## Abstract

Introduction:

Cross-sectional studies consistently find that cigarette smokers consume fewer fruits and vegetables each day than do nonsmokers. However, there are no published cohort studies on this relationship. This study evaluated the longitudinal relationship between fruit and vegetable consumption (FVC) and cigarette smoking, including measures of dependence and abstinence in a national population-based cohort analysis.

Methods:

A national random-digit-dialed sample of 1,000 smokers (aged 25 years and older) assessed baseline FVC and indicators of general health orientation. Multivariable analyses were used to assess whether baseline FVC was associated with smoking intensity, time to first cigarette (TTFC), and total score on an abbreviated version of the Nicotine Dependence Syndrome Scale (NDSS), adjusting for age, gender, race/ethnicity, education, and household income. The study also assessed whether baseline FVC predicted 30-day abstinence from all tobacco products at 14-month follow-up among baseline cigarette smokers, with additional adjustment for indicators of general health orientation (heavy drinking, exercise, and illicit drug use).

Results:

Higher FVC was associated with fewer cigarettes smoked per day, longer TTFC, and lower NDSS score. Those in the highest quartile of FVC were 3.05 times more likely (p < .01) than those in the lowest quartile to be abstinent for at least 30 days at follow-up.

Conclusions:

FVC was inversely associated with indicators of nicotine dependence and predicted abstinence at follow-up among baseline cigarette smokers. Further observational studies and experimental research would provide useful information on the consistency of the relationship and help elucidate possible mechanisms.

## Introduction

Cross-sectional studies have consistently observed relationships between various components of the human diet and cigarette smoking. For example, compared with those who do not smoke, people who do tend to consume more high-fat meats, caffeinated beverages, and alcohol, and fewer fruits and vegetables (F&Vs), high fiber grains, dairy, and lean meats (e.g., fish, poultry; Abood & Conway, 1994; Morabia & Wynder, 1990; Nuttens et al., 1992; Palaniappan, Starkey, O’Loughlin, & Gray-Donald, 2001; Serdula et al., 1996; Subar, Harlan, & Mattson, 1990; Whichelow, Golding, & Treasure, 1988). Nuttens et al. (1992) further found that after controlling for alcohol intake, the association of meat and bread consumption with smoking disappeared, while differences in fruit, vegetable, and dairy consumption persisted. Multiple researchers have observed an inverse relationship between fruit and vegetable consumption (FVC) and cigarette consumption; smokers who eat more F&Vs smoke fewer cigarettes per day (CPD; Birkett, 1999; Morabia & Wynder, 1990; Nuttens et al., 1992; Subar et al., 1990; Whichelow et al., 1988; Zondervan, Ocke, Smit, & Seidell, 1996).

Through meta-analysis of 51 published studies from 15 countries on the nutrient intake differences between smokers and nonsmokers, Dallongeville et al. (1998) observed that smokers, compared with nonsmokers, consumed more saturated fat, cholesterol, and alcohol, and less polyunsaturated fat, fiber, vitamin C, vitamin E, and beta-carotene. Lower intake of vitamin C, beta-carotene (Bolton-Smith, Casey, Gey, Smith, & Tunstall-Pedoe, 1991), and fiber (Nuttens et al., 1992; Woodward, Bolton-Smith, & Tunstall-Pedoe, 1994) has been attributed to lower FVC. Given the numerous other micronutrients and phytochemicals found in F&Vs that are critical for human physiological needs, differences in FVC between smokers and nonsmokers likely account for other differences in their nutritional intake as well.

The existing literature on the relationship between diet or specific FVC and smoking is limited to cross-sectional studies, which cannot address the direction of the relationship. Several published studies on diet and smoking have been conducted largely to point out the need to include dietary factors, especially FVC, in epidemiological studies of the etiology of tobacco-related diseases (Morabia & Wynder, 1990; Nuttens et al., 1992; Palaniappan et al., 2001; Serdula et al., 1996; Subar et al., 1990). Other studies have explained differences in food intake by differences in the general health orientation of smokers and nonsmokers (Goldbourt & Medalie, 1975; Woodward et al., 1994). These studies leave the question as to the direction of the relationship untested. Could nutritional factors influence the development and maintenance of nicotine addiction (Giovino, 2007) or is the relationship explained by general health orientation? In a preliminary analysis of U.S. Behavioral Risk Factor Surveillance System (BRFSS) data from 1996, 1998, and 2000, Yang and Giovino (2007) observed that former smokers abstinent less than 6 months were more likely than current smokers to consume at least five servings of F&Vs per day. This prompts the question of whether FVC increased postcessation or whether smokers who consumed large amounts of F&Vs were more likely to quit. Given the findings that smokers who consume large amounts of F&Vs smoke fewer CPD, another question is raised as to whether FVC is related to nicotine dependence.

