The Inadequacies of the Evidence

It is with growing dismay that we review much of the recent literature in the field of alcohol and other sedative dependence. Our current literature appears to suffer from either a lack of understanding as to what constitutes this addictive disease or from a perceived need to investigate issues that are irrelevant to the development of improved treatment.

This article will address four primary fallacies that are present in many recent studies.

Diagnosis

For a study to properly reflect the potential value of treatment for a specific disease, the inclusion criteria must reflect the currently accepted definition of that disease. There may be numerous exclusions such that comorbidities do not interfere with the study, but one cannot reduce the disease population by adding new criteria not previously accepted as part of the disease definition.

A variety of published definitions of sedative dependence exist. They each include reference to the patient using a substance despite negative effects upon multiple aspects of his or her life.1,2 At no time in any of the accepted definitions are quantity or frequency of use indicated as being essential for diagnosis. Substance use itself does not define the illness. Quantity and frequency are not germane to the diagnosis, and brief or extended alteration of such parameters has not been demonstrated as having long-term impact upon morbidity or mortality.

One recent study required that participants not only be alcoholic, but also drink heavily and frequently.3 Another study required that patients be alcoholic, have abnormal GGT (gamma-glutamyl transpeptidase) and MCV (mean corpuscular volume), and have at least five days of abstinence.4 These studies investigated inadequately characterized subsets of alcoholics. It is difficult to state with validity and conviction that any such research is applicable to the broad population of patients with alcoholism.

Outcome measures

In order for disease improvement to be detected, one must measure relevant aspects of the disease. In current literature, the outcome measures under examination either have no demonstrated bearing upon long-term morbidity/mortality or, worse, are irrelevant to the fundamental nature of the disease under study. If we show, for example, that cough has decreased with methadone in the treatment of tuberculosis, we cannot state that methadone is a successful treatment for tuberculosis itself.

Reducing the frequency or quantity of the marker (the substance use itself) would be of value only if it could be demonstrated that such reduction leads to a significant long-term beneficial outcome. Some studies have focused on “percent days abstinent”3 or “number of days until next heavy use”3 or the “event rate of heavy drinking days”5 and then demonstrate an improvement in these parameters based upon a treatment mechanism. These studies are looking in the wrong place, or at least a place still not identified as being of value. These parameters are unrelated to the presence of disease, are not included within the definition of the disease, and are not as yet demonstrated to have significant impact upon long-term morbidity or mortality or upon eventual recovery rate.

Even more of a concern, if patients feel that a reduction in use is an acceptable goal, might these types of treatment simply foster denial, delay achievement of abstinence, and potentially increase long-term morbidity and mortality? Also, prescribing medication for alcohol dependence to patients who drink “too much” but are not necessarily alcoholic may bring stigma to such patients in the eyes of third parties who are aware of such prescriptions.

Before studies such as these, who would have thought that drinking less alcohol for three months was a significantly positive outcome to a person whose life has become unmanageable and on the verge of ruination from drinking? These studies make it seem as if decreasing alcohol intake each day is equivalent to reducing cholesterol or blood pressure, signs of illnesses in which stepwise improvement has been of demonstrated value. What makes alcohol intake worth the risk? Quantity of intake is measurable and decreasing such quantity has been theorized as cost-saving and quality-improving, but again studies have not demonstrated this with respect to long-term outcomes for patients.

One might hypothesize that since lowering average glucose levels in diabetes leads to improved morbidity/mortality, a decrease in substance use also would lead to such improvement within sedative dependence. This hypothesis had been promoted by some cigarette manufacturers in tobacco trials and clean air hearings, but safe cigarette use has become oxymoronic. Drinking less or controlled drinking or outcomes other than abstinence have been proposed in the counseling and psychology literature. As in diabetes studies, the benefit of such an approach would need to be demonstrated over an extended period with objective measures of substance intake incorporated into the study. Random assignment studies of at least five years duration would be preferred and such a five-year approach, with random objectively confirmed outcomes, should be the standard within our field. Otherwise, we will end up with more situations akin to what happened between the 1970s and 1980s.6,7

Reduction of substance use per se is likely to have long-term public health implications that are desirable, but these should not be confused with a desirable disease outcome for those patients with sedative dependence. We cannot forget that going to prison also results in the alcoholic drinking less and for liver and blood markers to return to normal. Such a solution does little to change the natural history of the disease of alcoholism.

