The Thirteenth Step: Addiction in the Age of Brain Science

Anyone concerned with treating drinking problems must find that his patients often tell him more than is in the textbooks. Each tells a different story, but there are also repeated patterns. Much of the varied experience that is recounted can be interpreted as the patient’s astute observation of the alcohol dependence syndrome—a condition certainly far better described by the average alcoholic than in any book.

DEPENDENCE, ABUSE, substance use disorders, addiction. There are so many terms, used by professionals and laypeople alike, that do not necessarily have a clear meaning. Each attempts to capture substance use that is harmful, but drugs can be harmful in different ways. To build an understanding of the kind of substance use that gives rise to clinical problems, we need to know what we are talking about.

So let’s step back a little. A desire to chemically alter brain function and the mind states it produces is probably as old as humankind itself. We know that the ancient Greeks considered the knowledge of how to ferment grape juice into wine to be a gift from the god Bacchus and held annual celebrations to commemorate it. An even more striking example comes from the Russian tundra, where temperatures never rise high enough to allow fermentation of alcoholic beverages. The nomads of these lands devoted precious time and energy to seeking out mushrooms containing substances that, when ingested, produce a state of intoxication dominated by confusion, a dry mouth, and constipation.¹ A feature of this drug is that it is excreted in the urine without being chemically degraded. Because of that, it is said that toward the end of a long winter, a measure of an adult male’s urine could be exchanged for a reindeer cow, which must have represented a fortune.

Clearly the amount of effort an individual is willing to expend to obtain a drug tells us something important about its motivational value. We will see that this applies to humans and animals alike, allowing scientists, under controlled conditions in the laboratory, to assess motivation to obtain a drug. I hasten to say that the example of the mushrooms, while colorful, is not the most didactic. The nomads did not have access to other drugs. When alternatives are present, neither people nor animals will in fact work for this particular drug class. That observation, however, may be useful in itself. It tells us that the motivational value of a drug, or, for that matter, of anything, is not absolute or constant. Instead it is determined by many things. Among these, the choices available to the individual at the time are an important environmental variable.²

Little has changed in modern times. Mind-altering or “psychotropic” drug effects remain sought after. The means to achieve these effects vary and are in part determined by culture and availability, illustrating another set of important environmental variables. Many efforts to reduce drug use by educating the public are based on the assumption that substance use is a consequence of ignorance. For the most part, that is simply not the case. In modern Western countries, where the public is better educated on the potential consequences of alcohol use than ever in the history of our species, between 50 and 90 percent of the adult population consumes alcohol, and a significant minority develop problems because of their use. In other contemporary societies, various formulations of the hemp species Cannabis sativa or indica, or the opium poppy, Papaver somniferum, remain in widespread use. And nicotine use is on the rise in the third world as people become more affluent and better educated. It seems clear that whatever these drugs do, large numbers of people are willing to take them, even though they know the risks. What accounts for this willingness, of course, are the pharmacological, mind-altering properties of the substances that most people are after.

But here we must pause. Most people enjoy the relaxed social inhibitions that result from having a couple of drinks at a cocktail party. Most people who sample cocaine experience euphoria. Yet most people who try these drugs will not get in trouble because of their substance use. This is a fundamental issue when we try to understand the people who do develop problems, get sick, and in many cases die as a consequence of substance use. No matter which drug we look at, only a minority of people using it will develop medical or psychological problems. Meanwhile, a majority will be able to occasionally sample the same drug, or engage in intermittent recreational use over decades, without measurable harm.

Work published by James Anthony, at the time at Johns Hopkins University in Baltimore, uses data from two large epidemiological studies, and provides a perspective on this phenomenon.³ These studies examined large numbers of people from the general population and focused on those who had ever used potentially addictive substances or who were still active users, respectively. Anthony and his colleagues applied standard clinical diagnostic criteria to determine the fraction of those people who would qualify for a diagnosis of “dependence,” the official diagnostic category (defined later in this chapter). Among active users, the numbers ranged between 8 and 18 percent. The exception was nicotine, for which 34 percent of active users qualified for a diagnosis of dependence. As mentioned in the previous chapter, this does not necessarily imply the counterintuitive conclusion that is sometimes drawn from these data—that nicotine is more addictive than all other drugs. The barriers to obtaining and using cocaine are clearly much higher than those to obtaining nicotine. The important message here is that even among active users of the most addictive drugs, such as cocaine or heroin, only a minority develop a particularly destructive relationship with the drug. And in most cases, those who do develop this relationship take a long time—years or decades—to do so.

