Opinions are divided on the use of the term cervicogenic headache (CGH) in cases with no evidence of cervical damage. According to Sjaastad et al. (1990), CGH is diagnosed from three features: (1) unilateral headache triggered by head/neck movements or posture; (2) unilateral headache triggered by pressure on the neck; (3) unilateral headache spreading to the neck and the homolateral shoulder/arm. Other characteristics are not essential for CGH diagnosis, including pain improvement after greater occipital nerve (GON)/C2 block. However, other authors give different definitions of CGH, and this may explain why reported frequencies for this headache vary so widely. In this paper we critically review the major diagnostic criteria of Sjaastad et al. for CGH in the light of clinical studies conducted at our institute and other literature findings. In a study of 500 headaches we found only two patients with unilateral headache triggered by head/ neck movements or posture, and no cases of neck pressure-induced headache. No clear-cut criteria are given in the literature for differentiating CGH trigger points from myofascial trigger points. In another study of 440 primary headache patients we found that in the unilateral long-lasting headache group (64 migraines and 10 tension-type headaches), a pain involving the occiput/neck was present in 30 migraine and seven tension headache patients; thus, according to the CGH major criteria, 10% (30/307) of 'migraines' and 7% (7/96) of 'tension headaches' could be diagnosed as CGH. However, one cannot exclude that the association of unilateral pain with posterior irradiation is due to the high prevalence of migraine, tension-type headache and chronic neck pain. The relation between CGH and whip-lash injury has been put in doubt by a recent study which found no difference in headache frequency between trauma and control groups and reported no specific headache pattern in the trauma group. Other reports suggest that, when it occurs, CGH usually disappears within a year of whip-lash, throwing doubt on the appropriateness of surgery for post-traumatic CGH. The lack of specificity of GON/C2 block as a treatment for CGH adds further difficulties to the diagnosis of this headache. We conclude that, although neck structures play a role in the pathophysiology of some headaches, clinical patterns indicating a neck-headache relationship have still not been adequately defined. We believe that further rigorous studies are needed to definitively confirm the validity of CGH as a nosological entity.