Sweet intake, sweet-liking, urges to eat, and weight change: Relationship to alcohol dependence and abstinence

Addictive Behaviors 31 (2006) 622 – 631

Sweet intake, sweet-liking, urges to eat, and weight change: Relationship to alcohol dependence and abstinenceB Dean Krahn a,b,*, Jennifer Grossman a,b, Henry Henk a,b, Mary Mussey a,b, Ross Crosby c, Blake Gosnell c a

Wm. S. Middleton Memorial Veterans Hospital, Psychiatry Service, 2500 Overlook Terrace, Madison, WI 53705, Madison, Wisconsin, United States b University of Wisconsin Medical School, Madison, Wisconsin, United States c Neuropsych Research Institute (NRI), Fargo, North Dakota, United States

Abstract Introduction: Linkages between alcohol dependence (AD) and abstinence and aspects of food ingestion and preference have been described in animals and humans, including (1) eating sweets decreases urges to drink alcohol; (2) preferences for highly sweet tastants is associated with alcohol dependence; and (3) food deprivation leads to increased alcohol intake. Methods: We randomly assigned AD subjects in early abstinence to 3 different sets of dietary instructions (eat sweets for alcohol urges; eat a balanced diet; avoid sweets). We compared the groups on urges for alcohol, alcohol consumption, weight, and sweet preference at baseline, one, and six months. We also compared these AD subjects with light-drinking C’s and compared AD subjects who remained abstinent for 6 month follow-up with nonabstinent AD subjects. Results: Recruited AS subjects, 38 of 68, completed 6 month follow-up; 27 of 36 C’s completed the follow-up. 21 AD’s were abstinent while 17 were non-abstinent. There was no effect of dietary recommendations on urges to drink or alcohol consumption. AD’s were more likely than C’s to prefer highly sweet tastants. The proportion of AD’s preferring the sweetest tastant decreased over time. AD’s gained more weight than C’s over the 6-month follow-up.

B

This research was conducted at the Wm. S. Middleton Memorial Veterans Hospital, and was supported by a VA Merit Review grant to Dr. Krahn. * Corresponding author. Tel.: +1 608 280 7015; fax: +1 608 280 7203. E-mail address: [email protected] (D. Krahn). 0306-4603/$ - see front matter D 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.addbeh.2005.05.056

D. Krahn et al. / Addictive Behaviors 31 (2006) 622–631

623

Discussion: While the use of sweets did not affect urges to drink or drinking, important relationships between sweet preference, weight gain, and alcohol dependence or abstinence were identified. D 2005 Elsevier Ltd. All rights reserved. Keywords: Sweet-liking; Alcohol; Dependence; Abstinence; Diet; Urges

1. Introduction Several lines of research have documented linkages of possible clinical importance between sweet preference and eating behaviors and urges to drink and actual ingestion of alcohol in alcohol dependent subjects. The data supporting each proposed linkage between eating-related parameters and alcoholrelated parameters is reviewed below. 1.1. Ingestion of sweet foods or fluids can decrease urges to drink alcohol Animal and human studies show that provision of sweet, preferred substances as alternatives to alcohol results in decreased alcohol consumption. In the btwo-bottleQ paradigm for measuring alcohol preference, the provision of saccharin, sucrose, or fat solutions as a bthird choiceQ resulted in a decrease in alcohol consumption by laboratory animals (Lester & Greenbert, 1952). Similarly, the availability of a saccharin/glucose solution decreased drug self-administration (Carroll & Boe, 1982; Carroll, Lac & Nygaard, 1993) In humans, Yung, Gordis and Halt (1983) found that alcoholics with the longest periods of sobriety report markedly increased intake of sugar. While a definitive, prospective study of the effect of the ingestion of sweet-tasting food on urges for alcohol has not been done, alcohol-dependent patients early in abstinence are often told to use this dietary practice. For example, in the twelve-step facilitation arm of Project MATCH (Nowinski, Baker & Carol, 1995), clinicians recommended that subjects follow advice found in bLiving SoberQ (Alcoholics Anonymous, 1975) which states, bWe can only pass on the word that thousands of us—even many who said they had never liked sweets—have found that eating or drinking something sweet allays the urge to drink.Q Curiously, the same lesson in bLiving SoberQ notes that some alcoholics risk becoming ice cream or candy baddicts,Q and another book commonly used in alcohol treatment (Gorski & Miller, 1986) suggests that b(The alcoholic) should specifically avoid foods that cause stress such as concentrated sweets and caffeine.Q Unfortunately, virtually no data on the efficacy or risks of these recommended dietary strategies exist. 1.2. Preference for highly sweet tastants is associated with preference for alcohol and other drugs Rats with high preference for sweet saccharin solutions consume more alcohol than rats with low preference for such solutions (Gosnell & Krahn, 1992; Bell, Gosnell, Krahn & Meisch, 1994). Conversely, in rats bred for high and low alcohol self-administration, corresponding high and low intakes of saccharin have been noted (Kampov-Polevoy, Kasheffskaya & Sinclair, 1990; Overstreet et al., 1993; Sinclair, Le & Kiianmaa, 1989; Stewart, Russell, Lumeng, Li & Murphy, 1994). For

