|
Article Excerpt
THE LOW LEVEL OF RESPONSE (LR) TO alcohol is an important genetically influenced characteristic that predicts heavy drinking and alcohol-related problems, including alcohol-use disorders (AUDs; Barr et al., 2003; Heath et al., 1999; Schuckit, 2002; Wilhelmsen et al., 2003). Less response at a given blood alcohol concentration (BAC) is hypothesized to enhance the probability of drinking more alcohol to get the desired effects and to offer less feedback of impending intoxication per drink consumed (Schuckit, 2002; Schuckit et al., 2001b). There have been four follow-up studies of LR, all of which have reported that lower values earlier in life and before the onset of AUDs are associated with a greater risk for later heavier drinking and alcohol problems (Heath et al., 1999; Rodriguez et al., 1993; Schuckit and Smith, 2000; Volavka et al., 1996).
Alcohol challenge-based LR meets the Gottesman and Gould (2003) definition of an endophenotype as a low response to alcohol is demonstrated before AUDs develop, is genetically influenced (with heritabilities of .5-.6; Heath and Martin, 1992; Viken et al., 2003), and predicts later heavier drinking and alcohol problems. However, the precise underpinnings of LR are difficult to establish in humans. This occurs in part because alcohol challenges are rarely (if ever) carried out before the age of 18, and, although higher and lower AUD risk groups usually have been matched on drinking parameters (Schuckit and Gold, 1988), it is still possible that the LR differential between groups might reflect differences in tolerance that developed before testing. In this regard, it is reassuring that the greatest ability of LR to predict future AUDs rests with the lighter drinkers (e.g., those consuming an average of two drinks per occasion, with an average intake of one time per week; Schuckit et al., 2006). Those results indicate that it might be more difficult to identify a low LR as a risk factor for AUDs in younger heavier drinkers, perhaps because some individuals who should not be classified as low LR were placed into that category erroneously through acquired tolerance.
The ideal study establishing the underpinnings of LR in humans might require at least two sets of experiments in very young subjects who have never consumed alcohol. The first experiment would involve a classic alcohol challenge to determine whether the low LR is seen in a high-risk group (e.g., offspring of alcoholics), with future follow-ups determining alcohol outcomes. The second experiment might use intravenous alcohol to study intrasession tolerance by establishing a plateau of blood levels and observing how this affects performance on a task over time (Ramchandani et al., 2006). However, Human Protection Committees are likely to be reluctant to approve giving intoxicating doses of alcohol to such young individuals, especially in light of the fact that many drinkers have had their first experience with this drug by age 13 (Johnston et al., 2005). The relationship of LR to the sensitivity or acute tolerance to alcohol (as distinct from intersession tolerance) is a bit easier to study in animal models, with most investigations, but not all (Colombo et al., 2000), indicating that murine strains and selectively bred lines that demonstrate higher voluntary consumption of alcohol exhibit low sensitivity to alcohol and/or evidence more rapid or more intense acute tolerance when initially exposed to alcohol (Baldwin et al., 1991; Kurtz et al., 1996; Lumeng et al., 1982; Tabakoff and Ritzmann, 1979; Tampier and Mardones, 1999; Tampier et al., 2000; Waller et al., 1983).
Additional data to address this question might accrue in human research if alternative methods for determining LR could be developed. Although most evaluations of LR have used alcohol challenges (Eng et al., 2005; Erblich and Earleywine, 1999; Pollock, 1992; Wilhelmsen et al., 2003), that protocol is expensive and time consuming. In addition, such challenges cannot be carried out in an experiment with subjects too young to consume alcohol, individuals with some medical illnesses, people taking medications that might interact adversely with alcohol, or recovering alcoholics (Schuckit and Gold, 1988; Wall et al., 1999). The ability to test younger subjects is particularly important, because most individuals begin drinking in their early to middle teen years, and it is best to evaluate LR as early in the drinking career as possible (Schuckit, 2002).
The Self-Rating of the Effects of Alcohol (SRE) questionnaire was developed to facilitate gathering data on LR from subjects with more diverse backgrounds than can be tested with an alcohol challenge (Schuckit et al., 1997a,b). The questions were developed by observing the items most frequently endorsed in laboratory testing and, a priori, selecting those believed most likely to be remembered by subjects. Although chosen on the basis of clinical experience and a desire to create a retrospective measure that is as simple and brief as possible, the questions have high internal consistency, as will be described here. The SRE asks a person to think back to their drinking experiences earlier in life and to report the number of standard drinks (i.e., about 10 g of ethanol) required for any of up to four effects. These include the number of drinks for the first effect (or "buzz"); the drinks required for slurred speech; the amount of alcohol needed to produce stumbling gait, if this had been experienced; and how much alcohol was required to cause them to fall asleep against their wishes. The score for the earlier-in-life SRE is generated by totaling the number of drinks involved and dividing the sum by the number of effects reported.
The SRE score has a 1-year retest reliability of approximately .8, evidences internal coherence with a Cronbach's [alpha] of .97 (Schuckit et al., 2005b), and is relatively stable over time (Schuckit and Smith, 2004). At the higher and lower extremes, SRE scores identify up to 80% of individuals with very low LRs on alcohol challenges (Schuckit et al., 1997a,b). The correlation between an SRE-based retrospective LR score in 35-year-old subjects and results from their alcohol challenges at about age 20 ranges from about .4 to .6 for those with clearly high or low scores (Schuckit et al., 1997a,b, 2001b). In studies of men and women in the United States and Europe, a lower SRE score from the first five or so times of drinking (First 5) was associated with a fourfold higher rate of AUDs and correlated with both the maximum number of drinks consumed and with alcohol-related problems (Daeppen et al., 2000; Schuckit et al., 1997a, 2001b, 2003). A recent 5-year follow-up of almost 100 men and women revealed that SRE-based LR scores predicted later heavier drinking and alcohol problems (Schuckit et al., submitted for publication). In addition, LR scores as measured by the SRE may be genetically influenced, with correlations for the early drinking scores among first-degree relatives of about .20, second-degree relative correlations of about .12, but figures close to for unrelated individuals (Schuckit et al., 2001a, 2005c). Genetic linkage analyses using the SRE have highlighted several chromosomal regions as potentially related to LR, including some on chromosome 1 and chromosome 10 (Schuckit et al., 2001a). Additional support for biological underpinnings of SRE-based LR comes from functional magnetic resonance imaging studies that indicate that, whether measured by SRE or alcohol challenges, a low LR appears to relate to the intensity of signal in the prefrontal cortex and parietal cortex during a usual memory task (Paulus et al., 2006; Tapert et al., 2004).
However, even for the SRE, relatively few...
|