This study extends previous work on FVC and CPD as we evaluated the relationship between FVC and indicators of nicotine dependence. In addition, we assessed via longitudinal analysis whether those smokers with higher FVC at baseline would be more likely to abstain from smoking at follow-up than those with lower levels of FVC.

## Methods

We administered a random-digit-dialed telephone survey in a U.S. national sample of 1,000 current smokers aged 25 or older at baseline. Participants were recruited from all 50 states and the District of Columbia for a study of 1,000 current and 256 former smokers. The overall response rate for the study was 45.7% (the American Association for Public Opinion Research [AAPOR] Response Rate #4; AAPOR 2011). To be considered a smoker and included in the study, respondents must have smoked at least 100 lifetime cigarettes and have been currently smoking daily or on some days. Baseline assessments were completed over an approximate 1-year period and follow-up was conducted on average 14.5 months postbaseline, with a 75.1% follow-up completion rate (n = 751). This study was approved by the Roswell Park Cancer Institute and the University at Buffalo Social and Behavioral Sciences institutional review boards.

### Measures

#### Independent Variables of Interest

Independent variables of interest were frequency of consumption of fruits and fruit juice combined, vegetables (excluding potatoes), and aggregate FVC. Measures for FVC were taken from the BRFSS (2003 Questionnaire) including the questions: “How often do you drink fruit juices, such as orange, grapefruit, or tomato?” “Not counting juice, how often do you eat fruit; how often do you eat green salad?” “How often do you eat carrots?” and “Not counting carrots, potatoes, or salad, how many servings of vegetables do you usually eat (by serving, we mean a side portion or helping on your plate)?” Response categories were in times per day, per week, per month, or per year, and individual responses were converted into amounts consumed on a weekly basis with quartiles constructed for analysis. Using categories such as quartiles to quantify the independent variable is a standard method in nutritional epidemiology (Willett, 1998). The BRFSS F&V questions have been assessed at moderate reliability and validity (Nelson, Holtzman, Bolen, Stanwyck, & Mack, 2001).

#### Dependent Variables

Dependent variables included three indicators of dependence (CPD, time to first cigarette [TTFC], and the Nicotine Dependence Syndrome Scale [NDSS]) and, for longitudinal analyses, a measure of abstinence from cigarettes and not using other tobacco products within the past 30 days. CPD was categorized as less than 20 or 20 or more CPD. TTFC of the day after waking was categorized as 30 min or less from waking or greater than 30 min (Baker et al., 2007). We used the four items from the NDSS (Shiffman, Waters, & Hickcox, 2004) that were used in the Tobacco Use Supplement to the Current Population Survey 2003 (National Cancer Institute, 2003) to rate nicotine dependence as measured on a scale from 4 to 12. We considered a score of nine or greater to represent a higher level of dependence and less than nine a lower level of dependence. Question statements had response categories of (a) would you say not at all true, (b) somewhat true, or (c) very true and included: (i) you have trouble going for more than a few hours without smoking; (ii) even in a bad rainstorm, if you ran out of cigarettes, you would probably go to the store to get some more; (iii) when you go without smoking for a few hours, you experience craving; (iv) if you were in a public place where smoking is not allowed, you would probably go outside to smoke a cigarette, even in cold or rainy weather.