624

D. Krahn et al. / Addictive Behaviors 31 (2006) 622–631

example, Overstreet and colleagues demonstrated that saccharin and alcohol intakes were highly correlated (r N 0.61) over seven strains of rats selected for a wide range in alcohol preferences. It is possible that the correlations between sweet taste preference and alcohol preference reflects a more general phenomenon linking preferences for palatable food and fluids with preferences for other drugs. Krahn and Gosnell (Krahn & Gosnell, 1991) reported that rats that prefer a high-fat diet consumed significantly more alcohol than carbohydrate-preferring rats. Similarly, MarksKaufman and Lipeles (1982) found that those rats that drink a morphine solution consumed more dietary fat. However, the linkage between sweet preference and alcohol or other drug consumption is not due to similar tastes as sweet preference also predicts the rate of self-administration of intravenous morphine, amphetamine, and cocaine (DeSousa, Bush, & Vaccarino, 1996; Gosnell, Lane, Bell & Krahn, 1995). In humans, some observations suggest a relationship between sweet, preferred food preference and alcohol and other drug intake. Kampov-Polevoy, Garbutt and Janowsky (1997) reported that 65% of recently detoxified alcoholics preferred highly sweet sucrose solutions (0.83 M) compared with 16% of a control population. Janowsky et al. (2003) reported that a similar preference characterized cocaine dependent subjects. Subsequent studies by KampovPolevoy et al. (Kampov-Polevoy et al., 1997; Kampov-Polevoy et al., 1990) linked preference for highly sweet tastants to a positive family history for alcohol dependence. However, it should be noted that not all investigators have replicated these findings (Kranzler et al., 2001 and BoguckaBonikowska et al., 2001). 1.3. Weight change is associated with changes in alcohol intake Deprivation of food sufficient to cause weight decrease to 75% of baseline results in a potent increase in the rate of self-administration of alcohol and other drugs (Carroll & Meisch, 1984). Restoration of weight reverses this weight loss-induced stimulation of intake (Meisch, 1977). Deprivation of a sweet taste without caloric deprivation also increases the rate of self-administration of drugs (Carroll & Boe, 1982; Carroll et al., 1993). We have reported that human dieters who often avoid sweet, rich foods also report increased alcohol and other substance use and abuse (Krahn, Kurth, Demitrack & Drewnowski, 1992). Humans starved in research or in famines increase their intake of nicotine, caffeine, and cocaine (Franklin, Schiele, Brozek & Keys, 1948; Hanna & Hornick, 1977). We elected to test the efficacy of the dietary recommendations to use or avoid sweets to promote abstinence in alcohol dependent patients attempting to stop drinking. We hypothesized that alcohol dependent (AD) subjects receiving the recommendation to eat sweets in response to urges to drink would experience decreased urges to drink and decreased drinking compared to either a recommendation to eat in a balanced manner or avoid sweets during early abstinence. We also hypothesized that there might be subgroups of AD subjects characterized by different sweet preferences that might be differentially responsive to using sweets to dampen urges to drink. Simultaneously, we assessed the relationship between sweet preference, weight changes and alcohol dependence or abstinence. Our secondary hypotheses regarding these variables were: (1) that a higher percentage of AD subjects would prefer highly sweet tastants than control (C) subjects; (2) that the percentage of AD subjects and C subjects that preferred highly sweet tastants would remain stable over time and abstinence; (3) that AD subjects attempting to remain abstinent would gain more weight than C’s over the course of 6 month follow-up.

