A meta-analysis of the Papanicolaou smear and wet mount for the diagnosis of vaginal trichomoniasis

PURPOSE: To obtain reliable estimates of the sensitivity and specificity of the cervical Papanicolaou (Pap) smear and wet mount to diagnose vaginal tr ichomoniasis. METHODS: Articles indexed in MEDLINE (1976-1998) about diagnostic tests for trichomoniasis and their listed references were retrieved. Thirty studies (9,501 patients) that used trichomonas culture as a gold standard were sele cted. Studies were defined as level I if they fulfilled at least two of the following three criteria: consecutive patients were evaluated prospectivel y, the decision to culture was not influenced by test results, and there wa s independent and blind comparison with culture. Studies were classified as level II if one criterion was fulfilled, and as level III otherwise. RESULTS: The pooled sensitivity of the Pap smear for the diagnosis of trich omoniasis among level I studies was 57% (95% confidence interval [CI]: 51% to 63%) and the pooled specificity was 97% (95% CI: 93% to 100%). The likel ihood ratio for a positive Pap smear was 19 among level I studies (range: 8 to 62). The pooled sensitivity of the wet mount among level I studies was 58% (95% CI: 51% to 66%); among level II studies, the sensitivity was 72% ( 95% CI: 62% to 81%), and among level III studies, the sensitivity was 82% ( 95% CI: 67% to 97%). The overall specificity of the wet mount was 99.8%. CONCLUSIONS: A positive Pap smear for trichomonads in settings in which tri chomoniasis is common (prevalence greater than or equal to 20%) requires tr eatment. A positive Pap smear is indeterminate when the prevalence of trich omoniasis is about 10%; thus, clinicians should either confirm the diagnosi s by culture or treat all such patients, recognizing that some patients wil l be treated unnecessarily. A culture should be obtained in women with a po sitive Pap smear who are unlikely to have trichomoniasis (prevalence less t han or equal to 1%). While a positive wet mount is diagnostic, a negative w et mount does not exclude trichomoniasis. Am J Med. 2000;108:301-308. (C) 2 000 by Excerpta Medica, Inc.