In resource constrained societies and where HIV is a problem, oral thrush causes significant morbidity. In adults, ketoconazole is used and sometimes oral nystatin. Both drugs are relatively expensive compared to GV solution and ketoconazole has significant side effects especially in association with some of the treatments for HIV related problems.

In children, either GV solutions or nystatin are used, GV is a fraction of the cost of nystatin.

GV at 1% solution discolours the mouth (blue) and in the older child and adult would mark them out as having HIV infections. A much more dilute solution of GV has proved equally effective in vitro and would not carry the same cosmetic problem.

In this study of children, the investigators have compared the 3 solutions, 1% GV, 0.00165% GV and nystatin oral drops - all masked so that they look the same - to see if GV is more effective than nystatin, and to see if the weaker solution of GV is as effective as the stronger solution.

A double blind randomised trial of 2 strengths of GV solution and nystatin oral drops in the treatment of oropharyngeal candidiasis in children.

Children with oral thrush were enrolled from the paediatric wards of the Queen Elizabeth Central Hospital after permission and full information was given to the guardians.

Children of any age up to 14 years were included.

Mothers or guardians gave permission after pre counselling for HIV testing, and a saliva sample collection on enrollment. A full history and examination was carried out. The extent and severity of the candidal infection recorded on oral pictorial graphs and graded.

The child was then prescribed a treatment of A, B or C solution which was introduced into the mouth with a pipette. One ml of the solution was prescribed 3 times a day for 10 days.

The children were reviewed on day 3 to ensure no worsening of the condition and on day 12 when another saliva sample was taken.

A further review was carried out on day 21 of a limited number of children to repeat the saliva test.

Exclusions to the study were children who were already on an antifungal agent or those who had evidence of infection beyond the pharynx into the peritonsillar bed, suggesting the presence of oesophageal infection. These children were prescribed ketoconazole tabs.

If the oral infection was worse on day 3 miconazole gel was prescribed and the study medication stopped.

Sample size to achieve 80% power to detect a difference in failure rates of 20% and 10% (20% in the nystatin group and 10% in the GV groups) is 155 in each group.This assumes an HIV positivity of 50%. As a drop out rate of 20% is expected from death (in some HIV infected patients) or failure to attend for follow up, a sample size of 186 per group is to be recruited. This gives an overall number to be enrolled of 558 patients.

Recruitment has been completed - analysis is in progress.

Inclusion Criteria:

• All children up to 14 years of age with proven oral candidiasis

Exclusion Criteria:

• Children already on an antifungal agent or who had received an antifungal agent in the last week.