BACKGROUND Heteroduplex scanning techniques usually detect all heterozygotes, including common variants not of clinical interest. METHODS We conducted high-resolution melting analysis on the 24 exons of the ACVRL1 and ENG genes implicated in hereditary hemorrhagic telangiectasia (HHT). DNA in samples from 13 controls and 19 patients was PCR amplified in the presence of LCGreen I, and all 768 exons melted in an HR-1 instrument. We used 10 wild-type controls to identify common variants, and the remaining samples were blinded, amplified, and analyzed by melting curve normalization and overlay. Unlabeled probes characterized the sequence of common variants. RESULTS Eleven common variants were associated with 8 of the 24 HHT exons, and 96% of normal samples contained at least 1 variant. As a result, the positive predictive value (PPV) of a heterozygous exon was low (31%), even in a population of predominantly HHT patients. However, all common variants produced unique amplicon melting curves that, when considered and eliminated, resulted in a PPV of 100%. In our blinded study, 3 of 19 heterozygous disease-causing variants were missed; however, 2 were clerical errors, and the remaining false negative would have been identified by difference analysis. CONCLUSIONS High-resolution melting analysis is a highly accurate heteroduplex scanning technique. With many exons, however, use of single-sample instruments may lead to clerical errors, and routine use of difference analysis is recommended. Common variants can be identified by their melting curve profiles and genotyped with unlabeled probes, greatly reducing the false-positive results common with scanning techniques.