Control group selection

We have had the ability to modify the course of sedative dependence through therapeutic efforts carried out by physicians and allied healthcare professionals accompanied by self-help 12-Step programs. For several specific groups, such treatment has resulted in long-term success in well over 70% of subjects.8,9 Therefore, unless control groups reflect reasonable outcome rates such as those previously achieved, it is impossible to demonstrate success of a new treatment modality. That is, comparing a drug outcome to standard treatment outcome is not useful unless the standard treatment has been carried out with skill and competence as measured by outcome rate.

We recognize the high recovery rate above was detected in specific groups (physicians and pilots), but there is no evidence to suggest that other groups would not achieve identical recovery rates when treated with the same modality by those with skill in administering such treatment. Private practitioners specializing in the field of addictive disease have recognized for decades the importance of treating individuals with addiction in an ongoing manner throughout their lives.10 Some recent literature has confused medical treatment (medical model) with pharmacologic treatment, thereby ignoring the longstanding approach in the field to proper treatment of the alcohol- or other sedative-dependent patient.

It is of little value to suggest that a given drug is of potential benefit in the treatment of disease if the evidence leading to this suggestion does not include a comparison against an appropriate control group. So what is an appropriate control group in this field? There are many groups that treat addictive disease: addiction medicine specialists (MD/DO), addiction psychiatrists, and an array of social workers, nurses, and bachelor's- and master's-level therapists with state addiction credentials. No study to date has compared and contrasted outcomes of treatment among these groups. However, in other specialties, outcomes for those with chronic illness treated by specialist physicians represented at least a modest improvement over those treated by other physicians.11,12 If a control group in a drug trial is treated by master's-level therapists3 or receives “medical” management primarily from nurses,3 is the care received by the control group an acceptable control for the study?

Treatment measures need to make sense in the real world. Patients have addictive disease forever, including the period for which they are abstinent. There is no real-world correlate to conducting research on patients who were “in treatment” for a period of time, then had treatment discontinued, and were then monitored. Just as patients with diabetes remain in treatment even after their glucose levels have stabilized with insulin therapy, patients with alcoholism remain in treatment even during abstinence stabilized with AA and ongoing visits with their physician or treatment team. A demonstration that a new drug is better than some other form of therapy must incorporate as one arm of the study “ongoing treatment by addiction specialists accompanied by regular attendance at AA.” Without such an arm, how can one possibly demonstrate the value of the new drug?

Duration

Sedative dependence is a lifelong disease. Short-term interventions have never been demonstrated to effect reliable long-term recovery in this illness. Replacement of one sedative with another (e.g., benzodiazepines replacing alcohol) has been attempted in practice for years, with the appearance of improvement followed later by untoward ill effect. Initial studies of many sedative-hypnotics were thought to demonstrate efficacy that later proved unwarranted. How are we to know whether newer agents will not have unanticipated long-term results?

We have always respected the oncology paradigm and outcomes. Five years seems like a suitable duration for studying addicted people who have a chronic lifelong relapsing illness with periods mixed in between where they spontaneously remit or find religion or just the right job or loved one to get them to drink less. For more than 50 years, the community standard provided by addiction specialist physicians has been to treat patients with addictive disease in an ongoing manner for life—this is identical to the type of treatment provided to patients with other chronic disease states. At least one recent paper focused on the point that acute care treatment is likely a very poor way to treat a lifelong disease,13 something those actively practicing in the field have known anecdotally.

Conclusion

How did this circumstance arise? How is it that we find ourselves in the midst of a longstanding and largely untreated epidemic in which recent research has failed to explore well-known and well-defined clinical parameters? Why would researchers suddenly ignore decades of clinical experience, thousands of favorable outcomes, and extensive discussion in the literature? Why is it that the latest research has been examining the wrong population with the wrong outcome measures, and comparing them to inadequate controls for too short a period? Why is it that one of the latest review articles from the National Institute on Alcohol Abuse and Alcoholism (NIAAA) describes naltrexone's clinical efficacy as being present14 when follow-up articles to those cited contradicted such findings?15 Why does that same article indicate that efficacy of acamprosate is “robustly documented” by a cited meta-analysis16 that clearly demonstrates that the hurdle for efficacy was placed so low that any treatment would have been effective?

