Misdiagnosing Substance Use

© adiruch | 123rf.com

Allen Frances doesn’t like the DSM-5 and you can hear him say so here.  He said our mental health system was in a mess. And he is afraid that with DSM-5, it will get even worse. “People who are essentially normal are being diagnosed with mental disorders they don’t have.” Small changes in the diagnostic system can result in tens of millions of normal people qualifying for a diagnosis. He used himself as an example, stating how he would qualify for several of the DSM-5 disorders. Typical symptoms of grief over his wife’s death, lasting beyond two weeks, would have signified him as having a Major Depressive Disorder.

Anther mistake was combining what had been two different diagnoses of substance use in the DSM-IV—Substance Abuse and Substance Dependence—into one: Substance Use Disorder. Substance Abuse was when someone had recurrent, but intermittent, trouble from recreational binges. Substance Dependence was a continuous and compulsive pattern of use, often with tolerance and withdrawal. The majority of substance abusers “never become addicted in any meaningful sense.”

The two DSM IV diagnoses have radically different implications for treatment planning and for prognosis. Artificially lumping them together in DSM-5 forces inaccurate diagnosis, loses critical clinical information, and stigmatizes as addicts, people whose substance problem is often temporary and influenced by contextual and developmental factors.

Hasin et al., “DSM-5 Criteria for Substance Use Disorders: Recommendations and Rationale,” presents the rationale used by the DSM-5 workgroup for substance use disorders for its changes, particularly combining abuse and dependence into one disorder. They recommended the combination as well as dropping one diagnostic criteria (legal problems) and adding one (craving). Two criteria are required to diagnose a Substance Use Disorder. The number of criteria met will indicate mild (2 to 3 criteria), moderate (4 to 5), and severe disorders (6 or more). The following chart, taken from the article, illustrates the changes from DSM-IV to DSM-5.

Frances is not alone in seeing value with two distinct types of substance use disorder. Carleton Erickson in The Science of Addiction noted how the distinction allowed for the differentiation between individuals with drug-related problems who could stop using when they wished (substance abusers), and others who had the disease of chemical dependence. Chemically dependent people have a dysregulation of the mesolimbic dopamine system and generally cannot stop using drugs without intensive intervention into their drug use problems. “According to these criteria, drug abuse in intentional, ‘conscious,’ or voluntary. Drug dependence is pathological and unintended.”

In his article, “DSM-5 Made a Mistake Eliminating Substance Abuse,” Allen Frances indicated the DSM-5 workgroup for substance use disorders based its rationale for dropping Substance Abuse on studies suggesting the distinction was hard to make. He said the results of the studies were not definitive. Moreover, their interpretation was flawed by what he said was a basic DSM-5 misunderstanding of the nature of psychiatric diagnosis. “All DSM disorder overlap with other DSM disorders and also frequently with normality.” Fuzzy boundaries among near diagnostic neighbors are common and not a sufficient excuse to collapse clinically valuable distinctions.

Carleton Erickson’s discussion of the degrees of severity with drug problems helps to illustrate this misunderstanding. He indicated there were mild, moderate and severe forms of both drug abuse and drug dependence. Most people don’t think in terms of severity with substance use problems. You either have a problem or you don’t; you either abuse drugs or you don’t. He then illustrated their relationship to drug-seeking behavior as follows.

Drug Abuse

Drug Dependence








A Lot


Even more


All the Time

The overlap referred to by Frances occurs between severe drug abuse and mild drug dependence. The inability of psychiatric diagnosis to make a clear distinction here seems to have led to the decision to collapse the abuse and dependence diagnoses into one category in the DSM-5.

I think another overlap between drug abuse and drug dependence happens with regards to self-control. A distinction is necessary between self-control of thoughts, feelings and behavior when drinking and control of the drug intake itself. Any substance use can lead to a loss of self-control over an individual’s thoughts, feelings and behavior. When that loss of control results in recurrent, intermittent trouble, there is a drug abuse problem. The severity of this type of loss of self-control and the related intermittent trouble varies.

