• head injury
• Alzheimer’s disease

In this issue (see pp 857–862) Fleminger and colleagues¹ provide a timely update on the status of the link between head injury and Alzheimer’s disease. Forty three case–control studies are identified in their review, and 15 met strict methodological criteria for inclusion in the overall meta-analysis. The balance of evidence, at least in males, is now firmly for the existence of an association.

The issue is important both theoretically and in clinical practice. Clinicians are increasingly called on to give advice on the link between head injury and Alzheimer’s disease. Should advice about the increased risk of developing Alzheimer’s disease be given routinely to head injured people? Advice might encourage the person to avoid the risk of further injury and aid them in seeking help early if symptoms of the disease became apparent. Clinical opinion is also sometimes sought on whether previous head injury is a major contributing cause in those who develop Alzheimer’s disease. In a landmark case in November 2002, a coroner’s court in Burton-upon-Trent recorded a verdict of death by industrial disease in the case of Jeff Astle. This former England footballer had died at the age of 59 after having developed Alzheimer’s disease. In this case neuropathological evidence of brain injury that was apparently caused by heading the ball was of key importance in the decision concerning the cause of death.

The epidemiological studies reviewed by Fleminger et al suffer from inherent weaknesses. Crucially, all the studies described rely on the report of an informant to ascertain whether a head injury causing loss of consciousness had occurred in the lifetime of the individual. Recall bias and inaccuracy are thus major limitations. In these studies it is impossible to establish reliably the nature and severity of the original injury. Such uncertainty may account for some of the inconsistencies in published reports. The difference between men and women, for example, may arise because men have typically had more severe injuries.

More generally the case–control approach is too imprecise to answer questions remaining concerning the link. For example, does severity of head injury affect the likelihood of developing the Alzheimer’s disease, and does this interact with genotype?² Does repetitive head injury lead to earlier onset?³ It is to be hoped that this review by Fleminger et al will help stimulate the follow up of cohorts of patients with well documented head injuries. The emphasis should now be on understanding the link better rather than on documenting its existence.