A prospective study of oral lesions and their predictive value for progression of HIV disease

OBJECTIVE: This report evaluates and compares individual oral lesions and combinations of lesions in predicting progression-free survival in a seroprevalent cohort of men and women with HIV infection. DESIGN: This was a prospective study of HIV-infected patients, initially AIDS-free, followed for approximately 30 months. SETTING: Patients were volunteers examined at an academic medical center and at an inner-city hospital in New York. Participants identified themselves as homosexual men or as injection drug users (IDU). OUTCOME MEASURES: The primary outcome being assessed is time from a baseline oral examination until the development of an AIDS-defining condition or death from any cause within 12 months of the last study visit. Correlation is measured by relative risk (RR). RESULTS: While oral lesions were not predictive of progression among subjects with CD4 > or = 200, they were highly predictive of progression among those with CD4 < 200. For subjects with CD4 < 200, the only individual lesion that was significantly associated with progression-free survival was oral candidiasis (RR = 4.12, P = 0.009). Positivity for one or more lesions in a set demonstrated greater prognostic value among those with CD4 < 200, with RR's of 6.03 (P = 0.018) for the set consisting of oral candidiasis, hairy leukoplakia, and necrotizing ulcerative gingivitis (NUG), and 8.77 (P = 0.036) for the set consisting of the above lesions plus linear gingival erythema (LGE). Analysis by cohort suggested that the improvement in correlation was stronger in homosexual men than in IDU, but this question could not be resolved conclusively with these data. CONCLUSIONS: Lesion sets might be better prognosticators of progression-free survival than individual lesions among HIV-infected subjects with CD4 < 200. Prognostic value of the core lesion set (oral candidiasis and hairy leukoplakia) was enhanced by the addition of other lesions (NUG and LGE) not usually included in HIV staging systems. These results suggest that staging systems for HIV might be improved by the inclusion of other, survival-related oral lesions.