To study the discriminatory ability of dacryoscintigraphy in differentiating between patients with epiphora and volunteers using a simple method. Twenty eyes in ten volunteers and 66 eyes in 55 patients with severe epiphora were studied. Dacryoscintigraphy (15 frames of 1 min) was performed after administration of 4 MBq (99m)Tc-pertechnetate (10 mul) in both eyes. By mapping a single region of interest (ROI) over the conjunctival sac we determined T1 (%dose in first minute) and linear clearance rate (LCR, defined as [see text] from the tracer disappearance curve. Reproducibility was determined in volunteers. Conjunctival resorption was determined from completely obstructed systems. Sensitivity and specificity were calculated and receiver operating characteristics (ROC) analysis was conducted. In volunteers mean T1 was 52.8+/-11.9% (95% CI 47.1-58.4%), and LCR was 74.2+/-11.1% (95% CI 69.0-79.4%). Reproducibility was good (mean difference 4.1+/-13.3% for T1 and 0.7+/-17% for LCR). Epiphora patients had clearly higher T1 (82.1+/-15.2%, P<0.0001) and lower LCR (38.9+/-22.5%, P<0.0001) values. Tracer resorption was 24%. Based on ROC analysis 70% for T1 and 50% for LCR were considered optimal cut-off levels to separate patients from volunteers. Sensitivity/specificity were 77/95% for T1 and 71/100% for LCR. T1 and LCR values did not correlate with symptom scores or Anel test results. Dacryoscintigraphy, using LCR and T1 as parameters, is a reliable and objective method to detect tear-flow abnormalities.