Effects of Cigarette Smoking and Family History of Alcoholism on Sweet Taste Perception and Food Cravings in Women

Abstract

Background

Despite popular beliefs that smoking affects the sensitivity and liking of sweet-tasting foods and beverages, few psychophysical studies have examined this phenomenon and none have taken into account the individual’s family history of alcoholism (FH+), a predictor of heightened sweet preferences.

Methods

A within- and between-subjects study was conducted to determine the effect of both cigarette smoking and an acute exposure to nicotine on sweet taste sensitivity and preferences in women. Two groups were studied on 2 days separated by 1 week: women who were current smokers (n = 27, 18 were FH+) and those who never smoked in their lifetime (n = 22, 9 were FH+). Current smokers smoked nicotine-containing cigarettes during 1 test session and nicotine-free cigarettes during the other. The procedures were identical during both test sessions for the group of never smokers, with the exception that they did not smoke. Two-alternative staircase methods and forced-choice tracking procedures were used to assess sucrose thresholds and preferences, respectively, during both test session. Standardized questionnaires were administered to assess food cravings as well as smoking and alcohol usage and dependence. The Family Interview for Genetic Studies was used to detect alcoholism according to the DSM III criteria for family members up to second-degree relatives.

Results

Acute exposure to nicotine did not affect sucrose detection thresholds or preferences, but smokers had significantly higher sucrose detection thresholds than never smokers. The greater the smoking dose in pack-years, the lower the sucrose sensitivity. Regardless of smoking status, women who were FH+ preferred significantly higher sucrose concentrations and craved sweets more often than women who were not.

Conclusions

Both smoking and having a family history of alcoholism had differential effects on sweet taste. Smoking was associated with decreased sweet taste sensitivity whereas having a family history of alcoholism was associated with heightened sweet preferences. These findings suggests that future research on the effects of smoking on food habits and cravings should take into account family history of alcoholism given its association with sweet liking and the increased likelihood to develop a tobacco disorder.

More than 250 million women worldwide smoke tobacco, one of the most heavily addictive drugs (Mackay and Eriksen, 2002). Not only do women suffer greater relative risks of smoking-related diseases (see Perkins, 2001 for review), but they smoke for different reasons and have more difficulty quitting than men (Scharf and Shiffman, 2004). Among the many sex differences related to smoking (see Perkins, 2001 for review), women derive more pleasure from the smell and taste of cigarettes (Perkins et al., 2001, 2002), have a stronger motivation to use nicotine to avoid or reduce negative affect (Hogle and Curtin, 2006), and are more likely to smoke to control their body weight (Levine et al., 2006; Perkins, 2001). In fact, relapse often occurs because of concerns with postcessation weight gain (Levine et al., 2006). As early as 1928, ad campaigns such as “Light a Lucky and you’ll never miss sweets that make you fat” capitalized on women’s concerns about their waistlines. Cigarettes were marketed as a fat-free alternative to diminish sweet cravings and satisfy hunger. This message was so effective that market shares of Lucky Strikes^™ increased by more than 200% (Amos and Haglund, 2000).

Research conducted during the past few decades suggested that the effectiveness of these marketing campaigns was due in part to the effects of nicotine on taste and its underlying reward systems, and not just the power of suggestion. Long-term nicotine exposure decreases (Grunberg, 1982), whereas abstinence from smoking (Hall et al., 1989; Perkins et al., 1990; Spring et al., 1991) or cessation of nicotine treatment in animals chronically exposed to nicotine (Grunberg et al., 1985) increases, consumption of palatable, particularly sweet-tasting, foods. Moreover, epidemiological data on per capita consumption patterns in the United States revealed that when cigarette consumption is high, sugar consumption is low and vice versa (Grunberg and Morse, 1984).

The reasons for the association between smoking and sweet consumption remain unknown. Perhaps smoking and smoking cessation alter cravings for sweets and foods in general, and eating or drinking something sweet and rich in carbohydrates reduces cravings for cigarettes (West et al., 1990), due to shared neurobiological and psychological mechanisms (Alsene et al., 2003). In addition, smoking may affect sensitivity to sweet taste and thereby alter preferences. However, research studies in this area remain equivocal. Whereas some studies found that smoking decreased sweet taste sensitivity (Sato et al., 2002), others found opposite (Pursell et al., 1973) or no (Peterson et al., 1968; Pomerleau et al., 1991; Redington, 1984) effects. Differences in psychophysical methodology, recency of smoking and lack of information on the subjects’ smoking history may have contributed to these mixed results.

