The overlap between drinking and smoking does not indicate the presence of an “addictive” or “oral” personality or any other such abstraction, only the concrete reality that conscience and judgment are soluble in ethanol. Alcoholics are fearless in barrooms and automobiles. They’re not careful about safe sex, and they don’t worry about cigarettes, either. Deaths from cirrhosis, accident, suicide, and pharyngeal cancer were also far more common among alcohol abusers. Study findings in this regard are very similar to those from eight other longitudinal studies of premature mortality in alcoholics reviewed elsewhere.12
2. Is alcoholism environmental or genetic? In 1938, the year the Grant Study began, Karl Menninger, America’s most famous psychiatrist, made a bold statement: “The older psychiatrists . . . considered alcoholism to be a hereditary trait. Of course, scarcely any scientist believes so today. Although it’s still a popular theory, alcoholism cannot possibly be a hereditary trait, but for a father to be an alcoholic is an easy way for a son to learn how to effect the retaliation he later feels compelled to inflict.”13
Menninger was wrong. When it comes to alcoholism, genes trump environment. Our data revealed that a family history of alcoholism more than doubled the chance that a Study member would become alcoholic, even when all other suspected etiologies (such as ethnicity, social class, and multiple family problems) were statistically controlled. But children with alcoholic stepparents were not more likely to become alcoholic. Here the data from the Inner City men was even more convincing than that from the Grant men.
Until very recently, most social scientists believed unhappy childhoods to be a cause of alcoholism. Our data showed that the association of childhood environmental weaknesses with future risk of alcohol abuse exactly paralleled the extent of parental alcohol abuse: that is, the more severe a parent’s alcoholism, the more it will be reflected in his child’s environment, and in the severity of alcoholism in the child. Contrary to Menninger’s belief, however, alcoholism does not arise in children in response to an unhappy childhood or even to life with an alcoholic stepparent. It is the fact of an alcoholic biological parent, whether or not the child lives under the same roof, that increases the child’s risk. Of the fifty-one men who had few childhood environmental weaknesses and an alcoholic parent, 27 percent became alcohol-dependent. Of the fifty-six men with many environmental weaknesses and no alcoholic parent, only 5 percent became alcohol-dependent. Alcoholic parents do not have to live with their children to pass on the disease.
The presence of a genetic component does not free us from issues of nature and nurture. While the number of alcoholics in one’s ancestry increases the likelihood of alcohol abuse for genetic reasons, it also increases the likelihood of lifelong teetotalling, presumably for environmental reasons. Almost half of the forty-eight teetotalers of Anglo-Irish-American descent in the Glueck Study had an alcoholic parent.
3. Are alcoholics premorbidly different? This question essentially explores whether alcoholism is a symptom of mental illness or a cause. It has long and widely been considered the former. In the year the Grant Study was conceived, Robert Knight, a prominent psychoanalyst at the Austen Riggs Center, said it directly: “Alcohol addiction is a symptom rather than a disease. . . . There is always an underlying personality disorder evidenced by obvious maladjustment, neurotic character traits, emotional immaturity or infantilism.”14 In 1940 Paul Schilder, an Austrian psychiatric researcher who has given his name to four diseases, concurred. “The chronic alcoholic person is one who from his earliest childhood on has lived in a state of insecurity.”15 Two decades later, E. M. Jellinek, Yale’s great alcoholism scholar, wrote: “In spite of a great diversity in personality structure among alcoholics, there appears in a large proportion of them a low tolerance for tension coupled with an inability to cope with psychological stresses.”16 And in 1980, psychiatrist Michael Selzer wrote more generally in the leading textbook of psychiatry: “Despite occasional disclaimers, alcoholics do not resemble a randomly chosen population.”17 None of these world-renowned experts, however, possessed any prospectively gathered data as to what alcoholics were like before they became alcoholic.
Three premorbid personality types have repeatedly been proposed as causal contributors to alcoholism: the dependent, the depressed, and the sociopathic.18 The Grant Study confirmed none of these hypotheses.