Demographic variables that were adjusted for in statistical models included age (continuous variable), gender (male, female), race-ethnicity (non-Hispanic White, Hispanic, black, other, or unknown), education (<12 years, 12 years, 13–15 years, ≥16 years), and household income (<$25,000;$25,000–$49,999;$50,000–$74,999;$75,000–$99,999; ≥$100,000; unknown). Alcohol use was measured and categorized as no use, use within the last 30 days without heavy drinking, and heavy drinking in the last 30 days, as categorized by five or more drinks consecutively or within a few hours of each other. Illicit drug use was measured as use within the last 30 days of marijuana, cocaine, ecstasy, or LSD. Exercise was measured as average hours per week (0, 1–3, 4–7, and ≥8 hr) with the question “How many hours per week on average do you exercise? This includes playing sports, working out, aerobics, running, swimming, brisk walking, and other exercise activities.” We had also tested self-efficacy of quitting, motivation to quit, and previous quit attempts for inclusion in the model, reasoning that they may influence the relationship with FVC and smoking. However, they were not correlated with FVC (p values > .05), had very weak correlations with each other (−.10 to .18), and thus were not included.

### Analysis

A generalized linear model was used to estimate risk ratios for measures of nicotine dependence and abstinence against fruit (plus juice), vegetables, and combined FVC. Adjustment variables included age, gender, race-ethnicity, education, household income, and indicators of general health orientation (i.e., heavy drinking, illicit drug use, and exercise). For descriptive purposes, adjustment variables were left in their original categories as listed in Table 1, but for analysis, all adjustment variables were coded as dichotomous to maintain a higher degree of freedom and model parsimony, with the exception of age (retained as continuous). Cutoffs for dichotomy were made at the closest categorical median prevalence which also fell at theoretically reasonable points; race-ethnicity was categorized as non-Hispanic White and minority, education at 12 years or less and greater than 12 years, household income at $49,999 or less and$50,000 or greater, alcohol as no heavy drinking or no alcohol use and heavy drinking with in the past 30 days, and exercise as 3 hr or fewer per week and more than 3 hr per week. Models were run using both listwise deletion, and multiple imputation given that 6.4% of the data for income were missing. However, there was not a significant difference in the results, so the results from the more straightforward model using listwise deletion are reported in this article. Data were analyzed using IBM SPSS Statistics version 19.

Table 1.

Sample Characteristics

 Baseline Baseline with follow-up Baseline without follow-up δ1 n = 1,000 n = 751 n = 249 Smoking intensity (CPD mean) 17.3 17.4 17.0 0.4 TTFC (% ≤ 30 min) 57.9 58.3 56.6 1.7 NDSS (%) Low dependence 27.4 27.5 27.3 0.2 Mild dependence 29.9 30.6 27.8 2.8 Medium dependence 21.3 21.0 22.0 −1 High dependence 21.4 20.9 22.9 −2 Age in years (mean) 44.7 45.6 42.0 3.6*** Gender (% male) 43.8 41.1 51.8 10.7** Race/ethnicity (%)*** White, non-Hispanic 74.5 77.5 65.5 12 Hispanic 6.3 4.5 11.6 7.1 Black 13.5 13.7 12.9 0.8 Other or unknown 5.7 4.3 10.0 5.7 Education (%)*** <12 years 13.7 11.3 20.9 9.6 12 years or GED 35.0 34.8 35.7 0.9 13–15 years 34.2 36.0 28.9 7.1 ≥16 years 17.1 18.0 14.5 3.5 Household income (%)*** <$25,000 26.6 24.0 34.2 10.2$25,000–$49,999 35.3 34.3 38.0 3.7$50,000–$74,999 21.2 22.0 18.6 3.4$75,000–$99,000 8.7 9.3 6.8 2.5 ≥$100,000 8.3 10.3 2.5 7.8 Unknown 6.4 6.9 4.8 2.1 Alcohol use within 30 days No alcohol 39.4 39.7 38.4 1.3 Alcohol without heavy drinking 33.0 34.8 27.3 7.5 Alcohol with heavy drinking 27.7 25.5 34.3 8.8 Used street drug within 30 days (% yes) 9.2 7.7 13.8 6.1** Exercise (hours per week) 0 26.4 26.6 26.0 0.6 1–3 24.3 25.5 20.7 4.8 4–7 24.2 23.9 25.2 1.3 8 or more 25.0 24.0 28.0 4
 Baseline Baseline with follow-up Baseline without follow-up δ1 n = 1,000 n = 751 n = 249 Smoking intensity (CPD mean) 17.3 17.4 17.0 0.4 TTFC (% ≤ 30 min) 57.9 58.3 56.6 1.7 NDSS (%) Low dependence 27.4 27.5 27.3 0.2 Mild dependence 29.9 30.6 27.8 2.8 Medium dependence 21.3 21.0 22.0 −1 High dependence 21.4 20.9 22.9 −2 Age in years (mean) 44.7 45.6 42.0 3.6*** Gender (% male) 43.8 41.1 51.8 10.7** Race/ethnicity (%)*** White, non-Hispanic 74.5 77.5 65.5 12 Hispanic 6.3 4.5 11.6 7.1 Black 13.5 13.7 12.9 0.8 Other or unknown 5.7 4.3 10.0 5.7 Education (%)*** <12 years 13.7 11.3 20.9 9.6 12 years or GED 35.0 34.8 35.7 0.9 13–15 years 34.2 36.0 28.9 7.1 ≥16 years 17.1 18.0 14.5 3.5 Household income (%)*** <$25,000 26.6 24.0 34.2 10.2$25,000–$49,999 35.3 34.3 38.0 3.7$50,000–$74,999 21.2 22.0 18.6 3.4$75,000–$99,000 8.7 9.3 6.8 2.5 ≥$100,000 8.3 10.3 2.5 7.8 Unknown 6.4 6.9 4.8 2.1 Alcohol use within 30 days No alcohol 39.4 39.7 38.4 1.3 Alcohol without heavy drinking 33.0 34.8 27.3 7.5 Alcohol with heavy drinking 27.7 25.5 34.3 8.8 Used street drug within 30 days (% yes) 9.2 7.7 13.8 6.1** Exercise (hours per week) 0 26.4 26.6 26.0 0.6 1–3 24.3 25.5 20.7 4.8 4–7 24.2 23.9 25.2 1.3 8 or more 25.0 24.0 28.0 4