The present study aimed to determine the effects of acute nicotine exposure as well as smoking history on sucrose taste sensitivity, preferences, food cravings, and sweet-taste analgesia; sweet taste analgesia data will be presented elsewhere. To analyze the acute effects of nicotine per se, smokers were tested 30 minutes after smoking nicotine cigarettes on 1 test day and nicotine-free cigarettes on the other. To determine whether a history of smoking affected these variables, women who had been smoking for varying numbers of years were compared to women who had never smoked. Information on alcohol use and family history of alcoholism was obtained for each subject because of the significant comorbidity between tobacco and alcohol use (Sher et al., 1996), the suggested inverse relationship between alcohol and carbohydrates consumption (Mela 1989; Gruchow et al., 1985), and the observation that individuals with family history of alcoholism have an increased sweet taste preference (Kampov-Polevoy et al., 2003).

MATERIAL AND METHODS

Subjects

Forty-nine women (26.5% Black, 36.7% White, 12.2% Asian, 24.5% Mixed or Other Races), who were between the ages of 21 and 40 years, were recruited from advertisements in local newspapers. During initial screening, women were excluded if they had diabetes, high blood pressure, arthritis, conditions associated with chronic pain, irregular menstrual cycles, or were taking any medication with the exception of oral contraceptives.

Two groups of women were formed (see Table 1). The first group consisted of 27 women who smoked 4 or more cigarettes daily for at least the past year, whereas the second group consisted of 22 women who had never smoked in their lifetime. Those who smoked had been smoking for 12 (SD = 7; range 1 to 25) years with a current mean smoking rate of 10 (SD = 5; range 4 to 30) cigarettes per day. The groups were matched for the proportion of women who were Black since previous finding from our laboratory revealed that Blacks have higher sucrose preferences (Pepino and Mennella, 2005). Eleven additional women started the study but were excluded; 9 of the 11 only completed 1 testing day (current smokers = 7, never smokers = 2), and there were procedural difficulties for the remaining 2 women. All testing procedures were approved by the Office of Regulatory Affairs at the University of Pennsylvania and each woman gave informed written consent prior to testing.

Table 1.

Subjects’ Characteristics

Variablea	Current smokers	Never smokers	p-Value
Age (years)	29.4 ± 6.5	26.5 ± 4.3	0.08
BMI (kg/m2)	28.4 ± 11.6	24.0 ± 4.2	0.10
Years of education	13.4 ± 1.8	16.1 ± 1.7	<0.01
Yearly income
<$25,000	40.7%	31.8%	0.80
$25,000 to $50,000	37.1%	40.9%
>$50,000	22.2%	27.3%
Number who tasted PROP (tasted≤180 μM PROP)	21/27 (78%)	19/22 (86%)	0.35
Number who are FH+ (first and/or second generation)b	18/24 (75.0%)	9/21 (43.0%)	0.03
Number who are alcoholics (MAST score > 4)	8/27 (29.6%)	0/22 (0%)	<0.01
Percent who consumed (in parenthesis) and, if so, their monthly consumption rates
Alcohol (no. standard drinks)	40.0 ± 43.2 (96%)	17.1 ± 13.9 (91%)	0.02
Soda (no. cans)	22.6 ± 34.8 (100%)	13.3 ± 16.2 (82%)	0.29
Coffee (no. 8 oz. cups)	55.1 ± 71.3 (85%)	15.0 ± 14.3 (59%)	0.05
Cookies	5.2 ± 5.0 (93%)	6.4 ± 6.8 (91%)	0.51
Brownies	5.1 ± 18.6 (85%)	1.8 ± 2.6 (95%)	0.41
Cereal	11.5 ± 9.2 (89%)	13.4 ± 12.4 (100%)	0.55
Cake	4.8 ± 8.3 (93%)	2.2 ± 2.8 (100%)	0.17
Chocolate	17.6 ± 22.7 (93%)	10.2 ± 9.2 (95%)	0.17
Food craving inventory
Subscales
High fats	2.4 ± 0.9	1.9 ± 0.6	0.04
Carbohydrates/starches	2.7 ± 0.7	2.1 ± 0.6	<0.01
Sweets	2.8 ± 0.7	2.6 ± 0.8	0.56
Fast food fats	2.9 ± 0.8	2.6 ± 1.0	0.27
Total score	2.6 ± 0.5	2.3 ± 0.5	0.04
Number of subjects	27	22

BMI, body mass index; FH+, family history of alcoholism; PROP, propylthiouracil; MAST, Michigan Alcohol Screening Test.