These observations in turn raise an important and much debated question. Is there really any important difference between people who develop a destructive relationship with a drug and those who do not? Is there, as some would have it, a line, however bright, the crossing of which signifies an important transition and justifies distinguishing those who have crossed it from those who have not? Or is all substance use in principle equal, and harmful in varying degrees that depend only on the amounts used and the social, economic, or medical circumstances of the individual?

The medical profession comes down firmly on one side of this fundamental question. Someone is either healthy or ill, and in the latter case there had better be a diagnostic label for the disease or the doctor may not get reimbursed for its treatment. Indeed, the World Health Organization (WHO), in the tenth revision of its International Classification of Diseases and Health Problems (ICD-10) from 1992,⁴ defines a “substance dependence syndrome”⁵ that is similar to that described in the classic paper by Edwards and Gross quoted at the outset of this chapter. This syndrome is described as a cluster of physiological, behavioral, and cognitive phenomena in which the use of a substance or a class of substances takes on a much higher priority for an individual than other behaviors that once had greater value. A central characteristic of the dependence syndrome is said to be a strong, often overpowering desire to obtain and use the psychoactive drug, which for the rest of this book I will refer to as “craving.” Dependence, as defined by the WHO, refers to both psychological and physical phenomena. The criteria refer to “psychological dependence” as the experience of impaired control over drinking or drug use. “Physiological” or simply “physical” dependence is used to denote tolerance and withdrawal symptoms.⁶

Well, the doctor is always right, so what else is there to say? Quite a bit, of course, as anyone familiar with the history and the controversies of this field over the years well knows. For starters, the medical profession cannot have the final say on this subject. Although doctors devote more time than they may realize to treating medical consequences of drug and alcohol use, they are not a major presence in treating whatever it is that makes people engage in this use in the first place. For better and worse, psychologists, counselors, and mutual-help groups play a much greater role than doctors in treating people with substance use problems. Increased involvement of the medical profession is certainly needed, in particular as medications become an increasingly important component of an effective treatment package. But there will always be a major need for the contributions from other professions, and all of us will always need to establish an alliance and a dialogue with the patients themselves. So leaving scientific and philosophical issues aside, we have an important need for a common language, simply from the standpoint of clinical practicality. In seeking this common language, it is necessary to understand that for many sociologists and psychologists, a categorical view such as that espoused by the medical profession is quite foreign and problematic. A philosophical discussion of this fundamental yet controversial issue can quickly become involved, and both sides have valid arguments. In the end, my view is that whether there is a difference between people who develop a destructive relationship with a drug and those who do not needs to be determined by clinical utility. If there are certain characteristics that tend to occur together in people with particular treatment needs, or in whom similar mechanisms are at play that could be targeted by current or future treatments, then it is practical to have a specifically labeled category (or group of categories) for those people.

Almost all drugs that are abused by humans are also self-administered by experimental animals.⁷ The classical model involves placing a rat or mouse in a box with levers that can be pressed or buttons that can be poked using the nose. When the animal presses a lever or pokes a button, a pump is started and delivers a dose of a drug as a reward—through an intravenous catheter in the case of cocaine or heroin, or into a little cup in the case of alcohol. Animals readily lever press or nose poke for drugs such as cocaine, heroin, or alcohol, so these drugs can be called rewarding. Or at least, as a stricter experimental psychologist would probably prefer, they are reinforcing, meaning that the behavior that led their delivery becomes more frequent than in the absence of the outcome.

But here is a challenge. Almost all rats or mice will readily and within days self-administer heroin or cocaine. Alcohol takes a bit longer, in part because animals initially dislike its taste, and also because it takes longer for the brain effects caused by alcohol to be experienced, making it harder for the animal to learn that they are a consequence of its behavior. Nevertheless, within a few weeks, most rats and mice will self-administer alcohol as well.

