Current issues of ACP Journal Club are published in Annals of Internal Medicine


Nystatin did not alter systemic symptoms but improved vaginal symptoms in women with presumed candidiasis hypersensitivity syndrome

ACP J Club. 1991 Mar-Apr;114:49. doi:10.7326/ACPJC-1991-114-2-049

Source Citation

Dismukes WE, Wade JS, Lee JY, Dockery BK, Hain JD. A randomized, double-blind trial of nystatin therapy for the candidiasis hypersensitivity syndrome. N Engl J Med. 1990;323:1717-23.



To compare oral and vaginal nystatin with placebo for the treatment of premenopausal women with presumed candidiasis hypersensitivity syndrome.


Randomized, double-blind, crossover trial.


University of Alabama at Birmingham School of Medicine. The method and clinical site of recruitment not specified.


Patients were premenopausal women, aged 21 to 40 years, with a history of candida vaginitis that had been aggravated by antibiotics and had responded to nystatin or other antifungal treatments. Patients had at least 3 of 5 additional features thought to be common in patients with the candidiasis hypersensitivity syndrome. Of 50 patients recruited, 42 (84%) completed at least 1 treatment period.


During the 32 weeks of the study, each patient followed 4 different treatment regimens, for 8 weeks each, including each of the 4 possible combinations of oral nystatin (starting at 500 000 units 4 times daily, doubled after 2 weeks and again after 4 weeks) or placebo with vaginal nystatin (100 000 units once daily) or placebo.

Main outcome measures

Patients completed general and disorder-specific symptom questionnaires and were examined every 4 weeks, when fungal cultures were taken.

Main results

The 3 active-treatment regimens and the placebo regimen all reduced vaginal symptoms (P < 0.001); nystatin regimens were more effective than placebo (P < 0.001). For systemic, nonvaginal symptoms, all 4 treatment regimens reduced symptoms (P < 0.001), but the active regimens were no more effective than the placebo regimen (mean difference between all 3 active regimens and placebo was 2%, 95% CI -3% to 7%). For overall symptoms, the all-nystatin regimen was better than the all-placebo regimen (P < 0.013), because of the effect on vaginal symptoms and on 4 of 15 nonvaginal symptoms. There were significant carryover effects of all regimens on the overall symptom score (P < 0.04). There were no effects of the treatments on physical findings. The treatments did not affect the rate of positive fungal cultures.


In women with presumed candidiasis hypersensitivity syndrome, nystatin did not alter systemic symptoms attributed to the syndrome, but did improve vaginal symptoms.

Sources of funding: Critical Illness Research Foundation; National Institutes of Health; National Institutes of Allergy and Infectious Diseases.

Address for article reprint: Dr. W.E. Dismukes, Division of Infectious Diseases, Department of Medicine, University of Alabama Medical Center, Birmingham, AL 35294, USA.


This study stems from controversy between the medical community and the lay public about whether this syndrome is real and treatable, as noted in Bennett's editorial (1). For the medical community, the syndrome lacks "unambiguous definition," laboratory tests are unhelpful, and most published reports are in the lay press and are anecdotal, with no randomized controlled trials to evaluate therapies. The lay press accuses physicians of failure to listen to patients about how they feel, fear of assessing atypical therapies, and negligence in studying the illness.

This study by Dismukes and colleagues attempts to objectively evaluate oral or vaginal antifungal therapy compared with placebo by using presumptive diagnostic criteria. The methodology is strong; it comprises a priori power calculations, randomization, blinded evaluation, and crossover strategy. The results indicate a placebo effect or spontaneous improvement for both vaginal and systemic symptoms, further improvement in vaginal symptoms with antifungal therapy, but no significant effect of antifungal therapy compared with placebo on systemic symptoms.

However, there are several possible explanations for a negative study. If diagnostic criteria are not precise, then all patients may not have had the syndrome. No data were given about the reliability of the symptom questionnaires or observer variability in the physical examinations. The significant carryover effect among regimens suggests an inadequate washout period. The study could not control for diet, which is thought by some to be an important factor in exacerbations. Of 42 patients, 13 received systemic antibiotics for up to 14 days, which might affect exacerbations.

The authors should be credited for a well-designed study of rational therapy for this controversial syndrome, but any of these factors might have biased the study toward its finding of no difference in systemic symptoms among regimens.

Val Lawrence, MD
The University of TexasSan Antonio, Texas, USA


1. Bennett JE. Searching for the yeast connection. N Engl J Med. 1990;323:1766-7.