Not everyone who abuses a drug experiences the classic sense of losing of control over how much of the drug they use. A loss of control over drug intake—a continuous and compulsive pattern of use—is only evident within drug dependence. And again, the severity of this loss of control over drug intake varies. So I’d adopt Erickson’s degrees of severity with drug abuse and dependence problems as seen below.

Loss of Self-Control in Abuse

Loss of Control over Drug Intake in Dependence







A substance abuse problem with severe trouble related to loss of self-control may be indistinguishable from a substance dependence problem with mild loss of control over drug intake. Both people would look at their severe “trouble” and attribute it to drinking or drugging too much. Given an equal motivation to avoid further “trouble,” the substance abuser would likely have an easier time maintaining abstinence. Carleton Erickson said chemical dependence is not a “too much, too often, withdrawal” disease; it’s a “I can’t stop without help disease.” There is a pathological, compulsive pattern to substance use.

There does seem to be a “fuzzy boundary” between Substance Abuse and Substance Dependence. Nevertheless, the distinction still carries some clinical and diagnostic value. I agree with what Allen Frances said: “The change was radical, creates obvious harms, and provides no apparent benefit.” What should clinicians do? Frances suggested they simply ignore the DSM-5 change. He said it was appropriate and clinically preferable to continue making the distinction.

There is nothing sacred or official about the DSM-5 choices — I know because I made the choices for DSM-IV. The ICD coding system is official; the DSM codes are just one groups’ fallible adaptation of them. It is of great significance that the official coding in ICD-10-CM does not follow the DSM-5 decision to eliminate Substance Abuse. Instead, ICD-10-CM retains the DSM-IV terminology and continues to provide separate Substance Abuse and Substance Dependence codes for each of the major classes of substances.

The ICD-11 workgroup, currently in the final stage of development before field tests, will continue to separate Substance Dependence and Harmful Substance Use. The guidelines for dependence are revised and simplified into three diagnostic features: impaired control over substance use; increasing priority in life and physiological features. Severity qualifiers were suggested only for alcohol intoxication. They also introduced a new diagnostic category, with no equivalents in ICD-10 or DSM-5: single episode of harmful use. Frances commented:

The DSM-5 mistake thus places it out of line with ICD-10, ICD-11, previous DSM’s, and well established clinical practice. Clinicians remain truer both to clinical reality and to ICD coding when they ignore the new DSM-5 lumping of substance use disorders and instead continue to distinguish Substance Abuse from Substance Dependence. DSM’s are explicitly meant to be used only as guides, not worshiped as bibles. Clinicians are free to ignore DSM whenever it makes mistakes that go against clinical common sense and the International coding system.


Binge Drinking Biomarkers

© Konstantin Kulikov | 123rf.com

© Konstantin Kulikov | 123rf.com

Over two years ago, the Chair of the DSM-5, David Kuyper admitted that despite telling patients for decades that they anticipated finding biomarkers for psychiatric disorders, that discovery continues to be “disappointingly” out of reach. The hope is that at some future time, the promise will become a reality. “In the future, we hope to be able to identify disorders using biological and genetic markers that provide precise diagnoses that can be delivered with complete reliability and validity.”

In “What are Biomarkers?”, Strimbu and Tavel said by definition, biomarkers are objective, quantifiable characteristics of biological processes. The WHO said a biomarker is almost any measurement that reflects the interaction between a biological system and a potential hazard. “The measured response may be functional and physiological, biochemical at the cellular level, or a molecular interaction.” They cautioned that the overreliance on biomarkers “presents a serious and persistent risk of producing misleading, and in some cases dangerous, erroneous conclusions.” And yet, in many areas of research, it appears this warning is ignored.  Keep this in mind as we review here three different studies into biomarkers for—of all things—binge drinking.