Note. CPD = cigarettes per day; GED = General Educational Development certificate or diploma; NDSS = Nicotine Dependence Syndrome Scale; TTFC = time to first cigarette; δ1 = difference between baseline with follow-up and those lost to follow-up. *p < .05; ** = p < .01, *** = p < .001 for difference between retained and those lost to follow-up for variable overall.

## Results

The population at baseline ranged in age from 25 to 105 years (mean: 44.7; SD: 12.9), was 43.8% male, had a mean education level of category 13–15 years, and a mean household income in the category $50,000–$74,999 (Table 1). Past 30-day alcohol use was distributed as 39.4% not consuming alcohol, 33% consuming alcohol without heavy drinking, and 27.7% drinking heavily. Exercise levels were distributed relatively equally across categories of 0, 1–3, 4–7, and 8 hr or more per week. Participants lost to follow-up were more likely than those who were retained to be younger male minority, to have completed fewer years of formal education, to have lower household incomes, and to use illicit drugs. When the alcohol variable was dichotomized (i.e., heavy drinking vs. no heavy drinking), persons with a history of heavy drinking at baseline were less likely to have been retained. There was no statistically significant difference overall between those retained and those lost to follow-up for measures of nicotine dependence or exercise.

### Measures of Dependence

Those in the highest three quartiles of fruit consumption versus the lowest quartile were less likely to smoke ≥ 20 cigarettes per day (p < .001), to smoke within 30 min of waking (p < .05), or have an NDSS score of at least nine (p < .01) at baseline (Table 2). Those in the highest quartile of vegetable consumption were less likely to exhibit all three indicators of dependence on cigarettes than were those in the lowest quartile (p < .01 for CPD and p < .05 for TTFC and NDSS). For combined FVC, those in the highest two quartiles of consumption were less likely to exhibit all three indicators of dependence (p < .05 for TTFC and p < .01 for CPD and NDSS).

Table 2.

