The Many Levels of Relapse to Drinking: Commentary on Meyerhoff and Durazzo (ACER 2020)

Traditionally, the desired successful recovery from alcohol dependence has required complete abstinence, without transgression, for a sustained period, ideally the rest of one’s life. This stringent definition marks a goal that is unachievable by a majority of men and women with Alcohol Use Disorder (AUD) despite sincere efforts to stop drinking. While perhaps ideal, that goal of complete and permanent abstinence puts behavioral and pharmacological treatment at substantial risk for failure. Here we ask, are there acceptable alternatives to abstinence, and how are they established and substantiated?

WHO “Reduced Levels of Drinking” Metric

Recent efforts in clinical trials have challenged the use of sustained abstinence as the success endpoint and have supplanted it with new endpoints of pre-defined levels of drinking relapse. Specifically, these new treatment endpoints were put forth by the World Health Organization (WHO) and are operationally defined drinking levels of post-treatment relapse: very high, high, moderate, and low. These categories are based on a comparison of post-level drinking with pre-treatment drinking and applied separately to men and women; the daily consumption limits are lower for women than men. It is notable that the reduction in drinking level could still be considered high consumption in absolute terms, i.e., exceeding health guidelines of no more than 14 drinks/week for men and 7 drinks/week for women. Using the WHO “risk-drinking level reduction” approach, a recent study found that individuals who reduced their levels of drinking by one or two WHO levels could experience a degree of improved quality of life functioning and reduction in number of alcohol-related consequences related to hepatic functioning and mental health (Witkiewitz et al., 2020). Importantly, drinking reduction with the aim of harm reduction might be more readily achieved and maintained than abstinence for some individuals, although Witkiewitz et al. (Witkiewitz et al., 2020) note that complete abstinence has still been associated with better outcomes.

The Meyerhoff-Durazzo Study

Whether this revised concept of recovery could apply to changes in brain structural condition was tested with a novel analysis conducted by Meyerhoff and Durazzo (Meyerhoff and Durazzo, 2020). These authors repurposed their legacy structural MRI brain data to test regional brain volume differences in individuals with diagnosed AUD who had undergone the treatment programs at VA hospitals and area clinics. The original study (Durazzo et al., 2015) was longitudinal and examined only treatment-seeking volunteers who had maintained abstinence over 8 months and non-AUD control participants who underwent longitudinal assessment at the same initial and final MRI sessions as the AUD group. The abstaining AUD group showed significant increases in frontal gray and white matter volumes and shrinkage of lateral ventricular volumes from their initial days of sobriety but remained below tissue recovery levels of the non-AUD control group.

The re-analysis (Meyerhoff and Durazzo, 2020) used only the cross-sectional data from the examination time at 8 months post-treatment to test different WHO levels of relapse in comparison with abstinence. The study objective was to ask whether total abstention was required to have improvement in brain structure or whether some degree of relapse could also realize significant recovery of brain tissue. The new analysis reclassified the AUD group with the WHO risk drinking levels as follows. At the 8-month session, the relapsers were sorted into two WHO drinking level categories: low relapsing men drank on average less than 40 g of alcohol per day (about 3 drinks) over the 8 months (women <20 g per day) (N=16 men, 1 woman) and had reduced their drinking levels from pre-treatment levels by nearly 3 steps on average; and high relapsing men drank on average more than 40 g per day (women >20 g per day) (N=9 men, 2 woman) and had reduced their drinking levels from pre-treatment levels by an average of only 1.2 steps. The authors note that in European treatment trials, an endpoint criterion is a reduction in drinking by two levels rather than abstinence (Agency, 2000). Repurposing legacy data is not an easy task, but it was made possible by consistency in acquisition of brain structural imaging data with concurrently collected quantitative clinical descriptive and neuropsychological performance data. The new analysis revealed a strikingly hopeful finding for AUD treatment-seekers who reduced drinking either to “safe” levels, according to the Center for Disease Control [https://www.cdc.gov/alcohol/fact-sheets/alcohol-use.htm] and NIAAA [https://www.niaaa.nih.gov/alcohol-health/overview-alcohol-consumption/moderate-binge-drinking], or even to levels that exceeded those guidelines.

Using this re-classification, Meyerhoff and Durazzo found that only the high relapsers had significantly smaller gray matter volumes of the frontal cortex than either the low relapsers or the controls, whereas the low relapsers did not differ from controls. A similar trend was noted for the thalamus. The differences could not be attributed to number of drinking days but rather the amount, which translated into the WHO drinking levels metric. Despite the robust differences in regional brain volumes related to level of resumed drinking, cognitive testing showed only trends toward subgroup differences.

