After the first descriptions of late onset alcoholism in the early 1970s, no systematic attention followed for more than a decade. In the past 10 years, however, that has changed, and this report seeks to summarize recently acquired knowledge about late onset problem drinking, including the author's own work on this phenomenon. The incidence of new alcoholism cases does decline with age but remains appreciable into the late sixties. Samples of ageing alcoholics in treatment show that as many as 1/4 to 213 of cases had onset after age 60. Risk factors for late onset alcoholism include female gender, higher socioeconomic status and (in some but not all studies) life stressors, but neither psychiatric comorbidity nor positive family history of alcoholism appears to contribute in a majority of cases. Compared to longstanding alcoholics, late onset cases tend to be milder and more circumscribed, and they may also fluctuate more, with an apparently high likelihood of spontaneous remission, at least over the short term. These characteristics have implications for the use of brief and informal interventions to prevent and reduce late onset problem drinking, but also suggest caution in interpreting the response of late onset cases to treatment in uncontrolled studies. Present knowledge of late onset alcoholism is fragmentary; more systematic clinical research on its characteristics and treatment is needed.
KEY woms-Geriatric psychiatry, geriatric alcoholism, late onset drinking problems.
Studies of ageing alcoholics in treatment, published in the early 1970s both in the UK (Rosin and Glatt, 1971) and the US (Zimberg, 1974), suggested that about 113 of these patients had onset of their first drinking problem in late middle age or beyond, a finding that very much cut against the grain of conventional wisdom at the time. For in the 1960s a perspective had emerged, based on influential papers by Winick (1962) on opioid addiction and by Drew (1968) on alcoholism, that the 'window of risk'-the age span during which the new development of any sort of addictive problem could be expected to occur-was more or less closed forever well before age 50 in nearly all cases.
This point of view has been quite durable, finding * Based on a presentation at a symposium entitled 'Physiologic and Clinical Research on Aging and Alcohol Use' at the annual meeting of the Research Society on Alcoholism,