Indicators of Dependence by Fruit and Vegetable Consumption at Baseline

 Total ≥20 CPD TTFC ≤ 30 min from waking NDSS ≥ 9 Total weekly n % RR 95% CI % RR 95% CI % RR 95% CI Fruit consumption Quartile 1 (0.00–2.10 TPW) 255 67.3 1.00 – 66.0 1.00 – 65.0 1.00 – Quartile 2 (2.11–6.90 TPW) 240 50.2 0.75*** 0.64, 0.87 56.9 0.86* 0.75, 1.00 51.3 0.79** 0.68, 0.92 Quartile 3 (6.91–12.00 TPW) 269 43.3 0.64*** 0.54, 0.76 52.4 0.80** 0.69, 0.92 39.8 0.61*** 0.51, 0.73 Quartile 4 (≥12.01 TPW) 224 42.2 0.63*** 0.52, 0.75 53.3 0.81** 0.69, 0.94 50.0 0.77** 0.65, 0.91 Vegetable consumption Quartile 1 (0.00–8.21 TPW) 240 56.4 1.00 – 60.6 1.00 – 55.6 1.00 – Quartile 2 (8.22–12.44 TPW) 237 59.2 1.05 0.90, 1.23 64.0 1.06 0.91, 1.22 52.5 0.95 0.80, 1.12 Quartile 3 (12.45–18.60 TPW) 241 48.8 0.87 0.72, 1.03 55.0 0.91 0.77, 1.07 52.5 0.95 0.80, 1.12 Quartile 4 (≥18.61 TPW) 254 40.6 0.72** 0.59, 0.87 50.4 0.83* 0.71, 0.98 45.2 0.81* 0.68, 0.98 Fruit and vegetable consumption Quartile 1 (0.00–13.00 TPW) 249 61.6 1.00 – 64.7 1.00 – 60.6 1.00 – Quartile 2 (13.01–20.04 TPW) 242 56.4 0.92 0.79, 1.07 59.7 0.92 0.80, 1.06 53.5 0.88 0.75, 1.04 Quartile 3 (20.05–29.61 TPW) 242 45.4 0.74** 0.62, 0.88 53.7 0.83* 0.71, 0.97 45.9 0.76** 0.63, 0.90 Quartile 4 (≥29.62 TPW) 234 40.0 0.65*** 0.54, 0.78 50.4 0.78** 0.67, 0.92 44.6 0.74** 0.62, 0.88
 Total ≥20 CPD TTFC ≤ 30 min from waking NDSS ≥ 9 Total weekly n % RR 95% CI % RR 95% CI % RR 95% CI Fruit consumption Quartile 1 (0.00–2.10 TPW) 255 67.3 1.00 – 66.0 1.00 – 65.0 1.00 – Quartile 2 (2.11–6.90 TPW) 240 50.2 0.75*** 0.64, 0.87 56.9 0.86* 0.75, 1.00 51.3 0.79** 0.68, 0.92 Quartile 3 (6.91–12.00 TPW) 269 43.3 0.64*** 0.54, 0.76 52.4 0.80** 0.69, 0.92 39.8 0.61*** 0.51, 0.73 Quartile 4 (≥12.01 TPW) 224 42.2 0.63*** 0.52, 0.75 53.3 0.81** 0.69, 0.94 50.0 0.77** 0.65, 0.91 Vegetable consumption Quartile 1 (0.00–8.21 TPW) 240 56.4 1.00 – 60.6 1.00 – 55.6 1.00 – Quartile 2 (8.22–12.44 TPW) 237 59.2 1.05 0.90, 1.23 64.0 1.06 0.91, 1.22 52.5 0.95 0.80, 1.12 Quartile 3 (12.45–18.60 TPW) 241 48.8 0.87 0.72, 1.03 55.0 0.91 0.77, 1.07 52.5 0.95 0.80, 1.12 Quartile 4 (≥18.61 TPW) 254 40.6 0.72** 0.59, 0.87 50.4 0.83* 0.71, 0.98 45.2 0.81* 0.68, 0.98 Fruit and vegetable consumption Quartile 1 (0.00–13.00 TPW) 249 61.6 1.00 – 64.7 1.00 – 60.6 1.00 – Quartile 2 (13.01–20.04 TPW) 242 56.4 0.92 0.79, 1.07 59.7 0.92 0.80, 1.06 53.5 0.88 0.75, 1.04 Quartile 3 (20.05–29.61 TPW) 242 45.4 0.74** 0.62, 0.88 53.7 0.83* 0.71, 0.97 45.9 0.76** 0.63, 0.90 Quartile 4 (≥29.62 TPW) 234 40.0 0.65*** 0.54, 0.78 50.4 0.78** 0.67, 0.92 44.6 0.74** 0.62, 0.88

Note. CPD = cigarettes per day; NDSS = NDSS = Nicotine Dependence Syndrome Scale; RR = risk ratios; TPW= times per week; TTFC = time to first cigarette. Estimated RR adjusted for age, gender, race, education, income. *p < .05; ** = p < .01, *** = p < .001.