Frontal Systems Insult and Recovery

As Meyerhoff and Durazzo acknowledged, the frontal cortex--perhaps to a lesser extent the thalamus (Pfefferbaum et al., 2018b) —has been considered a principal target of chronic, excessive alcohol consumption. A reasonable chicken-and egg question has loomed over the assumption that brain tissue volume deficits, like other functional deficits, arise from excessive drinking or are pre-existing and perhaps contribute to impaired ability to inhibit drinking in the first place. Although evidence exists for brain functional and structural differences to distinguish positive compared with negative family history of alcoholism and therefore are pre-existing (Begleiter and Porjesz, 1990), equally compelling longitudinal evidence now abounds for excessive alcohol use to result in regional brain volume declines (Pfefferbaum et al., 2018b, Sullivan et al., 2018) and selective neuropsychological impairments (e.g.,Durazzo and Meyerhoff, 2020). Clearly, change toward normal levels of brain tissue volumes (Durazzo et al., 2015) and white matter integrity (Pfefferbaum et al., 2014) together with improvement in neuropsychological functioning beyond simple practice effects (Durazzo and Meyerhoff, 2020, Sullivan et al., 2000 ) have been observed with quantitative, longitudinal assessment and provides evidence for a role for changes in drinking levels to be causative. Conversely, prospective studies of adolescents before and after initiating appreciable alcohol drinking reported accelerated gray matter volume loss predominantly in frontal cortex compared with adolescents who remained no-to-low drinkers (Pfefferbaum et al., 2018a). What the actual biological mechanisms of change remain to be discovered and may rely on animal models (e.g., Zahr et al., 2013).

Limitations

Like most studies, this one has limitations, also noted by the authors. Salient ones are small sample sizes, too few women to test sex differences, lack of premorbid (pre-AUD) baselines, and unknown effects of lower yet continued drinking in the presence of comorbid health conditions. Another limitation is the recognition that treatment-seekers comprise only about 10% of those with AUD [https://www.niaaa.nih.gov/publications/brochures-and-fact-sheets/alcohol-facts-and-statistics], thereby restricting generalization of results (Fein et al., 2010). Of course, the findings need replication. Despite any shortfalls of this retrospective analysis of prospectively acquired data, the direction for future research and treatment studies is clear: reduction in alcohol consumption—short of abstinence— tailored to levels of the individual can result in improvement in brain structural, cognitive, and mental health (cf.,Witkiewitz et al., 2020).

Ramifications for Concepts of Relapse

A desirable endpoint for treatment, then, appears to be harm reduction, which may be attainable and sustainable by some who will never achieve abstinence. From a clinical perspective, as reminded by Gastfriend et al. (Gastfriend et al., 2007) and McLellan (McLellan, 2007), goals must be set at treatment or study initiation with explicit consideration of the current level of functioning and intention of drinking reduction of individual participants. A goal may be “controlled drinking” at levels below those reported pre-treatment, and the goals may change, perhaps toward further reduction, over time. Indeed, a change in goals may occur with improved brain structure and function that can accompany drinking reduction or sobriety and possibly enable recovery (cf.,Le Berre et al., 2013). The likelihood of this possibility is enhanced by recognizing that frontal cortical systems, which support cognitive functions of inhibition, decision making, and problem solving, are the ones likely to be affected with drinking and recover with drinking reduction or sobriety. Exemplary evidence for these points derives from the observation that dysregulation of ventromedial prefrontal cortical activity measured with functional MRI 5 weeks into treatment-enforced sobriety was predictive of early post-treatment relapse and severity of relapsed drinking (Seo et al., 2013).

Finally, we should note that some proportion of people who struggle with AUD reduce drinking or stop it altogether without treatment (Humphreys and Bickel). One might wonder how this could be accomplished. Considering treatment-free recovery through the lens of the Meyerhoff-Durazzo finding suggests the hypothesis that such recovery might occur in steps, where, for example, a sustained drinking holiday initiates brain structural recovery that in turn enables resilience and resistance to drinking. As alcohol-free or alcohol-reduced days proceed, more brain structural and functional recovery ensues. In this way, what typifies the downward spiral and self-perpetuating cycle of addiction reverses itself (Koob and Le Moal, 2006, Sullivan and Pfefferbaum, 2019). To the extent that the frontal cortex is selectively affected by AUD (Meyerhoff and Durazzo, 2020), is vulnerable to the age-alcoholism interaction (Sullivan et al., 2018), and plays a role in AUD recovery raises a further question regarding whether older men and women who seek sobriety can show the same resilience in recovery as their younger counterparts.