D. Krahn et al. / Addictive Behaviors 31 (2006) 622–631

625

2. Methods 2.1. Subjects Alcohol dependent subjects (AD) were recruited from the Addictive Disorders Treatment Program at the William S. Middleton Memorial Veterans Hospital in Madison, WI and from a local, private addiction treatment program. The majority of subjects were outpatients with the rest being in a residential program with frequent community, real-world contacts. All male subjects were used to avoid women who might limit sweet use due to worries about weight gain or who might be at greater risk for development of eating disorders. Importantly, neither program made routine recommendations regarding using or avoiding sweets. Inclusion in the study as an AD subject required: (1) age between 18 and 65; (2) passing reading test given to all new program entrants at a 6th grade or higher level; (3) alcohol dependent by DSM-IV criteria as determined by responses to the Mini International Neuropsychiatric Interview (MINI) (Sheehan & Lecrubier, 1994); (4) between 4 and 30 days from last drink with Clinical Instrument for Withdrawal Assessment-Alcohol Revised (CIWA-AR) score = 0; and (5) provision of informed, written consent. Exclusion criteria included: (1) active psychotic or suicidal state; (2) medical condition for which use of sweet, rich food would be contraindicated such as insulin-dependent diabetes; (3) on naltrexone, antidepressants (other than trazodone for sleep), and Li; (4) meeting DSM-IV criteria for an eating disorder on MINI; (5) attempting to quit smoking or opiates. Informed consent was then obtained. 2.2. Independent variables—dietary recommendation intervention and time Subjects were randomly assigned to groups that received different training regarding dietary interventions. However, based on response to the question bDo you think that eating sweets will decrease your urge to drink alcohol?Q with Yes or No, the randomization process was stratified so that each group had similar numbers of subjects that endorsed each response. Subjects assigned to each group viewed videotaped 15 min lessons documenting the presumed advantages of using sweets, eating healthy, or avoiding sweets. Those told to eat sweets were instructed to keep a supply of these sweets (provided by the study) with them at all times. They self-monitored for any urges for alcohol, recorded those urges and ate sweets to dampen the urge. 2.3. Dependent and mediating variables Urges for alcohol, moods, and alcohol consumption were assessed by use of traditional questionnaire instruments at baseline, one month, and six months. Urges for alcohol were assessed using the Alcohol Urges Questionnaire (AUQ) and a 10-point visual analog scale (VAS) of urge severity anchored on the low end by bno urgeQ and on the opposite end by bmost severe possibleQ 4. Moods were assessed with the Positive and Negative Affective Scale (PANAS) (Watson, Clark & Tellegen, 1988). Alcohol use was measured with the Time Line Follow Back for Alcohol Use (TLFB) (Sobell & Sobell, 1992). AD subjects were classified as abstinent (no drinks on TLFB) vs. non-abstinent (1 or more drinks on TLFB). Drop-outs were considered non-abstinent. Data from study drop-outs were not included as we predicted differences in changes in urges, taste, and weight over time making a last observation carried forward analysis inappropriate.

