A variety of headaches are frequently associated with the occurrence of neck pain. The purpose of this paper was to describe the adherence to diagnostic criteria of a series of patients enrolled on the basis of two clinical criteria: (1) unilateral headache without side-shift, and (2) pain starting in the neck and spreading to the fronto-ocular area. One hundred and thirty-two patients (36 male and 96 female) entered the study. Sixty-two patients were assigned to Group A (patients fulfilling criteria 1 and 2), 40 to Group B (criterion 2 only) and 12 to Group C (criterion 1, only). Eighteen subjects were excluded because X-rays of the neck were not available. Patients were evaluated regardless of whether or not they fell into one or more of the following diagnostic categories: cervicogenic headache (CEH), migraine without aura (M) and headache associated with disorders of the neck (HN) (IHS definitions). Fulfillment of the diagnostic criteria for CEH was found to be particularly frequent in Group A. A higher frequency of CEH diagnosis was found when two criteria were used (Group A) than in Group B (P = 0.001); in the former group a higher mean number of diagnostic criteria for CEH were also present (P = 0.001). Group A patients more frequently presented pain episodes of varying duration or fluctuating, continuous pain and moderate, non-excruciating, non-throbbing pain than Group B patients (P = 0.04 and P = 0.08, respectively). In Group C patients, the frequency of these two criteria was relatively low (17%) especially of the first mentioned variable. The presence of at least five of the seven 'pooled' CEH criteria (present in > or = 50% of the patients) might be deemed a reliable cut-off point, allowing the headache to be diagnosed as 'probable' CEH. If patients fulfilling M or HN criteria in addition to the CEH criteria are added to the 'pure' CEH group a total of 74% of Group A patients may have a CEH picture. The temporal pattern of pain and the quality of pain in Group A showed good sensitivity and specificity (> or = 75) when compared with Group B; therefore, the chances of diagnosing a definite CEH are significantly more frequent in patients presenting with unilateral pain that also begins as a neck pain. Head/neck trauma and radiological abnormalities in the cervical spine were not significantly associated with CEH, M or HN diagnoses. An improvement of the current diagnostic IHS criteria might make it possible to avoid the existing, partial overlap of CEH with HN and M. Extensive use should be made of the GON, and other, blockades in the routine work-up of CEH, both in the differential diagnosis and in the mixed forms (CEH + M, and CEH + HN), in order to improve the efficiency of the current diagnostic system.