AIM OF THE STUDY: To evaluate correlations between pre-operating epidemiological, clinical and diagnostic data and relapses concerning to patients operated on for benign thyroid nodules. MATERIAL AND METHODS: The AA. carried out a new ultrasonographic exam in 155 patients selected at random among 1012 treated by thyroid lobectomy between September 1976 and December 2002; 107 relapses were found (69%). The obtained data are analysed by χ2 test and a significant correlation between recurrence and data preceding first operation was found: non-homogeneous thyroid structure, multiple nodules, presence of peri-lesional halo, anti-thyroid antibodies, and non-adenomatous histological structure. CONCLUSION: The AA. conclude that it is possible to establish a “high risk” population; in spite of that, the high risk of relapse related to low surgical risk of total thyroidectomy suggest a more aggressive surgical procedure at the moment of first operation.