626

D. Krahn et al. / Addictive Behaviors 31 (2006) 622–631

Preference for sweet tastants was assessed using the method of Kampov-Polevoy, et al. (1997). Sweet solutions were made by dissolving sucrose in room temperature distilled water to achieve sucrose concentrations of 0.05, 0.10, 0.21,0.42, and 0.83 M. The taste stimuli were presented in blocks that were replicated five times with the order of presentation randomized within blocks. Subjects rated each sucrose solution on visual analog scales (VAS’s) for sweetness intensity and hedonic value. One VAS was anchored at one end by bnot sweet at allQ and at the other by bextremely sweetQ. The other VAS was anchored at the ends by bdisliked very muchQ and bliked very muchQ and anchored at the midpoint by bneither liked nor disliked.Q The subject sipped solutions from a tray of numbered medicine cups containing 10 ml of sucrose solution, swished them about in their mouth, then spat out the solutions. They were not permitted to swallow or taste any solution more than once. Subjects were classified as sweet-likers if they maximally liked the sweetest (i.e. most concentrated) sucrose solution. At baseline, 4 weeks, and six months, subjects were weighed in light clothing appropriate for indoors without shoes. Control subjects (C) were recruited from ambulatory patient areas in VA and a nearby hospital affiliated with the private addiction program. C subjects met all inclusion and exclusion criteria except that they did not meet criteria for past or present alcohol or other substance dependence or abuse (except tobacco dependence) and did not drink more than 14 drinks per week. 2.4. Statistical analysis Analyses in which the three dietary recommendation groups were compared to test our primary hypotheses used subjects who had completed at least the baseline and one month evaluation prior to drop-out. Those subjects who dropped out between the one month and six month assessments had their last observation carried forward used in analyses. Because our secondary hypotheses focused on the change in measures such as preference for highly sweet tastants and weight over time of abstinence, we included only subjects who completed all three time points. Analyses in which the three dietary recommendation groups were compared to test our primary hypotheses used subjects who had completed at least the baseline and one-month evaluation prior to drop-out. Those subjects who dropped out between the one month and six month assessments had their last observation carried forward used in analyses. Because our secondary hypotheses focused on the change in measures, such as preference for highly sweet tastants and weight over time of abstinence, we included only subjects who completed all three time-points. 2.5. Palm pilot data analysis Dietary recommendation groups were compared for levels of AUQ over 56 possible reporting times. The unit of analysis was the individual participant’s observation, but the analysis accounted for the nesting of multiple observations within subject (repeated measures), using a random coefficients model (i.e., hierarchical linear model). Participants are assumed independent of each other; however, observations are correlated within each participant due to the repeated measures. The analysis treats AUQ as a continuous dependent variable to be modeled as a function of time, dietary group, and a timeby-group interaction. The intercept in the model is specified in such a way that it represents a baseline or initial level (by coding time= 0 for the 1st observation time). The estimated average change over time for each cluster and difference in the rate of change overtime between clusters are obtained by the coefficients on the time, cluster and the interaction term (time-by-cluster). Analyses were conducted

D. Krahn et al. / Addictive Behaviors 31 (2006) 622–631

627

using SAS PROC MIXED (SAS Institute Inc., Cary, NC). PROC MIXED has an advantage over ANOVA repeated measure in it provides a better mechanism for handling missing data and unbalanced designs. The empirical option was also specified, which provides robust estimates of the standard errors (Huber, 1967). 2.6. Survey and sweet preference data analysis Comparisons of age, AUQ, and the percent preferring highly sweet tastants were made for C vs. AD and abstinent vs. non-abstinent AD subjects were completing the 6-month follow-up. Comparison of age and AUQ were made as a change from baseline value use a standard t-test. Percent of subjects preferring sweets was compared at each survey period (baseline, 1-month and 6-months) using a Chi-squared test.

3. Results Of 68 AD subjects who were entered into the study after informed consent, 38 completed the 6 month follow-up successfully. Of these original 68 subjects, 18 were randomized to the balanced diet recommendation group, 13 were assigned to the avoid sweets group, and 24 were randomized to the sweet use group. Of the 37 completers, 19 were in the sweet use group, 7 were in the avoid sweet group, and 11 were in the balanced diet recommendation group. The randomization resulted in groups that did not differ in age, weight, level of urge for alcohol, negative affect, or percentage of sweet-likers at baseline. Fig. 1 depicts the level of AUQ-measured urge to drink alcohol at 4 time points each day for the subjects in each of the three dietary recommendation groups. As can be seen, there is no difference between urge levels at baseline or in change in urge level over the two weeks of EMA assessment. Mean Change in AUQ Over Time 25

20

15

AUQ

adlib avoid pro

10

5

0 0

2

4

6

8

10

12

14

Day Number

Fig. 1. The level of urges to drink as measured by AUQ in the EMA-phase of the study in the three treatment groups is shown. No significant differences were seen.

628

D. Krahn et al. / Addictive Behaviors 31 (2006) 622–631

Table 1 Changes in weight and sweet-liking over time in alcohol dependent and control subjects Controls In at 6-month N Weight Baseline 1-month 6-month Sweet liking Baseline 1-month 6-month*

AD In at 6-month

27

Difference

P-value

38

190.44 189.74 190.89

195.79 200.41 203.21

37.04% 37.04% 40.74%

71.05% 59.46% 50.00%

5.32 6.97

0.0005 0.0249

34.02% 22.42% 9.26%

0.0058 0.0786 0.4922

*Several AD have no sweet-liking data at 1-month (27 vs. 37) and (27 vs. 30) at 6-months.