A recent press release announced that a biomarker distinguishing binger drinkers from moderate drinkers among young adults has been found by researchers at the University of Illinois at Chicago. The press release suggested the biomarker, “called phosphatidyl ethanol (PEth), could be used to screen young adults for harmful or heavy drinking such as binge drinking.” The lead author for the study, Mariann Piano, said: “Using a biomarker of heavy alcohol consumption such as PEth along with self-reporting could provide an objective measure for use in research, screening and treatment of hazardous alcohol use among young adults.” But I’m not convinced that it would be an appropriate or even helpful screening tool as suggested.

The association of PEth to alcohol consumption has been known for some time, so the suggestion this was a “groundbreaking” research study examining “the relationship between biomarkers and alcohol consumption” by Brent McCluskey in The Fix was a bit misleading. McClusky later clarified that the study was “groundbreaking” because this was the first study to investigate PEth levels in young adults. A search of just the journal Alcohol and Alcoholism, where the Piano et al. study was published, indicated there were 34 total articles meeting the search criteria of “PEth.”

Dr. Lena Gustavsson was the winner of the ESBRA Award in 1994 for her work on phosphatidyl ethanol (PEth). Aradottir et al. reported in 2006 that PEth was a promising new marker for ethanol abuse. They found that blood concentrations of PEth were highly correlated to alcohol intake. “Its diagnostic sensitivity is higher than that for previously established alcohol markers.”

In 2011, Isaksson et al. noted that since the formation of PEth was specifically dependent upon ethanol, “the diagnostic specificity of PEth as an alcohol biomarker is theoretically 100%.” They added that the half-life of PEth in blood is around four days.

Helender et al. (2012) concluded that PEth was the most sensitive biomarker of current alcohol consumption and prior drinking because the PEth test can detect lower consumption levels.

Jain et al. in 2014 found strong associations between PEth and self-reported measures of alcohol consumption among young injection drug users. They suggested PEth may be a useful marker “in settings where alcohol consumption is difficult to assess,” or to confirm or refute self-reported measures of alcohol consumption.

Kechagias et al. (2015) concluded that PEth was the only marker that could discriminate between abstinence and a moderate daily consumption of alcohol.

So the PEth biomarker would help to confirm whether or not someone is honestly reporting their recent consumption of alcohol. It would NOT be a biomarker to identify at risk binge drinkers. And it seems to have little application for treatment other than as an “honesty test” in treatment to affirm or rule out recent alcohol consumption. It may be a more accurate or sensitive test than a breathalyzer, but that doesn’t strike me as groundbreaking news.

Biological Psychiatry had another study in its in June 2015 issue by Warnault et al. that suggested a gene variant, Met68BDNF, reduced the release of brain-derived neurotrophic factor (BDNF) in mice. These mice would consume excessive amounts of alcohol and continue to drink despite negative consequences. An NIH news release said the researchers treated the alcohol with bitter-tasting quinine. “This suggests Met68BDNF carriers compulsively drink alcohol despite aversive consequences.”

The implication is that in humans, this gene variant may put an individual at greater risk of developing an alcohol use disorder. By administering a pharmaceutical compound developed to mimic the action of BDNF, the researchers were able to stop the compulsive drinking behavior in the mice. “This compound (LM22A-4) may have potential as a therapeutic for humans. It appears to reduce compulsive drinking without a generalized effect on motivation.”

BDNF is a protein in humans that is encoded by the BDNF gene. It acts on certain neurons, “helping to support the survival of existing neurons, and encourage the growth and differentiation of new neurons and synapses.” In the brain, it is active in areas vital to memory, learning and higher thinking. The BDNF gene may play a role in the regulation of stress response and in the biology of mood disorders. The expression of BDNF is reduced in Alzheimer’s and Huntingdon disease. Linda Gabriel described BDNF as Miracle-Gro for the brain.” If you’re interested, Alomone Labs is offering a free sample of human BDNF.

