Clinical and microscopic diagnosis of vaginal yeast infection: a prospective analysis

To evaluate the accuracy of clinical findings and direct microscopy in the diagnosis of yeast vaginitis.

Prospective.

Urban teaching hospital emergency department and walk-in clinic.

Seventy-one consecutive nonmenstruating women with a presenting complaint of vaginal discharge, itching, or pain.

Trained emergency physicians and nurse practitioners, blinded to culture results, collected clinical information, examined vaginal secretions, and obtained yeast cultures. Treatment was based on clinical impression and not altered by the study. Associations between clinical and laboratory parameters and yeast culture results were evaluated.

Twenty-three patients (32.4%) had positive yeast cultures. Accepted risk factors were seen in patients with and without culture-positive yeast vaginitis. Clinical criteria for yeast vaginitis--including pruritus, caseous discharge, perineal edema or erythema, and patient self-diagnosis--were more common in women with yeast vaginitis but were seen with other causes of vaginitis. Watery discharge and amine odor (negative "whiff" test) were good predictors of an alternative diagnosis, but, again, accuracy was poor. Gram stain showing yeast was the most accurate laboratory method of diagnosis. Saline microscopy, potassium hydroxide, and methylene blue were all equal and only moderately accurate, with a significant proportion of false-negative and false-positive results.

Standard criteria for diagnosis of monilial vaginitis are relatively unreliable. Gram stain, absence of watery discharge, and patient self-diagnosis of "another yeast infection" are the best independent predictors of a positive culture.