AUQ levels for the three groups were: 14.81 for the sweet use group; 11.98 for the balanced diet group; and 13.06 for the avoid sweet group ( F 2,42 = 1.04, p = 0.36, ns). The avoid sweet group had an 83% abstinence rate on the 6-month TLFB assessment vs. 58% for the balanced diet group and 53% for the sweet use group (x 22 = 1.79, p b 0.41, ns). Likewise, there was no significant difference in number of drinking days between 3 and 6 months. Because of the small number of completers of the 6-month follow-up, the study was underpowered for detection of significant effects of dietary recommendations at 6-months. Likewise, further analysis didn’t reveal subgroups of AD subjects characterized by differences in sweet preference, urge for palatable foods, or weight change that were particularly responsive to dietary interventions. Again, due to small numbers of completers, power was not sufficient to be very sensitive. The proportion of AD subjects (71%) who expressed maximal liking for the tastant with the highest sucrose concentration at baseline was significantly higher (t 63 = 2.86, p = 0.0006) than the proportion of C subjects who preferred that concentration of sucrose (37%). However, by 6 months, there was no significant difference (t 55 = 0.69, p = 0.4922) between the percentage of C subjects (40.7%) and AD subjects (50%) that maximally preferred the sweetest tastant. Importantly, AD subjects who failed to maintain abstinence through 6 months were much more likely to prefer the sweetest tastant than were those who maintained abstinence. These proportions were

Table 2 Changes in weight and sweet-liking over time in abstinent and non-abstinent AD subjects AD abstinent N Weight Baseline 1-month 6-month Sweet liking Baseline 1-month 6-month*

22 201.16 207.02 210.41 54.55% 47.62% 33.33%

AD non-abstinent

Difference

P-value

16 188.41 191.31 193.31 93.75% 75.00% 75.00%

*Several AD have not sweet-liking data at 6-months (18 vs. 13) and (21 vs. 16) at 1-month.

2.95 4.34

0.1884 0.3906

39.20% 27.38% 41.67%

0.0075 0.0979 0.0251

D. Krahn et al. / Addictive Behaviors 31 (2006) 622–631

629

significantly different at both baseline (t 36 = 2.83, p = 0.008)) and 6 month assessments (t 28 = 2.37, p = 0.25) (Table 2). Weight changes differed significantly (1-month: t 85 = 3.60, p = 0.0005; t-month: t 64 = 2.30, p = 0.25) between AD and C subjects at the 1 month and the 6 month follow-ups. Subjects attempting to maintain abstinence gained 7.42 pounds over 6 months while C subjects increased only .45 pounds (t 63 = 2.30, p = 0.25) (Table 1). Abstinent AD subjects gained 9.25 pounds in 6 months vs. 4.9 pounds gained by non-abstinent AD subjects, but this was not a significant difference (t 36 = 0.869, p = 0.3906) (Table 2).