The College alcohol abusers did not exhibit more premorbid evidence of dependent personality disorder than men who remained social drinkers all their lives, and 58 percent of the College men who became alcohol abusers did not lose control of their alcohol use until after age forty-five.
There were men who displayed many dependent traits as young adults and who showed lifelong difficulties with loving, with perseverance, and with postponement of gratification. Such so-called oral-dependent men were also more anxious and more inhibited about expressing aggression. Yet these traits in young adulthood were not significantly more common in future alcohol abusers than in everyone else. Oral-dependent traits did become very common in the College men, however, once they began to abuse alcohol.
It is also true that alcohol-abusing College men were five times more likely to report severe depression than men who did not abuse alcohol. Furthermore, of the thirty-one men who ever appeared to manifest major depressive disorder, fourteen (44 percent) also manifested alcoholism. Following these men for twenty-five years, I received the impression that many of the fourteen had turned to alcohol to relieve their depression. In 1990, however, the longitudinal data were subjected to blind analysis, and that impression proved to be illusion.19 One psychiatrist, blind to age of onset of depression, reviewed each man’s entire record and estimated the year that he first manifested evidence of alcohol abuse. A second psychiatrist, blind to age of onset of alcoholism, reviewed each man’s record and determined the age of onset of major depressive disorder (or probable major depressive disorder). In four of the fourteen cases, the psychiatrist looking for evidence of primary depression believed that the depressive symptoms could be explained entirely by alcohol abuse. In six more cases, the rater noted that the first episode of major depressive disorder had occurred (on average twelve) years after the patient met the criteria of alcohol abuse. In only four cases had a man’s depression actually preceded his alcoholism. Given the prevalence of alcoholism and affective disorder among the 268 men in the College sample, chance alone could account for primary alcoholism and primary depression occurring independently in four cases or even more.
As for the sociopathic personality, alcoholics are somewhat more likely to be premorbidly antisocial than asymptomatic drinkers.20 And some antisocial adolescents initiate alcohol abuse as their antisocial behavior develops. But most alcoholics are not premorbidly antisocial; they become antisocial only after developing alcoholism. The role of sociopathy in alcoholism remains murky and I’ve discussed it extensively elsewhere.21
Two of the three supposed personality antecedents of alcoholism having been dispatched, and the third on hold for the time being, we looked for more general premorbid factors that might predict later alcohol abuse. However, data from the College sample revealed that neither bleak childhoods, childhood psychological problems, nor (more positively) psychological stability in college differentiated future social drinkers from future alcohol abusers.
Far more surprising, most future alcoholics do not appear to differ from future asymptomatic drinkers even in terms of premorbid psychological stability. A hypothesis like that could not even be seriously entertained until prospective studies were available, so compelling was the illusion that unhappy, nervous people turn to alcohol for self-medication, and so unlikely did it seem that depression and the so-called alcoholic personality might be secondary to the disorder of alcoholism. But alcohol in high doses is neither a stimulant nor a tranquilizer; it is the very opposite of a successful drug for these conditions. And alcohol ingestion makes both insomnia and depression worse.22
Table 9.1 Correlation of Alcohol
ism and Dependent Traits with Premorbid Variables
Very significant = p<.001; Significant = p<. 01; NS = Not Significant.
In the Inner City sample, the three childhood variables that most powerfully predicted positive adult mental health—warmth of childhood, freedom from childhood emotional problems, and boyhood competence—did not predict freedom from alcoholism; similarly, the three variables that most powerfully did predict alcoholism—family history of alcoholism, ethnicity, and adolescent behavior problems—did not predict poor future mental health. Alcoholism and poor mental health cannot be yoked invariably together.