### Measure of Abstinence

In longitudinal analyses that adjusted for demographics and indicators of health orientation, the likelihood of making a quit attempt was higher in quartile 4 of vegetable consumption and in quartiles 2 and 4 of combined FVC versus quartile 1 (Table 3). There was no significant difference in making a quit attempt across quartiles of fruit consumption.

Table 3.

Abstinence From Cigarettes and Not Using Other Tobacco Products at Follow-Up by Fruit and Vegetable Consumption at Baseline

 Total Made a quit attempt Abstinence ≥ 30 days among attempters Overall abstinence ≥ 30 days Baseline total weekly n % RR 95% CI % RR 95% CI % RR 95% CI Fruit consumption Quartile 1 (0.00–2.10 TPW) 183 48.6 1.00 – 8.3 1.00 – 4.0 1.00 – Quartile 2 (2.11–6.90 TPW) 179 45.1 0.93 0.74, 1.17 23.0 2.76* 1.21, 6.28 10.4 2.58* 1.10, 6.05 Quartile 3 (6.91–12.00 TPW) 210 47.8 0.99 0.79, 1.22 14.4 1.73 0.73, 4.14 6.5 1.62 0.65, 4.02 Quartile 4 (≥12.01 TPW) 170 58.1 1.20 0.97, 1.47 25.3 3.03** 1.37, 6.73 14.8 3.67** 1.61, 8.35 Vegetable consumption Quartile 1 (0.00–8.21 TPW) 172 45.1 1.00 – 11.4 1.00 – 5.2 1.00 – Quartile 2 (8.22–12.44 TPW) 174 44.7 0.99 0.77, 1.27 12.5 1.09 0.45, 2.67 5.7 1.09 0.43, 2.75 Quartile 3 (12.45–18.60 TPW) 188 48.5 1.08 0.85, 1.36 20.0 1.75 0.80, 3.84 9.2 1.77 0.77, 4.06 Quartile 4 (≥18.61 TPW) 197 59.5 1.32* 1.07, 1.63 22.7 1.99 0.95, 4.16 13.4 2.59* 1.20, 5.57 Fruit and vegetable consumption Quartile 1 (0.00–13.00 TPW) 177 50.3 1.00 – 9.6 1.00 – 4.8 1.00 – Quartile 2 (13.01–20.04 TPW) 184 38.1 0.76* 0.59, 0.97 20.0 2.08 0.92, 4.71 7.7 1.61 0.68, 3.77 Quartile 3 (20.05–29.61 TPW) 184 49.4 0.98 0.79, 1.22 15.2 1.58 0.68, 3.65 7.0 1.45 0.60, 3.50 Quartile 4 (≥29.62 TPW) 183 61.5 1.22* 1.01, 1.48 23.8 2.47* 1.18, 5.19 1.47 3.05** 1.42, 6.57
 Total Made a quit attempt Abstinence ≥ 30 days among attempters Overall abstinence ≥ 30 days Baseline total weekly n % RR 95% CI % RR 95% CI % RR 95% CI Fruit consumption Quartile 1 (0.00–2.10 TPW) 183 48.6 1.00 – 8.3 1.00 – 4.0 1.00 – Quartile 2 (2.11–6.90 TPW) 179 45.1 0.93 0.74, 1.17 23.0 2.76* 1.21, 6.28 10.4 2.58* 1.10, 6.05 Quartile 3 (6.91–12.00 TPW) 210 47.8 0.99 0.79, 1.22 14.4 1.73 0.73, 4.14 6.5 1.62 0.65, 4.02 Quartile 4 (≥12.01 TPW) 170 58.1 1.20 0.97, 1.47 25.3 3.03** 1.37, 6.73 14.8 3.67** 1.61, 8.35 Vegetable consumption Quartile 1 (0.00–8.21 TPW) 172 45.1 1.00 – 11.4 1.00 – 5.2 1.00 – Quartile 2 (8.22–12.44 TPW) 174 44.7 0.99 0.77, 1.27 12.5 1.09 0.45, 2.67 5.7 1.09 0.43, 2.75 Quartile 3 (12.45–18.60 TPW) 188 48.5 1.08 0.85, 1.36 20.0 1.75 0.80, 3.84 9.2 1.77 0.77, 4.06 Quartile 4 (≥18.61 TPW) 197 59.5 1.32* 1.07, 1.63 22.7 1.99 0.95, 4.16 13.4 2.59* 1.20, 5.57 Fruit and vegetable consumption Quartile 1 (0.00–13.00 TPW) 177 50.3 1.00 – 9.6 1.00 – 4.8 1.00 – Quartile 2 (13.01–20.04 TPW) 184 38.1 0.76* 0.59, 0.97 20.0 2.08 0.92, 4.71 7.7 1.61 0.68, 3.77 Quartile 3 (20.05–29.61 TPW) 184 49.4 0.98 0.79, 1.22 15.2 1.58 0.68, 3.65 7.0 1.45 0.60, 3.50 Quartile 4 (≥29.62 TPW) 183 61.5 1.22* 1.01, 1.48 23.8 2.47* 1.18, 5.19 1.47 3.05** 1.42, 6.57