4. Discussion This randomized, controlled trial comparing (1) recommendation to use of sweets to decrease urges for alcohol in early abstinence; (2) avoid of sweets in early abstinence; and (3) eat a healthy, balanced diet did not find a significant effect of these recommendations on urges to drink or actual consumption of alcohol. Given that we cannot find any other RCT data supporting the benefits of these recommendation, and our findings don’t support these recommendations, we suggest that texts and guidelines (Nowinski et al., 1995; Gorski & Miller, 1986) used in substance dependence treatment program be changed. It is possible, however, that our methods failed to detect a small or transient effect on urges. A larger study could have detected a smaller effect, but, given our four times daily assessment of urges, any effect that was missed due to infrequent assessment would have been quite brief. While our primary hypothesis wasn’t supported, our secondary hypotheses were supported to some degree. Like Kampov-Polevoy et al. (1997, 2004) we found that AD subjects, during the first month of abstinence, reported maximal pleasant response to the sweetest tastant significantly more frequently than C subjects. Importantly, however, our retests of sweet preferences at 1 month and 6 months revealed that the percentage of AD subjects preferring the maximum concentration of sucrose decreased over time. We had predicted a trait-like stability in this characteristic. By 6 months, the percentage of AD subjects that preferred the sweetest tastant had decreased to 50%, a level that was not significantly different from controls. AD subjects who reported abstinence throughout the 6 months of the study were significantly less likely to prefer the maximally sweet tastant than were the AD subjects who did not maintain abstinence. Thus, sweet preferences should be tested as a predictor of future abstinence. Moreover, the relationship between these preferences and the process of relapse should be studied further. It is not clear what the mechanism of change in sweet preference would be. If these findings are replicated, peripheral and/or central changes in taste hedonics could be occurring. Weight gain has been anecdotally reported as a problem for those attempting to maintain abstinence. AD subjects clearly experienced a significant weight gain over the first 6 months of attempted abstinence. When we compared abstinent AD subjects to that of non-abstinent AD subjects, abstinent subjects gained nonsignificantly larger amounts of weight. These findings must be replicated and the mechanism of this weight gain must also be understood. Whether dieting or exercising during this period would avoid weight gain or hurt abstinence rates is unknown. Thus, our study did not find support for the currently recommended dietary interventions. However, it did find support for important interactions between taste and weight variables and the process of abstinence. These findings should be replicated and use of sweet preference or weight gain as prognostic factors should be assessed further.

630

D. Krahn et al. / Addictive Behaviors 31 (2006) 622–631

Acknowledgements We would like to thank Sarah Kerr and Kristin O’Leary for their expert technical support.

References Alcoholics Anonymous. (1975). Living sober. New York7 Alcoholics Anonymous World Services. Bell, S. M., Gosnell, B. A., Krahn, D. D., & Meisch, R. A. (1994). Ethanol reinforcement and its relationship to saccharin preference in Wistar rats. Alcohol, 11, 141 – 145. Bogucka-Bonikowska, A., Scinska, A., Koros, E., Polanowska, E., Habrat, B., Woronowicz, B., et al. (2001). Taste responses in alcohol-dependent men. Alcohol and Alcoholism, 36(6), 516 – 519. Carroll, M. E., & Boe, I. N. (1982). Increased intravenous drug self-administration during deprivation of other reinforcers. Pharmacology, Biochemistry and Behavior, 17, 563 – 567. Carroll, M. E., Lac, S. T., & Nygaard, S. L. (1993). Autoshaping i.v. cocaine self-administration in rats: Effects of nondrug alternative reinforcers on acquisition. Psychopharmacology, 11(1–2), 5 – 12. Carroll, M. E., & Meisch, R. A. (1984). Increased drug-reinforced behavior due to food deprivation. In T. Thompson, J. E. Barrett, & P. B. Davis (Eds.), Advances in behavioral pharmacology, vol. 4 (pp. 47 – 88). New York7 Academic press. DeSousa, N. J., Bush, D. E. A., & Vaccarino, F. J. (1996). Individual differences in sucrose intake are predictive of intravenous amphetamine levels. Abstracts - Society for Neuroscience, 22, 705. Franklin, J. C., Schiele, B. C., Brozek, J., & Keys, A. (1948). Observations on human behavior in experimental semi starvation and rehabilitation. Journal of Clinical Psychology, 4, 28 – 45. Gorski, T. T., & Miller, M. (1986). Staying sober. Independence, MO7 Herald House/Independence Press. Gosnell, B. A., & Krahn, D. D. (1992). The relationship between saccharin and alcohol intake in rats. Alcohol, 9(3), 203 – 206. Gosnell, B. A., Lane, K. E., Bell, S. M., & Krahn, D. D. (1995). Intravenous morphine self-administration by rats with low vs. high saccharin preferences. Psychopharmacology, 117, 248 – 252. Hanna, J. M., & Hornick, C. A. (1977). Use of coca leaf in southern Peru: Adaptation or addiction. Bulletin on Narcotics, 29, 63 – 74. Huber, P. (1967). The behavior of maximum livelihood estimates under non-standard conditions: Fifth Berkeley Symposium in Mathematical Statistics and Probability. Berkeley7 University of California Press. Janowsky, D. S., Pucilowski, O., & Buyinza, M. (2003). Preference for higher sucrose concentrations in cocaine abusingdependent patients. Journal of Psychiatric Research, 37(1), 35 – 61. Kampov-Polevoy, A., Garbutt, J. C., & Janowsky, D. (1997). Evidence of preference for a high-concentration sucrose solution in alcoholic men. American Journal of Psychiatry, 154(2), 269 – 270. Kampov-Polevoy, A. B., Eick, C., Boland, G., Khalitov, E., & Crews, F. T. (2004). Sweet liking, novelty seeking, and gender predict alcoholic status. Alcoholism, Clinical and Experimental Research, 28(9), 1291 – 1298. Kampov-Polevoy, A. B., Kasheffskaya, O. P., & Sinclair, J. D. (1990). Initial acceptance of alcohol: Gustatory factors and patterns of alcohol drinking. Alcohol, 7(2), 83 – 85. Krahn, D. D., & Gosnell, B. A. (1991). Fat-preferring rats consume more alcohol than carbohydrate-preferring rats. Alcohol, 8(4), 313 – 316. Krahn, D. D., Kurth, C., Demitrack, M., & Drewnowski, A. (1992). The relationship of dieting severity and bulimic behaviors to alcohol and other drug use in young women. Journal of Substance Abuse, 4(4), 341 – 353. Kranzler, H. R., Sandstrom, K. A., & Van Kirk, J. (2001). Sweet taste preference as a risk factor for alcohol dependence. American Journal of Psychiatry, 158(5), 813 – 815. Lester, D., & Greenbert, L. A. (1952). Nutrition and the etiology of alcoholism: The effect of sucrose, saccharin and fat on the self-selection of ethyl alcohol by rats. Q J Study Alcohol, 13, 533 – 560. Marks-Kaufman, R., & Lipeles, B. J. (1982). Patterns of nutrient selection in rats orally self-administering morphine. Nutrition and Behavior, 1, 33 – 46. Meisch, R. (1977). Ethanol self-administration: Infrahuman studies. In T. Thompson, & P. B. Dews (Eds.), In advances in behavioral pharmacology (pp. 286 – 313). New York7 Academic Press.