^aData presented as mean ± SD or percentage.

^bFour subjects lacked information on some of their family members and therefore were excluded from this analysis.

Procedures

Each woman was tested on 2 days separated by 6 days (±1). Because there were suggestions in the literature that sucrose thresholds (Than et al., 1994) and responses to nicotine (Terner and de Wit, 2006) vary across the menstrual cycle, the first day of study for all participants occurred during the same phase of the cycle. On both test days, subjects arrived at the Monell Center and had not eaten or drunk anything but water for at least 1 hour. Those subjects who smoked were asked to refrain from smoking for 12 hours prior to testing. Smoking abstinence was verified by measuring carbon monoxide (CO) levels at the beginning of testing using the Vitalograph (Lenexa, KS); CO levels had to be ≤13 parts per million. Blood pressure was measured prior to testing (DINAMAP^® PRO Series 100; GE Medical Systems, Jupiter, FL) and only those with normotensive status (systolic blood pressure < 140 mmHg; diastolic blood pressure < 90 mmHg) participated in the study.

A between- and within-subjects design was utilized. At the beginning of both test days, current smokers smoked, in counterbalanced order, cigarettes with nicotine on 1 testing day (Marlboro^™; 1.1 mg nicotine, 15 mg tar) and nicotine-free cigarettes on the other (Quest 3^™; <0.05 mg nicotine, 9 mg tar). The smoking session was conducted in a controlled manner to standardize puff volume and duration of inhalation (Mendelson et al., 2005). Women were asked to rate how much they craved a cigarette before and after they smoked on a visual analog scale that ranged from 0 to 100 mm; higher number reflects higher craving (Mendelson et al., 2005). Never smokers did not smoke and procedures were identical on both test days. Immediately after smoking for women who smoked, or 10 minutes after arrival for never smokers, we determined the analgesic properties of intraoral sucrose, compared to water, via the cold pressor test; these data will be presented elsewhere. Thirty minutes later, we assessed sucrose detection thresholds and sucrose preferences.

At the end of the second testing day, a forced-choice psychophysical method (see Mennella et al., 2005 for details) was used to verify the genetically determined taste sensitivity to propylthiouracil (PROP) (Kim et al., 2003), a compound, like its chemical relative phenylthiocarbamide (PTC), that tastes bitter to some individuals whereas others either cannot taste them or require high concentrations to recognize their presence. PROP thresholds were determined for two reasons. First, unlike PTC, it is odorless and safer (Lawless, 1980). Second, previous research suggested that smokers were significantly more likely to be insensitive to PROP and PTC (Enoch et al., 2001; Kaplan et al., 1964; Snedecor et al., 2006) and that PROP sensitivity may be related to sweet taste sensitivity (Chang et al., 2006). To this end, subjects tasted, but did not swallow, in ascending order, water and then three 5 ml solutions of PROP (56, 180, and 560 μM), rinsing before and after each tasting. They were classified into groups based on which was the first sample, if any, they detected a bitter taste, if at all.

Sucrose Detection Thresholds

A 2-alternative staircase method was used to assess sucrose thresholds (Wetherill and Levitt, 1965). On each trial, subjects were presented with pairs of solutions, one of which was a sucrose solution and the other of which was deionized water. There were 17 sucrose solutions that ranged from 0.000056 to 1.0 M sucrose and were prepared in quarter log units. Testing began at solution 10 (0.0032 M) for all participants. During each trial, the subject was asked to determine which of the pair contained a taste. The subject rinsed her mouth with distilled water prior to tasting each sample. The concentration of the solution present was increased after a single incorrect response and decreased after 2 correct responses in a row. A reversal was considered to have occurred at points at which the concentration sequence changed directions. The procedure was terminated when 4 reversals meeting the following 2 criteria had occurred. First, there were no more than 2 dilution steps between the 2 successive reversals. Second, the series of reversals cannot form an ascending pattern (i.e. one in which positive and negative reversals are achieved at successively higher concentrations). These additional criteria ensure a more stable measure of the threshold attained (see Pribitkin et al., 2003). The threshold concentration was then calculated as the mean of the log values of the last 4 reversals.