Self-administration, studied in this way, has become the standard model by which scientists study brain mechanisms through which abused drugs work and test medications that might be helpful in treating their use. But doesn’t that seem odd? I just went over human research indicating that, by clinical diagnostic criteria, only a minority of active users of any substance will develop substance dependence, and that this takes them many years, sometimes decades. Are the diagnostic criteria arbitrary after all?

Probably not. A few years back, two different laboratories, one in France and one in the United Kingdom, evaluated some of the core features defined by the established clinical diagnostic criteria for drug dependence in animal experiments that used the self-administration model in a particularly clever way.⁸ In one of these studies, which took note of the fact that clinical dependence takes time to develop, rats were first allowed to self-administer cocaine for an unusually long period of time. A month of forced abstinence followed the self-administration phase. Finally, after the forced abstinence, the animals were tested for relapse to drug seeking, using a model I will describe in more detail in a coming chapter. Although all the animals had been given the same access to cocaine during the self-administration phase, about one in five showed an interesting constellation of features. While still in the self-administration phase, they had been particularly persistent in their drug seeking, nose poking even during breaks in the sessions, when the drug was not available. It was as if they could not stop themselves. They had also continued their drug taking even when this led to adverse consequences, in this case a mild electric shock that made other animals stop their self-administration. Finally, they had exhibited the highest motivation to obtain the drug, measured as the maximal number of nose pokes they were willing to make to obtain a dose of cocaine. The animals showing these three characteristics were also those that were particularly prone to relapse after the month of abstinence. With about one in five rats showing this special pattern of drug use, the fraction was clearly similar to that seen in human epidemiological studies. This pattern of drug use could be called compulsive because the animals no longer seemed able to flexibly change their behavior the way their fellow subjects were.

If compulsive drug use, as characterized by these features, is a hallmark of clinically important substance use syndromes, then it would seem clear that these syndromes are primarily conditions of malfunctioning motivational machinery and control of behavior, reflecting processes in the brain, even though what brings the patient to the doctor may be a cirrhotic liver or some other damaged organ.⁹ In fact, the most striking observation from many years of working with people who seek treatment for alcohol or drug problems is that their motivational machinery somehow seems broken. The job of this sophisticated brain circuitry is to guide behavior, so that a person can be reasonably successful in the pursuit of his or her goals and, if one believes in such a thing, of happiness.¹⁰ As we know from our own lives, this machinery is by no means perfect. But it generally does not make simple, disastrous blunders when better choices are readily and immediately available. Yet that is what happens over and over again for people like Eric in the previous chapter. Contrary to widespread public belief, this is not because of any moral flaw in these people or a lack of understanding on their part. When asked, no patient I ever met really valued getting drunk over having a job or a home. Yet they would easily and repeatedly lose both these things and then some for an opportunity to get drunk or high. Likewise, unless severely depressed,¹¹ patients don’t want to die when a decent life is an option. Yet like Eric, many patients continue to engage in substance use despite knowing full well that deteriorating health and ultimately death are likely to result.

Although this inability to steer behavior toward desired goals is at the core of compulsive substance use, that process plays out over years, and assessing it can easily become rather subjective in the individual case. The clinician needs to know whether compulsive drug use is manifested by some simpler features that can be observed here and now, and be less subject to debate. The patients who over the years engage in substance use that takes them far away from their goals do seem to share a number of simpler, clinically meaningful features that pass the utility test. None of these features on its own is necessary or sufficient to identify people with compulsive drug use, but when enough features are present, we can be fairly certain.

