has a much smaller side effect profile than ketoconazole and is generally safe and effective for treatment of a number of local and systemic fungal infections. 6 For treatment of oral candidiasis, it also has the additional potential benefit of high drug concentrations in the oral mucosa and saliva. 6 The use of fluconazole for the treatment of thrush has been restricted primarily to immunocompromised patients in whom it has been quite effective. 7 A study comparing nystatin and fluconazole in immunocompromised children demonstrated a clinical cure rate of 91% for fluconazole vs. 51% for nystatin and also a substantially higher rate of mycologic cure for fluconazole (76%vs. only 11% for nystatin). 4 Fluconazole also has a lower reported rate of inconvenience with administration than nystatin. 8 In our current study comparing the two drugs in healthy infants, fluconazole suspension was also significantly more effective than conventional nystatin suspension in the treatment of oral thrush. Both clinical and microbiologic cure rates were significantly improved in the fluconazole group, by 3.5- and 13-fold, respectively. This improvement was independent of potential risk factors for thrush such as pacifier use and type of feeding. The rate of clinical cure of 100% noted with fluconazole in this study is consistent with the observed rates of 87 to 100% noted in various other studies involving immunocompromised patients. 4, 6–8 The microbiologic cure rate of 73.3% for fluconazole in this study is also consistent with that noted in earlier studies with immunocompromised patients 4, 6–8 and exceeds the microbiologic cure rate noted with miconazole gel. 3 This study therefore confirms the high efficacy rates noted for fluconazole for the treatment of oral candidiasis in immunocompromised patients. We noted no adverse side effects to the use of fluconazole. This is consistent with other studies that demonstrated fluconazole to be a very well-tolerated medication. 4, 6, 7 Potential problems with the use of fluconazole are cost and selection of resistant isolates. The price of generic nystatin suspension to the Department of Defense is $1.43 for a 60-ml bottle compared with $18.41 for a 35-ml bottle of fluconazole suspension (10 mg/ml) (S Carlisle, personal communication). However, the initial cure rate observed in our fluconazole group would decrease the need for follow-up physician visits. The routine use of fluconazole in the treatment of thrush may result in the selective growth of non-albicans species resistant to fluconazole and the development of fluconazole-resistant strains of C. albicans as documented in studies of immunocompromised patients treated with fluconazole. 6 Indeed three patients in our study did grow non-albicans species from their oral mucosa after treatment with fluconazole. None of these patients developed clinical disease. Darouiche 6 observed that, in immunocompromised patients, these non-albicans species are primarily colonizers of the oral mucosa that almost never cause symptomatic thrush. Up 10% of patients with advanced HIV infection develop resistant Candida species including strains of C. albicans. Darouiche reported that this rate is lower in non-HIV-infected immunocompromised patients and that the development of resistance strains almost always occurs in severely immunosuppressed patients receiving either repeated or prolonged therapy with escalating dosages of fluconazole. 6 Although warranting further investigation with more long term studies, concerns regarding the development of resistant thrush in healthy infants receiving fluconazole may be more theoretical than practical. The small sample size of This study makes it impossible to draw statistically meaningful conclusions beyond clinical and microbiologic cure rates. Larger studies will have to be accomplished to test the validity of our preliminary observations with regard to recurrence rates, complication rates and the efficacy of fluconazole in the treatment of concurrent candidal diaper rash. All of these clinical factors will significantly effect a cost-benefit analysis to determine the potential cost effectiveness of various treatment regimens of fluconazole. Fluconazole is much more expensive than nystatin per unit volume. However, this cost does not reflect the cost of repeat physician visits or lost parental work time for failed initial therapy and these costs should be determined to accurately compare the costs of the two medications. Acknowledgment. Major David Bush, M.D., USAF, MC, Wilford Hall USAF Medical Center, conducted our statistical analyses.