Sucrose Taste Preferences

Sucrose preferences were assessed by a forced-choice tracking technique (Cowart and Beauchamp, 1990). Subjects were presented with pairs of solutions that differed in sucrose (Sigma-Aldrich, Inc., St. Louis, MO) concentration (0.09, 0.18, 0.35, 0.70, and 1.05 M). For comparison, sodas such as Coca Cola^™(Atlanta, GA) are equivalent in sweetness to a 0.33 M sucrose concentration. They were first presented with a pair of samples chosen from the middle range (0.18 M vs. 0.70 M) and were asked to taste each without swallowing, and then to point to which of the pair they liked better. Each subsequent pair was then determined by the subject’s preceding preference choice. The procedure continued until the subject had either chosen a given concentration of sucrose when it was paired with both a higher and lower concentration or had chosen the highest (1.05 M) or lowest (0.09 M) sucrose solution 2 consecutive times. The entire task was then repeated with stimulus pairs presented in reverse order (i.e. the weaker stimulus in the first pair was initially presented first and the stronger one was presented first in the second series). Subjects rinsed and expectorated with water before and after tasting each sample and a 1-minute interval separated each pair of solutions. The geometric mean of the sucrose concentrations chosen during the 2 trial series provided the estimate of sucrose preferences. One current smoker did not complete the sucrose preference test and was excluded from the analysis of sucrose preferences.

Anthropometrics and Questionnaires

Subjects were weighed and measured for height (Detecto, Model 439, Physician Scale; Webb City, MO) and body mass indices (kg/m²) were computed. On Day 1 of testing and 10 minutes before assessing sucrose detection thresholds, all but 1 woman completed the Food Craving Inventory (FCI), a 28 food-item validated questionnaire designed to measure the frequency of specific and general food cravings (White et al., 2002). The FCI consist of 4 subscales, each subscale consist of 4 to 8 items within the following food types: High Fats (i.e., fried chicken, gravy, sausage, hot dogs, fried fish, corn bread, bacon, steaks); Sweets (i.e., cakes, cinnamon rolls, ice cream, cookies, chocolate, donuts, candy, brownies); Carbohydrates/Starches (i.e., sandwich bread, rice, biscuits, pasta, pancakes/waffles, rolls, cereal, baked potato), and Fast Food Fats (i.e., pizza, french fries, hamburger, chips). Subjects were asked to rate how often they experienced a craving for each of the foods over the past month using a 5-point Likert scale (1 = never, 5 = always/almost every day); higher numbers for each of the subscales reflect greater cravings for that food type. A general food craving score was calculated by summing the 4 subscales. Subjects were also queried about how often they drank coffee and soda and ate chocolate, cookies, brownies, cereals, and cake per month.

The frequency of drinking, amount of alcohol consumed on a single occasion, type of alcoholic beverage (i.e., beer, wine, liquor) and size of beverage was determined by interviewing the participants with a time-line follow-back questionnaire. From these data, we estimated the number of standard drinks of alcohol consumed per month. In addition, each subject completed the Michigan Alcohol Screening Test (MAST) to screen for lifetime alcohol-related problems and alcoholism; a cutoff of MAST score >4 was used (Selzer, 1971).

The Family Interview for Genetic Studies was conducted to detect alcoholism according to the DSM III criteria for family members up to second-degree relatives (McLeod et al., 1994; Nurnberger et al., 1994) for all but 4 subjects because they lacked information on some of their family members. Based on these criteria, women who had a first and/or second-degree relative with a history of alcoholism are hereafter referred to as FH+, whereas those who had no history of alcoholism in 2 generations are hereafter referred to as FH−. Both first- and second-degree relatives were included since previous studies suggested that the presence or absence of a first-degree relative may be too narrow (McLeod et al., 1994). The incidence of women with only a first-degree relative with a history of alcoholism in this sample was 29% for current smokers and 14% for never smokers. When considering 2 generations, the incidence for FH+ increased to 75% for current smokers and 43% percent for never smokers.

Women who smoked were also queried about their smoking history (e.g., at what age they started smoking regularly, the current number of cigarettes smoked, the number of cigarettes smoked when smoking the heaviest). From these data, we calculated the cigarette dose in pack-years for each woman by multiplying the average number of packs smoked per day by the number of years of smoking (Frye et al., 1990). The modified Fagerstrom Tolerance Questionnaire (mFTQ) was administered to each current smoker to assess high nicotine dependence (Heatherton et al., 1991). Scores could range from 0 to 15; scores above 7 indicate high nicotine dependence whereas scores below 7 indicate low or moderate nicotine dependence.