One characteristic is simply that alcohol or drug use escalates over time, a feature that has been replicated in experimental animals.¹² As escalation occurs, it takes more and more drug to get the same buzz but also to become impaired—the classic signs of developing tolerance. Patients notice that when they stop, they become anxious, shaky, or miserable in other ways—withdrawal syndromes that are in part different for each drug but tend to be a mirror image of the acute drug effects. Patients attempt to limit the quantities of drug that are used but repeatedly find that, to no one’s surprise except their own, instead of having just one glass of wine with dinner they end up consuming every bottle available, falling asleep on the floor, and throwing up on the carpet. To control intake better, they impose periods of abstinence, often in the hope that they will be able to resume controlled alcohol or drug use after having stayed sober or drug free long enough. But they can’t. The next attempt to engage in social, controlled use that the rest of us can easily manage ends with the same rapid escalation of intake and consumption of much more than was intended. These are all objective, observable facts. The last characteristic may be more open to debate because it relies on subjective reports from patients. But after seeing heroin addicts who stayed clean for years become distressed and sweat profusely at the sight of needles and syringes, I have become convinced that craving—the persistent and overpowering desire to resume drug use even after prolonged abstinence—is as real as any of the other features and is what drives continued substance use despite adverse consequences.

Right here is the core of most attempts to capture a clinical syndrome that denotes an important relationship between a drug and the individual using it. In various constellations, this syndrome is what the WHO and the American Psychiatric Association have attempted to define in their diagnostic catalogs under labels such as “substance dependence” or “substance use disorder.” Whatever the name and the exact diagnostic criteria, what we are trying to capture is a condition associated with extensive medical, psychological, and social consequences. As I write these words, I catch myself because I am reminded that I have somehow lost my way—what I just wrote is hiding the fact that we are dealing with a set of conditions that make people sick, make them desperately sad and anxious, can even kill them, disrupt their families, harm their children physically and emotionally, and cause losses to productivity in the trillions—not to mention all the losses due to drug-related crimes. So yes, there is a category worth distinguishing and having a label for. But before moving on to the issue whether dependence, substance use disorder, or something else is the best label for this important category, we need to branch out and deal with two other categories that are frequently discussed.

Extensive research has shown that this label is essentially meaningless. People with this type of substance-related problem can be found anywhere along the spectrum of substance use severity. Some have what we have so far referred to as dependence. Others have nothing of the kind but may, for instance, live in environments where norm breaking, fights, or other dangerous behaviors are not contained by stable social structures. Most people receive a “diagnosis” of abuse because of a single behavior, such as being caught driving while intoxicated. Unless this is associated with compulsive substance use, I don’t think it is a medical diagnosis. Yes, it is dangerous, irresponsible behavior, but as a physician I do not have any more tools to help out with it than any other sensible adult does. Abuse is simply not a clinically meaningful concept. Large epidemiological studies, such as the National Epidemiological Survey on Alcohol and Related Conditions (NESARC) sponsored by the NIAAA, strongly support this conclusion. As a consequence, the latest edition of the diagnostic manual from the American Psychiatric Association has eliminated abuse as a diagnosis.

Second, it should be obvious that neither substance dependence nor abuse criteria capture everyone who is harmed by their alcohol or drug use. This is particularly important to consider for alcohol because it is so widely used, and because it is such a nasty, unusually toxic drug. It is, for instance, possible for someone who is still in control of their alcohol consumption to develop severe medical problems from alcohol. If a patient has a certain genetic predisposition and has not been told that consuming half a bottle of wine daily with dinner is a major factor behind his hypertension, he could easily end up receiving antihypertensive treatment for decades, until he ultimately develops congestive heart failure, yet never be the wiser. So it is important to recognize and identify a category that we may call hazardous alcohol use, which does not necessarily imply compulsivity. This concept is tremendously helpful in reducing medical problems caused by alcohol use in the general population. We will revisit it when discussing behavioral treatments. For now, let me just mention that if a good family physician assessed our patient’s alcohol consumption and told him about the connection, the patient would in most cases be able to change his behavior, with dramatic improvements in health outcomes as a result.¹³ After all, as I said, we are in pursuit of happiness. There is very little happiness to be found in daily intake of a beta blocker, or becoming unable to climb a flight of stairs. Yes, old habits die hard. Yes, it is a struggle. But ultimately, unless they have developed compulsive drug use, most people are able to change their behavior to achieve their goals. This is fundamentally different from Eric and his fellow patients.