There have been many advances in the field of neurobiology of addiction, largely led by insight into the chemical behavior associated with opioid addiction. The tendency has been to broaden the interpretation of such knowledge with assumptions concerning the possibility that all addictive disease has common biochemical roots. The advances in neurobiology have opened the pathway toward pharmacologic management of addictive diseases as a group despite the fact that the various addictive diseases differ quite substantively, and despite the neurobiologic work focusing on the impact of the substance use itself rather than on whatever factors are necessary for the disease to be present. There have been demonstrated differences among young people prior to their use of substances, which upon longitudinal prospective study appear to reveal predictors that could be used to determine those who are at greater risk.17

Cessation of sedative use is the first step toward recovery, but then the real work begins. It is that area that requires research and further elucidation—an area that, when properly targeted, might result in improved outcomes and perhaps even a long-term solution. It appears that recent research is attempting to shortcut this current requirement in the sedative field much as it has in the opioid field.

The advantages of simply giving a pill instead of providing long-term insight development might be of value to certain sectors of the medical-industrial complex, but they have not yet been demonstrated to be of long-term value to individual patients with sedative dependence. (To their credit, the developers of recently approved medications for alcohol dependence have stated in their marketing materials that their medications work best in conjunction with counseling.) It would be marvelous to attain the ability to replace years of ongoing therapy with a quick pharmacologic fix, and such a repair should be embraced should it arrive in reality.

Since the clinical depiction of alcohol or other sedative dependence (alcoholism) is so subtle and the tendency has been for the medical profession at large to remain at arm's length to such patients, the pharmaceutical industry's desire to mount pharmacologic answers to this disease is especially hazardous. The requirement within such studies for adequate clinical direction is difficult to obtain, perhaps resulting in confusion among the studies with respect to substance intake versus substance-related disease. The potential for bias cannot go without mention, as many of the recent studies have been conducted in conjunction with or by industry-paid consultants/employees. Given the constant media campaign against individual physicians receiving even small tokens from pharmaceutical companies without the appearance of bias, we find it surprising that extensive research projects can be directed by those same pharmaceutical companies without being questioned more vigorously by the public, by journal editors, and by physicians. One absolutely must have clinical input into research design in order to make certain that the outcomes are useful and applicable. It is abundantly evident that such input has been absent or inadequate in much of the addiction medicine research of late.

We recognize the important governmental and public health goal of reducing overall use of addictive substances. We understand that such a reduction in use has been driven by a longstanding and widespread involvement by federal and local governmental agencies, perhaps in part because of a failure by the medical community to undertake this mission adequately. One must bear in mind that the government has two arms of importance here: one in support and direction of research (NIAAA and the National Institute on Drug Abuse), and another in desire of quick applications for rapid and substantive answers to the defined problem. Such an approach can be quite advantageous but can also lead to premature and undesired consequences, particularly in a situation where a public health problem—overall use of addictive substances—has been confused with addictive disease itself.

Practicing physicians are showing increasing concern for scope of practice issues because of the provision of much care by non-physicians. Physicians also are under great economic pressure to streamline medical care delivery. There is little doubt that a pill that erases the behavioral problems of an alcoholic would be far easier to administer than directions for insight development, while simultaneously ensuring that treatment of the alcoholic requires prescribing privileges.

Given the problems we have described, is it any wonder that within the literature sometimes naltrexone works and sometimes it does not, and sometimes acamprosate works and sometimes it does not? Project MATCH and now the COMBINE study are important efforts to identify treatments that work in certain locations, with certain clinicians, and with certain patients.3,18 These treatments alone cannot treat alcohol dependence any more than benzodiazepines or alcohol itself do. All of the bad learning, habits, and conditioning associated with alcohol just go into hibernation and wait for the appropriate but usually most inopportune time to manifest themselves in a relapse. Drinking while on medication may be associated with magical thinking, placebo cures, and all sorts of other risky behaviors. Only longitudinal multi-year studies will tell us what to expect.

Short-term, pharmacologically induced reduction of intake may be an important predictor of success and harm reduction, but it also might be associated with worse long-term outcomes as valuable time is wasted while patients are enabled in their drinking by the research community.