Data Analyses

To determine the acute effects of nicotine exposure on sucrose sensitivity, sucrose preferences, and craving for cigarettes in current smokers, we conducted Mixed Analyses of Variance (ANOVA) with experimental condition (nicotine vs. nicotine-free cigarettes) as a within-subjects factor and order of testing (nicotine cigarette first vs. nicotine-free cigarette first) as the between-subjects factor.

To determine whether a history of smoking affected sucrose preference and sensitivity, separate ANOVA’s were conducted with group (current smoker, never smokers) as the between-subjects factors. Because thresholds and preferences for sucrose on Day 1 were not significantly different from those obtained on Day 2 for current smokers and never smokers, we calculated an average of each of these measures for each subject for these analyses. To determine whether a history of smoking affected food cravings, scores for each of the 4 FCI subscales and total FCI scores were compared for current smokers and never smokers using t-tests for independent samples (White et al., 2002). To further explore whether there was a relationship between having a family history of alcoholism on sucrose sensitivity, preferences and craving for sweets, two-way ANOVAs were conducted with group (current smokers, never smokers) and family history of alcoholism status (FH+, FH−) as the between-subjects factor. When significant, post hoc Fisher Least Significant Difference analyses were conducted.

Pearson correlation coefficients were used to assess relationships between sucrose sensitivity, sucrose preferences, FCI subscales, age, and smoking history (i.e., pack-years). Stepwise-forward regression analyses were also conducted to control for the effect of age when smoking history was examined as a potential predictor of sucrose sensitivity, preferences or food cravings. For this analysis, each variable was first ranked by its F-value. The variable with the highest significant F-value entered into the logistic regression model. To avoid the loss of nearly significant variables, p < 0.10 was considered as significant by this stepwise analysis.

RESULTS

Subject Characteristics

As shown in Table 1, there were no significant differences between current smokers and never smokers in their age, body mass index, income level, and PROP status. However, women who smoked were less educated [t(df = 47) = 5.37; p < 0.0001], and reported drinking significantly more alcohol [t(df = 47) = 2.39; p = 0.02] and coffee [t(df = 47) = 2.57; p = 0.01] when compared to women who never smoked. Current smokers reported first experiencing tasting with a cigarette when they were 14.9 years (SD = 3.6) of age (range: 3 to 21 years) and started smoking regularly when they were 17.4 years (SD = 2.6) old (range: 13 to 23 years). Seventy-five percent of the current smokers started smoking between 15 and 20 years of age. At the time of heaviest smoking, subjects smoked 20 cigarettes daily (SD = 10) (range: 6 to 40). Nicotine dependency, as measured by the mFTQ, was 7.9 (SD = 2.2) and 74% were categorized as being highly nicotine dependent (mFTQ > 6).

Effects of Acute Nicotine Exposure in Current Smokers

There was no significant effect of the experimental condition, order of testing, or interaction between these factors on sucrose detection thresholds or sucrose preferences in women who smoked (all p-values > 0.10). Women craved cigarettes more before smoking (M = 58; SD = 31 mm) when compared to after smoking [M = 39; SD = 42 mm; F(1,25) = 9.25; p < 0.01]. Although smoking either nicotine or nicotine-free cigarettes significantly decreased cravings to smoke, nicotine cigarettes were more effective [F(1,25) = 4.96; p = 0.04]. The higher the nicotine dependence (as scored on the mFTQ), the higher the cravings for a cigarette before smoking [r(df = 26) = 0.55; p = 0.003].

Effects of a History of Smoking: Current Smokers Versus Never Smokers

Figure 1 shows that a history of smoking was related to sucrose detection thresholds. Current smokers had significantly higher sucrose thresholds (i.e., decreased sensitivity) than never smokers [F(1,47) = 5.53; p = 0.02]. We point out that this analysis does not permit one to unequivocally conclude that a history of smoking is related to sucrose sensitivity as each smokers smoked a half an hour before we conducted the threshold tests whereas never smokers did not. However, this inference takes on more strength with the observation of a relationship between cigarette dose in pack-years and sucrose sensitivity: the greater the cigarette dose in pack-years, the higher the sucrose detection threshold [r(df = 27) = 0.46; p < 0.02, Fig. 2].

Fig. 1.