In dismissing these factors (unhappy childhoods, multiproblem families, depression, and anxiety) as major causes of alcoholism, I do not mean to say that they are not important. They are always important, and they will make any chronic disease worse. I simply wish to underscore that in a prospective design, when more salient variables like culture and familial alcoholism can be controlled by the choice of sample, premorbid personal and family instability do not make a statistical contribution to the risk of alcoholism. Let me also reiterate that the Inner City men with an alcoholic biological parent but an otherwise stable family were five times as likely to develop alcoholism as men from clearly multiproblem families but without an alcoholic biological parent.
Our data suggested only two areas in which alcoholics appeared to be premorbidly different from asymptomatic drinkers. Before they become abusers, future alcoholics can tolerate relatively larger amounts of alcohol than controls can without hangovers, vomiting, or staggering.23 This difference appears at least in part to be another reflection of the hereditary component of alcoholism.
The second difference is that future alcoholics are more likely to come from environments that tolerate adult drunkenness and discourage children and adolescents from learning safe drinking practices—Ireland and the United States are prime examples. Alcoholics were less likely to come from cultures that permit adolescents to consume low-proof alcoholic beverages at ceremonies and with meals, but condemn drunkenness—Italy, for example. Thus, the parents and grandparents of the alcoholics in our samples were several times more likely to have been born in English-speaking countries than in Mediterranean ones. Controlling for genetic risk, the Irish Study members had very significantly higher rates of alcohol dependence than the Italians. One highly controversial conclusion of the Study is that alcoholism is more affected by parental modeling of how children should drink than by legislation of when children should drink.
4. Should the goal of treatment be abstinence? When the Grant Study began, we knew far more about the natural history of most cancers than we did of alcoholism. Even the clinical course of alcoholism was poorly understood. What actually happens to alcoholics over time—not just those who attend clinics, but the whole constellation of treated and untreated alcoholics? Why does the prevalence of alcohol abuse decline sharply with age?24 Is the explanation for this decline “burnout” (no), or return to asymptomatic drinking (no), or stable abstinence (yes), or high mortality (yes)? Another question is how long abstinence or return-to-controlled-drinking must persist before an individual’s recovery from alcoholism can be considered secure. In smoking and even in cancer, remission must usually endure for five years before relapse is considered unlikely. In alcohol treatment studies, however, investigators often speak of “recovery” after the drinker has been symptom-free for six months or a year. The Harvard Study of Adult Development demonstrates that such a short time frame is not realistic. Only after five years of abstinence can remission from alcoholism be regarded as secure.25 And as with smoking, the remitted alcoholic can almost never return to social drinking.
Table 9.2 summarizes the outcome status for the 148 alcoholics from the College and Inner City cohorts successfully followed until death or age seventy. It clarifies why the numbers of alcoholics seem to decline with age, and why at age seventy only a quarter of the men were still abusing alcohol. Some had been stably abstinent (for a mean of nineteen years), and a few had returned to social drinking, less stably. But almost half of them were dead.
An interesting finding: the alcoholic Inner City men were more likely to become abstinent than the alcoholic College men. Only nine of the College men meeting criteria for alcohol abuse achieved three or more years of abstinence, while 51 alcoholic Inner City men achieved this status.
5. Can “real” alcoholics ever drink safely again? The answer to this question is “Yes, but . . .” My hesitation is based on four caveats: one from the general literature, and three based on Study findings. In fifty years of alcoholism literature, every study that reported the successful return of its clients to social drinking (often making the evening news in the process) has been found to have been in error after ten years’ follow-up.26 One highly publicized cautionary tale was the case of Audrey Kishline. Kishline founded Moderation Management in 1994, but in March of 2000, she swerved out of her lane and killed two occupants of an oncoming car in a drunken driving accident.27
Table 9.2 Outcome Status for Study Alcoholics at Age 70 or at Time of Death
Second, our data revealed that of the men who returned successfully to social drinking, most had only barely met the criteria for a diagnosis of alcoholism. This was true in both the Inner City and College samples. Third, of those who had maintained social drinking for three years or more, half relapsed or ended up seeking abstinence. Fourth, even the successes (the men who at last follow-up have had no further problems due to alcohol) often found it impossible to return to the carefree use of alcohol manifested by most social drinkers.
