Persisting Differences in the Sensitivity to the Effects of Alcohol: What We Know and Where to Go From Here

A large body of work has examined acute response to alcohol as a predictor of heavy alcohol use and alcohol use disorder (AUD) (1). However, the article by King et al. (2) is one of the few (3) to re-examine an individual’s response to alcohol repeatedly over time. In this commentary, I first review the notable accomplishments of this article and then put forth future directions that I believe will further build off of this line of important research.

The study by King et al. (2) has numerous methodological and theoretical strengths. The researchers used a well-designed, tightly controlled, within-subjects alcohol administration protocol. The study also had high retention rates over a 5-year follow-up period (87.6% of eligible subjects). The authors clearly delineated several specific patterns of alcohol response that could be found at follow-up based on neurobiological theories of addiction: tolerance (reduced alcohol response to negative effects), sensitization (increased stimulating response), early stage of the allostasis model (sensitivity is maintained), and incentive-sensitization (alcohol wanting increases). Analyses examined these possible prospective outcomes between heavy (n = 86) and light (n = 70) drinkers (categorized at baseline) as well as within the heavy drinkers based on high, intermediate, and low AUD symptoms.

These findings contribute significantly to the extant literature in several ways. First, consistency in a subject’s alcohol response over a 5-year period was demonstrated. Second, the authors found support for the early stage of the allostasis model (where sensitivity to rewarding effects is maintained), with the heavy drinking group continuing to experience heightened sensitivity to the stimulating and rewarding effects of alcohol compared with the light drinking group. The heavy drinking group also continued to experience lower sensitivity to the sedating effects and exhibited a lower cortisol response compared to the light drinkers.

A third important contribution of this article is that it builds on a growing body of research that examines sensitivity to the stimulating effects of alcohol. Whereas this article examined both stimulation and sedation, historically more attention has been paid to the low responder model (LRM). The work by Schuckit and others has led to low response to alcohol being considered an endophenotype for AUD and has shown that individuals who experience a dampened response to the sedating or negative effects of alcohol are at increased risk for AUD. Low response has been examined through the use of multiple methods (e.g., body sway, subjective report, cortisol, functional magnetic resonance imaging) and has been shown to be heritable and to predict AUD prospectively.

An alternative model is the differentiator model (DM) (4). DM extends the LRM by positing that both reduced sensitivity to the negative effects of alcohol on the descending limb of the blood alcohol curve and increased sensitivity to the rewarding effects on the ascending limb of the blood alcohol curve are important in understanding risk for AUD. A meta-analysis of 29 studies (1) found support for both the DM and the LRM in different populations. The LRM was supported when examining group differences based on family history status, such that men with a positive family history of alcoholism experienced reduced subjective response (for sedation and also tentatively stimulation). However, when examining groups as a function of light versus heavy drinking, as in the article by King et al. (2), there was support for the DM. One of the conclusions made in the meta-analysis (1) was that these subjective responses may occur in different at-risk populations (positive family history [FH+] vs. heavy drinker) or predict different pathways to AUD [see also (5)].

Given the potential for situational and contextual factors to influence the drinking experience, I contend that conducting research outside of the laboratory, away from tightly controlled experimental paradigms, and focused on naturalistic responses to alcohol and decisions made in real time through the use of ecological momentary assessment (EMA) is an important next step in this line of research. Studies using EMA methodology may further our understanding of how, why, for whom, and which acute alcohol effects (e.g., stimulation vs. sedation, ascending vs. descending limb effects, FH+ vs. heavy drinker) relate to problematic alcohol use. Therefore, these types of studies may clarify LRM versus DM debates.

Several recent examples illustrate the potential benefits of data collection in the natural environment for understanding alcohol response. First, in a heavy drinking adolescent sample (6), experiencing less “high” from alcohol at the end of a drink was associated with more subsequent alcohol use, which is consistent with the LRM. In contrast, increased stimulation was associated with less subsequent alcohol use, which is inconsistent with the DM. Second, in an adult sample (7), a retrospective report of low response to alcohol (assessed via the Self-Rating of the Effects of Alcohol measure) was related to drinking more drinks per occasion across a 21-day period. Interestingly, after controlling for the differences in amount of alcohol, low responders were less likely to experience hangover. Research has also shown that response to alcohol assessed in a subject’s own environment varies as a function of context (e.g., people report feeling more “buzzed” at a restaurant), a subject’s drinking companions, and what time of day a subject is drinking (8). Increasing our understanding of responses to alcohol in a subject’s actual drinking environment is an important next step to furthering our understanding of risk that may be conveyed by this endophenotype.

Despite knowing that response to alcohol is influenced by genetics, limited research has examined if individuals with certain genetic variants respond differently to alcohol in their own environment and in certain contexts. One such study (9) found differences in naturalistic responses to alcohol as a function of variants in the OPRM1 and DRD4 genes but did not find significant gene × environment (i.e., location, whether or not others were drinking) interactions in predicting response to alcohol. However, laboratory-based alcohol administration work has found that DRD4 is related to social bonding after alcohol consumption relative to placebo (10). These laboratory findings highlight the possibility that individuals with the seven-repeat allele DRD4 variant may experience more of the rewarding effects of alcohol when drinking with friends compared with when they are alone or relative to individuals without this variant. This hypothesis has not been directly tested. Increased research using EMA methods that link directly with laboratory-based response is needed to address existing inconsistencies and foster understanding of how individual differences in response to alcohol confer risk in a subject’s own environment.

Another benefit of EMA research on response to alcohol is that it can facilitate research on DM and LRM in populations that are largely excluded from alcohol administration research. For example, EMA allows for the real-time assessment (as opposed to retrospective reports) of alcohol response in adolescents (6) who are under the legal drinking age under the age of informed consent, who might otherwise be excluded from alcohol administration studies. Thus, EMA would allow for a prospective examination of response to alcohol similar to the study by King et al. (2) only at younger ages and closer to the initiation of alcohol use; this may be particularly relevant when examining theories of addiction development. For example, whereas King et al. (2) did not find that the heavy drinkers experienced sensitization across a 5-year period, it may be that these same drinkers experienced sensitization in adolescence, which increased their heavy drinking. Similarly, the sample in their study excluded individuals with an AUD, a common exclusion in alcohol administration research. These may be the individuals who might have a later stage allostasis, where reward sensitivity would diminish over time.

A final fit of real-world assessment of response to bene alcohol is that temporal associations between acute alcohol response and alcohol consumption and problems could be disentangled. It may be that in certain situations increased sensitivity to the stimulating effects of alcohol is related to consuming more alcohol and in other situations reduced sensitivity to the sedating effects is related to increased consumption. Furthermore, EMA research could examine both the between-subject (e.g., FH+ vs. FH−; heavy drinker vs. light drinker) and the within-subject variability in how a subject responds to alcohol.

In conclusion, the article by King et al. (2) and numerous other studies from this research group highlight the importance of sensitivity to the stimulating effects of alcohol. Their current study demonstrated persistence in differences in both stimulation and sedation between heavy and light drinkers. This study did not find support for the development of tolerance (with the exception potentially being the heavy drinking low AUD subgroup) or sensitization over time. These results are highly informative and add to the DM literature as well as the overall literature, showing that response to alcohol conveys risk for problematic drinking over time. I look forward to additional research examining stimulation and sedation as well as physiologic responses (e.g., heart rate variability) in naturalistic settings to address issues such as how increased stimulation to alcohol relates to drinking heavily within a drinking event or if tolerance to the rewarding effects develops after repeated heavy drinking days.