Sucrose detection thresholds (mM) of 27 female current smokers and 22 female never smokers during first (white bar) and second (black bar) test session. Fourteen of the current smokers smoked nicotine-free cigarettes thirty minutes prior to measuring detection thresholds during the Session 1 and nicotine cigarettes during Session 2. The order was reversed for the remaining 13 smokers. The group of never smokers did not smoke and procedures were identical during Sessions 1 and 2. Data are presented as means (±SEM). *Signifies current smokers significantly different from never smokers at p < 0.05 level.

Fig. 2.

Correlation between pack-years and sucrose detection thresholds [r(df = 27)=0.46; p < 0.02].

The correlation between age and dose pack-years was highly significant [r(df = 27) = 0.71; p < 0.001]. The older the smoker, the more years she has been smoking and therefore the higher the dose in pack-years. In addition, age was significantly correlated with sucrose thresholds for both groups [never smokers: r(df = 22) = 0.52; p < 0.02 and current smokers: r(df = 27) = 0.40; p < 0.04] such that the older the women, the higher the sucrose thresholds (i.e. the less sensitive). Nevertheless, stepwise-forward regression analysis including the age at what women started smoking regularly, cigarette dose in pack-years and current age revealed that dose pack-years was the variable that best predicted sucrose thresholds in current smokers [r = 0.46, F(1,25) = 6.8 p < 0.02].

There were no statistical differences between the groups in their bliss point for sucrose [F(1,46) = 1.46; p = 0.75]. Neither was there an association between the cigarette dose in pack-years and sucrose preference (p = 0.59). As a group, women who never smoked preferred a 0.45 ± 0.30 M (mean ± SD) sucrose solution whereas those who currently smoked most preferred a 0.48 ± 0.30 M solution. The level of sucrose preferred was not related to sucrose thresholds for either group or both groups combined (all p-values > 0.74).

Differences emerged, however, when we analyzed the effects of smoking history on the cravings for food during the past month (Table 1). Although the 2 groups of women were similar in the frequency in which they reported eating chocolate, cookies, brownies, cereals, and cake or drinking soda, current smokers craved carbohydrates [t(47df) = 2.79; p = 0.008], high fats foods [t(47df) = 2.16; p < 0.04] and foods in general (p < 0.05) more frequently when compared to never smokers. Cravings for sweet foods was positively correlated with sucrose preferences in the group of women who never smoked [r(df = 22) = 0.48; p = 0.02], but not for those who were current smokers (p = 0.40). In contrast, food cravings were related with cravings for cigarettes in smokers such that the greater the craving for a cigarette, the more frequent the craving for foods in general [r(df = 26) = 0.43; p = 0.03] and carbohydrates and high fats food, in particular [r(df = 26) = 0.43 and 0.47; both p-values < 0.03].

Family History of Alcoholism and Alcoholism

As shown in Table 1, women who smoked were more likely to be alcoholics (p < 0.006, Fisher Exact Test) and have a FH+ [chi-square (df = 1) = 4.14; p = 0.04]. However, the two-way ANOVAs revealed no significant interactions between smoking status and family history of alcoholism on sucrose thresholds, sucrose preferences or food cravings. Neither were there differences in sucrose thresholds, preferences or food cravings between current smokers who were alcoholics and current smokers who were not. Although having a family history of alcoholism did not affect sucrose detection thresholds, women who were FH+ preferred significantly higher sucrose concentrations [F(1, 42) = 3.99; p = 0.05] and craved sweets more often [F(1,43) = 3.9; p = 0.05; see Fig. 3] than FH− women. These differences in sweet preferences were not due to ethnic/racial differences among the groups: 30% of the FH+ and 22% of the FH− were Black (p = 0.42).

Fig. 3.

The sucrose preference (geometric mean) and frequency of craving sweets during the past month, as determined by the FCI, in women who have a family history of alcoholism (solid bars; n = 27) and those who do not (hatched bars; n = 17). Data are presented as means (±SEM). *Signifies significantly different at p = 0.05 level. Five women were excluded from this analysis; 4 lacked information on some of their family members and 1 did not complete the sucrose preference test.

