Objectives: To assess the prevalence and distribution of bleeding on probing (BoP) in a cohort of periodontally diseased patients, evaluate the influence of patient-related predictors on BoP variability, and characterize two subgroups of patients with a significantly different BoP score. Method and Materials: Six hundred and one patients were retrospectively included. The percentage of BoP-positive sites was calculated for the entire dentition (%BoPtotal) and within different areas of the dentition. The influence of patient-related predictors on intersubject variability in %BoPtotal was evaluated. Two subgroups of patients, belonging to the lowest (LB, low bleeding) and highest (high bleeding, HB) quartile of patient distribution according to %BoPtotal, were identified. Results: Median %BoPtotal was 26.4%, with 99% of the population showing at least one BoP-positive site. %BoPtotal was positively correlated with patient age and number of sites with pocket probing depth (PPD) >= 5 mm, and significantly more prevalent in mandibular vs maxillary sites, posterior vs anterior sites, interproximal vs oral/buccal sites, and sites with PPD >= 5 mm vs PPD = 4 mm. LB and HB patients were significantly different for %BoPtotal, as well as in sites with PPD = 4 mm and PPD >= 5 mm, but not for sex, age, smoking status, daily cigarette consumption, diabetic status, and number of sites with PPD >= 5 mm. Conclusions: BoP was highly prevalent and variable in both patients and different areas of the dentition; patient-related factors and site-specific characteristics (such as age, number of periodontal pockets, probing depth, tooth type, and aspects) seem to be partly responsible for the individual variation in BoP; and HB and LB were not significantly different with respect to the investigated predictors. Palabras clave: bleeding on probing, gingival inflammation, periodontal disease