The year 1977 was the end of the age twenty-to-forty-seven follow-up period for the Inner City men. Twenty-one of them had achieved stable abstinence of three years or more, and twenty-two had returned to controlled drinking for three years or more.28 Continued follow-up until 1992 revealed that eighteen (86 percent) of the twenty-one abstinent men maintained their abstinence until age sixty or until death.29 The range of known abstinence for these eighteen men was three to thirty-seven years; the mean length of abstinence was twenty years. In contrast, over the same fifteen-year follow-up period, seven of the twenty-two men with three or more years of controlled drinking—almost a third—relapsed to sustained alcohol abuse, with a mean duration of alleged controlled drinking before relapse of twelve years. Three of the twenty-two sought sustained abstinence, four withdrew from the Study, and three, because of the brevity of their abuse, were reclassified as non-alcoholic. Thus, in 1992 or at death, only five of the twenty-two men were still believed to be true alcoholics drinking in a controlled fashion. Figure 9.2 illustrates the relation of the number of alcohol-related problems to the likelihood of a return to social drinking.
Figure 9.2 Relation between number of alcohol-related problems and likelihood of a return to social drinking.
6. How do we prevent relapse? Observers of human nature have long puzzled over the phenomenon of individuals who become “converted” to a new religion or otherwise abruptly alter their life course. The Study sought to discover whether in the case of alcoholism this usually occurred through clinical treatment (no), through willpower (no), through becoming sick and tired of being sick and tired (no), or through nonclinical factors known to affect relapse prevention in other addictions (yes). Alas, going on the wagon is more often evidence of alcoholism than of recovery, just as dieting is more often evidence of obesity than of slimness. Nevertheless, prolonged relapse prevention is key, both in alcoholism recovery and in obesity. Thus, our continued study of two cohorts of alcoholics over fifty years provided a naturalistic way of uncovering who did manage to prevent relapse.30
In the vast majority of alcoholics, we found that while counseling, detoxification, and even hospitalization could be temporary lifesavers, they did not change the natural history of the disease. As with diabetes and obesity, permanent changes in self-care are the only way to extend life over the long run. Table 9.3 covers four well-established means of relapse prevention in
any addiction. Stably abstinent Study members on average employed two out of four of them during their first year of abstinence. About half of the ever-abstinent men found an alternative for alcohol, and some found more than one. Substitute dependencies varied from candy binges (five men) to benzodiazepines like Valium or Librium (five men); from compulsively helping others (two men) to returning to dependence upon parents (two men); from marijuana (two men) to mystical belief, prayer, and meditation (five men); from compulsive work or hobbies (nine men) to compulsive gambling (two men); from compulsive eating (three men) to chain smoking (seven men).
Table 9.3 Factors Associated with Abstinence for a Year or More for 49 Untreated and 29 Clinic-Treated Inner City Men
Another half of the men achieved relapse prevention through compulsory supervision or behavior modification. By that I mean the presence of factors independent of willpower that systematically altered the consequences of alcohol abuse. This was often probation, or a painful medical consequence that, as AA members say, “keeps the memory green.”
Two other means of relapse prevention were employed by almost half the men. One was to get involved in an inspirational group (for our sample, usually AA). The second was to find new relationships; sometimes these were love relationships and sometimes opportunities to help needy others. But they were always relationships with people toward whom the alcoholic did not feel guilty for past misdeeds. In their review of the literature on remission from abuse of tobacco, food, opiates, and alcohol, Stall and Biernacki identified these same four factors.31 The four factors in Table 9.3 appeared to be the most important keys to relapse prevention; only 30 percent of the men resorted to clinic attendance or hospitalization during their first year of abstinence.
Triumphs of Experience: The Men of the Harvard Grant Study Page 30