Measurement of human papillomavirus (HPV) DNA load has been suggested as a means for improving the positive predictive value of HPV testing for detecting cervical intraepithelial neoplasia 3 and cancer (CIN3+). We hypothesized that lesions surrounding CIN3+ (especially the extent of associated CIN1) are important determinants of load, which limit its clinical use. To test this hypothesis, we reviewed the pathology of women at enrollment in the Atypical Squamous Cells of Undetermined Significance/Low-Grade Squamous Intraepithelial Lesion Triage Study who were managed at enrollment with oncogenic HPV testing using Hybrid Capture 2 (Digene Corp., Gaithersburg, MD) or colposcopy. We focused our analysis on 133 cases of CIN3+ that were identified, stratified by the number of HPV types detected using a PCR-based method (62 single, 71 multiple infections). Infection with multiple types (P < 0.01) and identification of CIN1 or CIN2 in tissues surrounding CIN3 (P = 0.01) were each significantly associated with higher load (two-way ANOVA). Higher load was associated with more extensive CIN1 (P < 0.01), especially when associated with multiple HPV types. Intensive partial rescreening of enrollment thin-layer slides obtained from women with histopathological CIN3+ revealed that ASC-US/LSIL cells were more numerous than ASC-H/HSIL cells. ASC-US/LSIL cell counts were significantly correlated with load (r = 0.31; P < 0.01), whereas counts of ASC-H/HSIL cells were uncorrelated. We conclude that the extent of histopathological CIN1, number of HPV types, and the number of ASC-US/LSIL cells in exfoliative cervical samples vary greatly among cases of CIN3+ and strongly affect HPV load. These factors severely limit the clinical use of load measurements irrespective of HPV testing methodology.