Although an interesting finding, it has only been demonstrated within mice at this point in time. If further research were to show that Met68BDNF in humans was associated with a higher risk of alcohol use disorders, compounds like LM22A-4 may help in reducing the neurochemical element in craving or compulsive drinking. But there’s more to compulsive drinking than just biology and neurochemistry.

The most interesting study to me was reported in News & Views for The Scripps Research Institute. Herman et al. reported that the deletion of GIRK3 subunits (G-protein-gated inwardly rectifying potassium) in mice. They compared “knockout” mice, ones missing GIRK3, with normal mice. During a simulated “happy hour” when access to alcohol was limited to two hours a day, GIRK3 knockout mice consumed significantly more alcohol than the control group of normal mice. When mice were given continuous access to alcohol, conditions where mice do not get intoxicated, the effect was not evident.

There was no difference found between the GIRK3 and control mice in how alcohol was metabolized. Both groups also experienced a similar loss of balance, sleepiness and reduced body temperature in response to alcohol. The researchers thought there were two possibilities. Mice without GIRK3 could be drinking more because they felt more pleasure from alcohol, so they wanted to drink more. Or they felt less pleasure and thus needed to drink more to reach the same level of pleasure as normal mice.

In order to answer this question, they looked at the mesocorticolimbic dopaminergic pathway (the reward pathway in the brain) and found that the pathway was completely insensitive to alcohol without GIRK3. Even at high doses, alcohol did not alter the firing of neurons missing GIRK3. Alcohol also failed to trigger the release of dopamine in the ventral striatum of GIRK3 mice. The results suggested that mice drink more alcohol to boost the engagement of other neural pathways mediating alcohol’s rewarding effects.

By reintroducing GIRK3 in the knockout mice, the researchers were able to alter the binge drinking down to normal levels. Normal mice with increased GIRK3 drank even less. “This has led the researchers to believe that a compound selectively targeting GIRK3-containing channels may hold promise for reducing alcohol consumption in heavy binge drinkers.” An abstract for the study published in the Proceedings of the National Academy of Sciences can be found here.

PEth, BDNF, and GIRK3 are all biomarkers of some sort that are being applied to the potential hazard of “binge drinking.” As Strimbu and Tavel observed, the key issue is determining the relationship between any given biomarker and the relevant clinical endpoints, which in this case is binge drinking. In some cases, biomarkers may be shown to measure the process of a key pathway stage in reaching the clinical endpoint of binge drinking. But assuming this relationship risks mistaking correlation for causation. PEth seems to be an example of a correlational biomarker. And caution needs to be exercised at this point with any conclusions drawn from the findings of studies with BDNF and GIRK3.

While there is growing evidence of a genetic connection with alcoholism, the evidence is not conclusive at this time. Following Carleton Erickson, I’d say that while alcohol dependence runs in families, it is not purely a genetic disease. “Rather, the tendency to become alcoholic is inherited. Thus alcoholism can skip generations, or affect only certain individuals in an alcoholic family.” See “The Genetic Connection” for more discussion on genetic research into alcohol dependence.


The Genetic Connection

© Claudio Ventrella | 123RF.com

© Claudio Ventrella | 123RF.com

Researchers at the University of Texas Austin have announced that they found “convergent evidence” of gene networks related to alcohol consumption in humans. “These networks of coordinately expressed gene sets with common biological functions provide a systems level model of alcohol dependence.” But it would be wrong to conclude, as the Digital Journal did that: “Researchers at the University of Texas Austin successfully identified a genetic network which appears to determine alcohol dependence.” Other news outlets, such as Fox News, said something similar. This is another example of how the media reporting on research runs ahead of the research itself.