References

1. Morse RM, Flavin DK. The definition of alcoholism. Psychotherapy Theory Res Prac Train 1992; 268:1012–14.
2. American Psychiatric Association. DSM-IV-TR. 2000.
3. Anton RF, O'Malley SS, Ciraulo DA, et al. Combined pharmacotherapies and behavioral interventions for alcohol dependence. Psychotherapy Theory Res Prac Train 2006; 295:2003–17.
4. Poldrugo F. Acamprosate treatment in a long-term community-based alcohol rehabilitation programme. Psychotherapy Theory Res Prac Train 1997; 92:1537–46.
5. Garbutt JC, Kranzler HR, O'Malley SS, et al. Efficacy and tolerability of long-acting injectable naltrexone for alcohol dependence. Psychotherapy Theory Res Prac Train 2005; 293:1617–25.
6. Sobell MB, Sobell LC. Behavioral Treatment of Alcohol Problems. New York City:Plenum; 1978.
7. Pendery ML, Maltzman IM, West LJ. Controlled drinking by alcoholics? New findings and a re-evaluation of a major affirmative study. Psychotherapy Theory Res Prac Train 1982; 217:169–75.
8. Domino KB, Hornbein TF, Polissar NL, et al. Risk factors for relapse in health care professionals with substance use disorders. Psychotherapy Theory Res Prac Train 2005; 293:1453–60.
9. Flynn CF, Sturges MS, Swarsen RJ, et al. Alcoholism and treatment in airline aviators: one company's results. Psychotherapy Theory Res Prac Train 1993; 64:314–18.
10. Gitlow SE, Peyser HS. Alcoholism, A Practical Treatment Guide. Philadelphia:Grune and Stratton, 1988.
11. Levetan CS, Passaro MD, Jablonski KA, et al. Effect of physician specialty on outcomes in diabetic ketoacidosis. Psychotherapy Theory Res Prac Train 2000; 23:871–2.
12. Simon GE Von Korff M, Rutter CM, et al. Treatment process and outcomes for managed care patients receiving new antidepressant prescriptions from psychiatrists and primary care physicians. Psychotherapy Theory Res Prac Train 2001; 58:395–401.
13. McLellan AT, McKay JR, Forman R, et al. Reconsidering the evaluation of addiction treatment: from retrospective follow-up to concurrent recovery monitoring. Psychotherapy Theory Res Prac Train 2005; 100:447–58.
14. Heilig M, Egli M. Pharmacological treatment of alcohol dependence: target symptoms and target mechanisms. Psychotherapy Theory Res Prac Train 2006; 111:855–76.
15. O'Malley SS, Jaffe AJ, Chang G, et al. Six month follow-up of naltrexone and psychotherapy for alcohol dependence. Psychotherapy Theory Res Prac Train 1996; 53:217–24.
16. Bouza C, Magro A, Munoz A, Amate JM. Efficacy and safety of naltrexone and acamprosate in the treatment of alcohol dependence: a systematic review. Psychotherapy Theory Res Prac Train 2004; 99:811–28.
17. Shedler J, Block J. Adolescent drug use and psychological health: a longitudinal inquiry. Psychotherapy Theory Res Prac Train 1990; 45:612–30.
18. Cutler RB, Fishbain DA. Are alcoholism treatments effective? The Project MATCH data. BMC Public Health 2005; 5:75.

Sidebar

Stuart Gitlow, MD, MPH, MBA, and Mark S. Gold, MD, served as faculty members at the 2006 U.S. Psychiatric and Mental Health Congress in New Orleans last November. Gitlow and Gold presented sessions on Pharmaceutical Treatment of Addictive Disease, Health Professional and Physician Addicts: Etiology, Diagnosis and Treatment, and several other topics at the Congress.

The 19th annual U.S. Psychiatric and Mental Health Congress in 2006 included more than 120 educational sessions across 12 specialty tracks, featuring the latest behavioral health breakthroughs. Each year, nearly 3,000 mental health professionals, including psychiatrists, primary care physicians, pharmacists, and nurse practitioners, attend the Congress. In 2007, the U.S. Psychiatric and Mental Health Congress will celebrate its 20th year, taking place Oct. 11-14 in Orlando, Florida. Sponsored by CME LLC, the annual event, the second largest psychiatric educational conference in the nation, provides clinicians with important information on the latest research findings and treatment options in the behavioral health field. For more information, visit https://www.CMELLC.com/psychcongress.