DISCUSSION

Acute nicotine exposure had no effects on sucrose detection thresholds or preferences in women who smoked. However, women who currently smoked had higher sucrose detection thresholds, albeit within normal clinical range (Pribitkin et al., 2003), when compared to never smokers who were matched for age, race, body mass index, income, and phase of the menstrual cycle. The greater the cumulative smoking dose pack-years, the greater the decline in sucrose taste sensitivity. That there were significant relationships between age and sucrose thresholds among both groups of women, and age and pack-years among the smokers, warrants additional study. Given the small sample size of the current study, we can not address this important issue of determining the unique role of each of these independent variables on taste perception among smokers. However, the association between pack-years and declines in sweet taste sensitivity garners more support from larger clinical studies on the effects of smoking on olfactory functioning and bitter taste sensitivity. In a study of 638 adults of varying ages, smokers were twice as likely to evidence an olfactory deficit that persons who never smoked after adjusting for potential confounders such as age and sex (Frye et al., 1990). Two smaller-scaled studies evaluated the effects of smoking on bitter taste. One such study, which did not control for age, found that smokers were less sensitive to the taste of quinine, a bitter-tasting substance (Kaplan et al., 1964). That bitter taste sensitivity improves after smoking cessation is suggested from the other within-and between-subject design study which determined taste threshold in smokers at various time before and during enrollment in a group therapy program to stop smoking (Peterson et al., 1968). Taste thresholds for bitter improved in those individuals who had successfully quit smoking for 18 months but was unaltered in those who continued smoking. Whether smoking leads to long-term, but reversible (see Frye et al., 1990), sensory deficits for salty, umami, and sour tastes and/or trigeminal irritation, another component of flavor, is an important area for future research.

The present study focused on sweet taste because of the research relating nicotine administration/cigarette smoking with changes in consumption patterns of sweet-tasting, high caloric foods (Grunberg, 1982; Grunberg and Morse, 1984; Grunberg et al., 1985; Hall et al., 1989). Whether the decreased sucrose detection thresholds in current smokers would affect perception of sweetness when sucrose levels are above threshold (i.e., suprathresholds) or whether this decreased sensitivity can be partially reversed following long periods of smoking cessation, like that recovery observed for smell identification (Frye et al., 1990) and bitter taste sensitivity (Peterson et al., 1968) remain unknown.

Several hypotheses, not mutually exclusive, can be suggested to explain how smoking affected sucrose thresholds. First, taste perception may be altered because of poor oral hygiene (Ohno et al., 2003) or damage to the periphery (Taybos, 2003). Tobacco use can affect the surface epithelium of the tongue, irritate salivary glands on the hard palate, and increase risk for periodontal disease (Taybos, 2003). Although we did not assess oral care in these subjects, poor oral hygienic regimens, which are common among smokers (Andrews et al., 1998) as well as the elderly (Bartoshuk et al., 1986), may produce a background taste that causes poorer discrimination between water and the tastant, consequently resulting in poorer taste detection (see Bartoshuk et al., 1986). Perhaps good oral care would improve taste sensitivity in current smokers in a manner similar to that observed in the elderly (Ohno et al., 2003).

Second, changes in taste detection may be due to the ability of nicotine to inhibit neurons in the nucleus of the solitary tract (NTS), the first central relay in the gustatory pathway that plays a key role in the modulation of taste signals (Simons et al., 2006). Nicotine applied directly on the tongue depressed taste-evoked responses of gustatory neurons of the NTS in rats (Simons et al., 2006). The depressant effects were dose-dependent and reliant on trigeminal afferents (Simons et al., 2006). Although the present study did not find alterations in sucrose detection thresholds after acute nicotine exposure, the possibility that prolonged periods of smoking affects the circuit of gustatory transmission in NTS cannot be ruled out.

Third, the differences in taste detection thresholds were not due to smoking per se but may be a characteristic of individuals who are susceptible to become smokers. For example, it has been suggested that sensitivity to bitter compounds such as PTC and PROP may protect against cigarette smoking (Enoch et al., 2001; Kaplan et al., 1964; Snedecor et al., 2006). However, we found that the women who smoked were as likely to be sensitive to PROP as were women who never smoked, a finding which is consistent with a recent study in which 567 individuals were genotyped for polymorphisms that affect taste sensitivity to PROP and PTC (Cannon et al., 2005). That certain haplotypes of the TAS2R38 gene, and not PROP/PTC sensitivity in general, revealed a protective factor for smoking (Cannon et al., 2005) calls for more research to determine the role of other genes that affect sensitivity to tastes and other flavors present in cigarette smoke (Cannon et al., 2005). Importantly, the genetic contributions to smoking may vary within the different stages of dependency and what is protective for smoking initiation may have no effect on smoking maintenance or cessation (Kendler et al., 1999). To be sure, tobacco dependence, like other drug dependencies, is a complex trait that is affected both by environment and multiple genes (Batra et al., 2003).

