The aims of tumor follow-up in head and neck cancer patients are (1) evaluation of therapeutic efficacy, (2) management of impairments, (3) detection of new tumor manifestations, and (4) psychosocial care. In general standardized 5-year-protocols are used for all such patients. However, it is questionable whether a rigid follow-up schedule is optimal for a very heterogeneous tumor population. Therefore 603 patients with sqamous cell carcinoma of the oral cavity, pharynx or larynx, or with cervical metastasis from an unknown primary site (CUP syndrome), who had been diagnosed and treated curatively by an operation with or without radiotherapy (n = 523) or just by radio(chemo)therapy (n = 80) between 1985 and 1994, and who had been followed-up regularly according to a standardized plan, were worked-up retrospectively. Data were evaluated for the manifestation and prognosis of curable new tumor manifestations as well as for tumor-specific factors likely to select groups which should be followed more or less intensively. Within a 5-year follow-up period new tumor growth was detected in 152/603 (25%) patients: 79 local and 31 regional recurrences, 18 systemic metastases and 24 second primary cancers. Where follow-up was extended beyond the 5th year, 168/603 (28%) patients presented a new tumor manifestation. One hundred and sixteen of the 152 (28%) patients had another operation with or without radiotherapy or had radio(chemo)therapy alone. So far 18/116 (14%) patients have survived their new tumor manifestation for more than 5 years and 30/116 for more than 2 years. Tumor-specific data on the initial tumors (T stage, N stage, site) did not indicate the risk of a new tumor manifestation, but 87% of patients who survived their new tumor manifestation for more than 2 years initially had T1 or T2 tumors and only 30% initially had N+ necks. Occurrence of distant metastasis or a second primary outside the head and neck region limited survival to < or = 2 years after detection. In terms of survival, follow-up efforts should therefore concentrate on detection of locoregional recurrence, particularly if an option for further curative local therapy exists. The limited success of detection of new tumor manifestations in terms of survival does not justify a reduction in tumor-follow-up examinations, since the benefit of the other efforts cannot be determined from survival figures.