But the researchers themselves didn’t make the same mistake. The WCAAR newsletter from the University of Texas Austin, which was cited above, said: “The purpose of this study was to arrive at a comprehensive picture of the gene sets driving the phenotype of alcohol dependence.” They hoped the gene networks they identified could be useful as a reference for future studies, “such as exploration of specific drug targets in the network or identification of new gene isoforms unique to humans.”

The study consisted of doing post mortem biopsies of the prefrontal cortex brain tissues of two groups of individuals—alcoholics and non-alcoholics. They discovered that at a global level, gene connectivity, “representing fully functional biological processes,” was weaker within the alcoholic group. “This dysregulation of normal biological function is a hallmark of alcohol dependence.” Looking at the abstract for the published study in Molecular Psychiatry, the team of researchers said:

Overall, our work demonstrates novel convergent evidence for biological networks related to excessive alcohol consumption, which may prove fundamentally important in the development of pharmacotherapies for alcohol dependence.

The problem seems to have been that the news outlets were attributing causation to what is at this stage of the research only a correlational finding. Here is a short video of Ionica Smeets explaining the difference between correlation and causation; and the dangers of confusing correlation for causation.  In her talk she said: “Jumping to an incorrect conclusion about causality when you see a correlation, is the most often made logical mistake.”

This was a study of the brain tissue of dead alcoholics and non-alcoholics. The post mortem presence of the gene networks was suggestive of a possible predisposing genetic factor for alcoholism, but could not said to be causative.  “This strong interconnectedness of gene networks linked to alcohol dependence was absent in the control tissue.”

The alcohol consumption networks in the alcoholic tissue suggest subtle transcriptome reorganization in response to alcohol dependence. They may also indicate an allostatic mechanism – a physiological process to regain homeostasis – to overcome changes associated with alcohol dependence.

An alostatic mechanism refers to “the wear and tear” or physiological consequences of chronic exposure to heightened neural or neuroendocrine response that is the result of chronic stress—in this case from the lifelong consumption of large amounts of alcohol. So the researchers are suggesting their results indicate the presence of “a physiological process to regain homeostasis”—an attempt by the brains of the alcohol dependent individuals “to overcome changes associated with alcohol dependence.” They see adaptation, but not necessarily causation.

Alcohol dependence runs in families, and is suggestive of genetic causality, but not conclusively so. The Addiction Science Research and Education Center of the University of Texas Austin noted that while alcoholism runs in families and is suggestive of a genetic cause, many things run in families that aren’t genetic, such as speaking Spanish. Although twin and adoption studies indicate susceptibility for alcohol dependence in families, there are some individuals like Jay Joseph who question the scientific validity of twin studies (See “Chasing Ghosts” and “Troubling Twin Studies“). But even Erickson fails to make a causal link between genetics and alcoholism:

Alcohol dependence is not a genetic disease (which suggests destiny); rather, the tendency to become alcoholic is inherited. Thus alcoholism can skip generations, or affect only certain individuals in an alcoholic family. (number 103 of the Alcohol Facts)

An earlier theory of the “cause” of alcoholism was the “THIQ theory.” As Carleton Erickson indicated in his list of “Drug Myths,” the THIQ is an older myth popular in the early 1970s that when alcoholics drink they form opiate-like THIQs in the brain, to which they become “addicted.” There is a short, but helpful description of the THIQ myth on the HAMS Network. I don’t agree with much of what is there about addiction and recovery, as I have a 12 Step, abstinence-based approach to recovery, where they are harm reduction-based. But this seems to be a balanced summary of what is now a discarded theory.

So while there is growing evidence of some kind of genetic connection with alcoholism, that connection is not conclusive at this point in time. Nor can it be said to be causative. And if or when it becomes more conclusive, it seems that a better way of describing it is not as a genetic disease, but an inherited tendency. So the expressed hope of the study’s findings by its researchers can be anticipated, namely that it “may prove fundamentally important in the development of pharmacotherapies for alcohol dependence.”  But it doesn’t provide evidence for a genetic cause of alcohol dependence.