Fourth, smoking may alter sucrose as well as bitter taste detection thresholds (Kaplan et al., 1964; Peterson et al., 1968) because of nicotine-induced alteration in serotonin, a neurotransmitter that modulates cellular responses of taste receptors (Herness et al., 2005). Not only has a history of smoking been shown to reduce serotonin levels centrally (Benwell et al., 1990), but the acute administrations of Paroxetine, a specific serotonin reuptake inhibitor, in healthy individuals decreased sweet and bitter taste thresholds (Heath et al., 2006). Moreover, individuals with seasonal affective disorder (SAD), a disorder postulated to be due to a decreased central serotoninergic function (Rosenthal et al., 1987), had higher taste thresholds for sweet, bitter, and salty tastes during the winter, when they are depressed, but thresholds normalized during the summer, when the symptoms of depression disappeared (Arbisi et al., 1996).

Although women in the present study craved carbohydrates more often if they were current smokers, there were no differences in consumption patterns for foods such as cake, cookies, brownies, and cereal. This is consistent with previous research that shows that carbohydrate intake is not different between current smokers and never smokers (Dallongeville et al., 1998). The more they craved cigarettes, the more they craved certain types of foods. Perhaps it is only when these women are deprived of nicotine or seeking relief from nicotine withdrawal that they give in to these cravings (Spring et al., 1991). Like nicotine (Ribeiro et al., 1993), diets rich in carbohydrates, but poor in protein, increase brain serotonin levels (Fernstrom and Wurtman, 1971). Alternatively, current smokers may confuse hunger with nicotine cravings during periods of nicotine withdrawal (West, 2001). That there is a crossover between cravings for cigarettes and food (Alsene et al., 2003) is suggested by the findings that smoking abstinence increased food consumption (Hall et al., 1989; Perkins et al., 1990) and food restriction increases cigarette use in both men and women (Cheskin et al., 2005).

Women who smoked consumed significantly more alcohol and coffee than never smokers, findings which are consistent with previously published research in the US (Perkins, 1992; Perkins et al., 1990) and China (Koh et al., 2005). They were also more likely to be alcoholics and have a history of alcoholism in their family. The high frequency of FH+ among smokers in the present study is remarkably similar to a recent report from the San Diego Prospective Study (74%; Schuckit et al., 2007) as well as a study of nonalcoholic women with generalized anxiety disorders (76%; McLeod et al., 1994).

A family history of alcoholism, rather than smoking or being an alcoholic per se (Kampov-Polevoy et al., 2003), was associated with greater liking for sweets. That is, FH+ women preferred significantly higher sucrose concentrations and craved sweets more often than FH− women. The comorbidity of alcohol and nicotine dependence and the associations of these 2 drugs of abuse with sweet taste preferences may be attributed to common neurobiological bases of reward (Levine et al., 2003), pharmacological interactions resulting in one drug facilitating the other (Acheson et al., 2006), and/or shared genetic factors (see Li et al., 2007). Rats genetically selected for high alcohol consumption are more sensitive to the rewarding properties of nicotine (Le et al., 2006) than rats genetically selected for low alcohol consumption. Likewise, consumption of sweet-tasting solutions predicts alcohol intake among rodent strains and within their segregating crosses (Bachmanov et al., 1996) and alcohol-preferring animals find sucrose more rewarding than animals that show low preference to alcohol (Le et al., 2006). Clearly, future research on the effects of smoking or smoking abstinence on sweet taste preferences and food habits should take into account family history of alcoholism given its associations with sweet liking and increased likelihood to develop tobacco dependence.

The results of the present study extend previous findings that smoking is associated with altered chemosensory perception (Frye et al., 1990; Kaplan et al., 1964; Peterson et al., 1968) and suggest that smoking and having a family history of alcoholism have differential effects on sweet taste. Whether the associations are a consequence or predictor of smoking and drinking remain unknown. Longitudinal studies, which track measures of taste thresholds and sensitivity in young children before they have experienced any drugs of abuse, are needed to determine whether sweet taste thresholds and preferences could serve a marker for those more vulnerable to develop addictions. More knowledge of how genetic and experiential factors interact with smoking and alcohol addictions on the perception of food-related cues may assist in designing programs to help women, a population especially concerned with weight gain (Perkins, 2001), to stop smoking.