Some diseases, such as Alzheimer’s and type II diabetes, have what is called an "age effect." This means that rates of illness increase as individuals become older. For other diseases such as alcohol dependence or depression, however, this conceptualization becomes problematic as lifetime rates in younger individuals often exceed those found in older individuals. A study in the January issue of Alcoholism: Clinical & Experimental Research confirms that more-recently born individuals have higher rates of alcohol dependence. These findings, however, may be influenced by underreporting of problems among the elderly, and early deaths due to alcoholism.
"I am a mathematician who works on genetic and family analyses," said John P. Rice, professor of mathematics in psychiatry at Washington University School of Medicine, and corresponding author for the study. "One of the first steps in studying a disorder is to determine how familial it is. This involves comparing rates in the relatives of cases, also called probands, with rates in the general population. We noticed that the rates of alcoholism were considerably higher in the siblings of the probands compared to the parents. In the process of examining these rates, we noticed the strong influence of year of birth."
Rice and his colleagues decided to examine the effects of age and date of birth on rates of alcohol dependence. They used data collected from the multi-site Collaborative Study on the Genetics of Alcoholism (COGA) to examine three groups: a relative sample (4,099 interviewed parents and siblings of alcohol-dependent individuals born prior to 1979); a control sample (1,054 members of families chosen without regard to a diagnosis of alcohol dependence); and a family history sample (6,159 parents and siblings older than 17 yeas of age, dead or alive, born prior to 1970). Several different data analysis techniques were used.
"There were three key findings," said Rice. "One, we observed much higher rates in more-recently born individuals; two, analysis of family history information indicated a higher rate of alcoholism in relatives who were deceased; and three, co-morbid diagnoses of antisocial personality or depression, as well as cigarette smoking, were predictors of risk. The family history data suggest that alcoholics may die younger, so that a bias is introduced when only living individuals are surveyed. This doesn't totally explain the cohort differences, but may be a large contributor."
"Rice’s study serves an important purpose by identifying the bias he suggests while still confirming a cohort effect," said Deborah S. Hasin, professor of clinical public health at Columbia University. "Identifying changes over time, or birth cohorts, in rates of alcoholism shows that factors affecting the occurrence or chronicity of alcoholism are changing. This then presents an opportunity to identify the changing factors, as well as providing information that can be used in policy, such as planning prevention measures or provision of services."
"A major question that remains," noted Rice, "is whether these findings suggest an ‘epidemic’ of alcoholism in the young, or whether they can be explained by the elderly underreporting their alcohol experience, or high rates of alcoholism among deceased relatives. A special strength of our COGA study is that we have diagnostic information on family members who are deceased; most surveys do not have these data." Despite the apparent influence of these two variables on rates of dependence among the young, Rice said the high rates were nonetheless "troubling."
Rice called for continuing examination of date of birth in relation to rates of alcohol dependence. "There is a strong statistical relationship, no matter what the source," he said. "The value of using birth cohorts is similar to that of using gender as a covariate, since we know it is important." For example, he said, "the rate of alcoholism in female relatives born between 1950 and 1959, as well as those born between 1960 and 1979, was about 30%, compared to 4.5% in female relatives born before 1929. In the controls, the rate was 13% in recently-born women, compared to 2.3% in women born during the 1930's. These are dramatic differences. A similar effect is seen in the men, but not as pronounced."