This brings us back to compulsive use, and the special relationship between the person and the drug we have so far delineated as substance dependence syndrome. By now I hope to have made a convincing argument that it is, after all, justified to reserve a category for people with compulsive use—Eric’s kind of relationship with a substance. There are nevertheless some important caveats to consider.

First, if such a category should be delineated, it should be done knowing that we have only imperfect tools to do so. There is no blood test, brain scan, or other objective measure. The diagnostic criteria are all reasonable and reflect a great deal of clinical wisdom. But we have some ways to go. When scientists ask questions of the general population, they find that about 12 percent of Americans qualify for ever having had a diagnosis of alcohol dependence, and about 4 percent receive this diagnosis at the time of being assessed. The numbers are typically twice as high for men as for women. Adding people with dependence on other drugs does not greatly increase the 4 percent number, indicating both that vastly more people have alcohol problems than drug problems and that drug problems frequently occur in people who also have alcohol problems.¹⁴

Wait a moment. Yes, drug and alcohol problems are terribly common, but are they really this common—one in twenty-five adults with a serious substance use diagnosis at any given time? I’m not so sure. The fact is that a majority of these people never seek or receive any treatment. In many cases that is clearly because access to treatment and insurance coverage are sorely lacking. But it is equally clear that in many other cases, these are people who continue to live their lives without negative medical or social consequences. That makes one wonder just how good a job our diagnostic criteria do at capturing what we are after. It turns out that it’s probably not so good if applied outside of a clinical setting. When people in the general population received a diagnosis of ever having had alcohol dependence and then were reinterviewed five years later, two-thirds of the women and one-third of the men no longer received the same diagnosis.¹⁵ Remember, this was about ever having had alcohol dependence. So unless a fairy had allowed these people to relive their lives and change already lived history, the real number of people with a diagnosis simply could not have decreased.

The situation becomes very different, however, if we add one more requirement. Among people who in the first round had received a diagnosis and sought treatment, almost all, or nine out of ten, still qualified for the diagnosis five years later. Clearly, as indicated by Edwards and Gross in the quote that opened this chapter, the patients know something about their alcoholism that the textbooks do not know. They tell us about it when they seek treatment. Accordingly, other studies indicate that treatment-seeking alcoholics and those who qualify for the same diagnosis but never seek treatment largely come from two different populations. To emphasize that distinction, I frequently refer to “clinical alcoholism” when discussing the former group.

Having come this far, we are ready to ask, at last, what might be the most appropriate name for the category, or group of categories, I have so far delineated. I have already given away the name that for decades was the official one, “dependence,” used by Edwards and Gross above. I love that classic paper, and its message that patients with alcohol problems know something important about their condition that we may not be able to find in a textbook. But the choice of the word “dependence” is terrible. Outside a limited circle of scientists and specialized clinicians, dependence is typically interpreted as physical dependence, characterized by tolerance and withdrawal. As a result, laypeople and professionals alike continue to view these two phenomena as integral to clinically significant substance use disorders. And yet we know now that compulsive drug use reflects pathology of motivational brain circuitry, not the shakes of coming off alcohol. Dependence, in the physiological sense most commonly given to it, is neither necessary nor sufficient as an ingredient in a substance use disorder.

The failure to make that distinction for decades has been doubly problematic. Here are illustrations of the two sides of that coin: On one hand, Δ⁹-THC, the active ingredient of marijuana, is eliminated by the body so slowly that it essentially provides a built-in taper, masking the expression of withdrawal and frequently leading people to believe that cannabinoids are somehow less addictive ¹⁶ than other classes of drugs. On the other hand, even though the vast majority of benzodiazepine-treated patients do not develop addiction, their predictable needs for dose increases over the long term, and the emergence of accentuated symptoms on discontinuation has given rise to an entire cottage industry of “taper treatments.” Based on this misconception, patients with debilitating anxiety disorders frequently have their successful treatments discontinued to address their presumed benzodiazepine dependence. Worse still, the mixup has allowed ideologues generally opposed to antidepressant treatments to claim that these are addictive, too, because—similar to antihypertensive drugs—there is a discontinuation syndrome.