D. Krahn et al. / Addictive Behaviors 31 (2006) 622–631

631

Nowinski, J., Baker, S., & Carroll, K. (1995). Twelve step facilitation therapy manual. In M. E. Mattson (Ed.), U.S. Department of Health and Human Services (pp. 177 – 193). Rockville MD7 National Institute on alcohol Abuse and Alcoholism. Overstreet, D. H., Kampov-Polevoy, A. B., Rezvani, A. H., Murrelle, L., Halikas, J. A., & Janowsy, D. S. (1993). Saccharin intake predicts ethanol intake in genetically heterogeneous rates as well as different rat strains. Alcohol Clinical and Experimental Research, 17(2), 366 – 369. Sheehan, D. V., & Lecrubier Y. (1994). Mini-international neuropsychiatric interview (MINI). Tampa FL, University of South Florida, Institute for Research in Psychiatry; Paris, INSERM-Hospital de la Jalpetriere. Sinclair, J. D., Le, A. D., & Kiianmaa, K. (1989). The AA and ANA rat lines, selected for differences in voluntary alcohol consumption. Experientia, 45(9), 798 – 805. Sobell, L. C., & Sobell, M. B. (1992). Timeline follow-back: A technique for assessing self-report alcohol consumption. In R. Z. Litten, & J. P. Allen (Eds.), Measuring alcohol consumption: Psychosocial and biochemical methods (pp. 286 – 313). Totowa, NJ7 Humana Press. Stewart, R. B., Russell, R. N., Lumeng, L., Li, T. K., & Murphy, J. M. (1994). Consumption of sweet, salty, sour and bitter solutions by selectively bred alcohol-preferring and alcohol-nonpreferring lines of rats. Alcohol Clinical and Experimental Research, 18, 375 – 381. Watson, D., Clark, L., & Tellegen, A. (1988). Development and validation of brief measures of positive and negative affect: The PANAS Scales. Journal of Personality and Social Psychology, 54, 1063 – 1070. Yung, L., Gordis, E., & Halt, J. (1983). Dietary choices and likelihood of abstinence among alcoholic patients in an outpatient clinic. Drug and Alcohol Dependence, 12(4), 355 – 362.