Note.RR = risk ratios; TPW= times per week. Estimated RR adjusted for age, gender, race, education, income, heavy drinking, street drug use, and exercise. *p < .05; ** = p < .01, *** = p < .001.

Multivariable analyses indicate that among persons who made a quit attempt, those in quartiles 2 and 4 of fruit consumption were more likely to be abstinent at follow-up than were those in quartile 1 (Table 3). There were no differences in success among attempters across quartiles of vegetable consumption. Those attempters in quartile 4 of combined FVC were 2.47 times (95% CI: 1.18, 5.19) more likely to report abstinence from cigarettes and other tobacco products for at least 30 days compared with those in the lowest quartile of FVC. Among all baseline smokers who were followed up, abstinence was higher in the second and fourth quartiles of fruit consumption, the fourth quartile of vegetable consumption, and the fourth quartile of combined FVC, relative to their respective referent groups. Those in the highest FVC quartile were 3.05 times (95% CI: 1.42, 6.57) more likely than those in the lowest FVC quartile to report abstinence from all tobacco products of at least 30 days.

## Discussion

At the cross-sectional level, our findings clearly depict a relationship between FVC and all three indicators of nicotine dependence. The findings for fruit consumption suggest a low threshold for effect, while those for vegetable consumption suggest an effect only for those in the highest quartile of consumption. A somewhat similar pattern was observed for both abstinence among attempters and overall abstinence for fruit consumption, such that persons in quartiles 2 and 4 of fruit consumption were more likely to be abstinent than those who consumed the least amount of fruit. For vegetables, overall abstinence was only elevated significantly among those that consumed the highest amount. Abstinence among attempters and overall abstinence were only elevated for those in the highest categories of combined FVC. Future studies are needed to determine whether these relationships are consistently observed.

### Possible Mechanisms

Given the cross-sectional findings of an association between FVC and dependence, we postulated that FVC may mediate abstinence through an overarching mechanism of dependence; that is, smokers with higher versus lower FVC may be less dependent on cigarettes, with lower dependence facilitating abstinence. However, in a simple statistical mediation test, baseline dependence did not significantly impact the model coefficients evaluating the relationship between FVC and abstinence at follow-up (data not shown). Experimental studies are needed to test whether FVC influences dependence and in turn abstinence.

Mechanisms proposed to explain cross-sectional research findings in prior studies linking diet and smoking include socio-economic status (Goldbourt & Medalie, 1975; Woodward et al., 1994), a more general unhealthy lifestyle of smokers (Goldbourt & Medalie, 1975; Woodward et al., 1994), food palatability to smokers (Grunberg, 1982; McClernon, Westman, Rose, & Lutz, 2007), and lower health knowledge among smokers (Woodward et al., 1994). Other mechanisms that might also provide explanation for a relationship between diet and smoking include the dopaminergic reward system (Balfour, 2004), differences in brain serotonin associated with mood and nicotine withdrawal (Bowen, Spring, & Fox, 1991), satiety (Giovino, 2007), and constipation (Giovino, 2007).

Smoking is consistently associated with indicators of socio-economic status (Dube et al., 2010; Giovino, 2002), as is diet quality (Darmon & Drewnowski, 2008). However, after adjusting for socioeconomic status indicators in cross-sectional studies, the relationship between FVC and smoking persists (e.g., Serdula et al., 1996; Subar et al., 1990; Woodward et al., 1994). In this study, indicators of dependence and abstinence from tobacco remained significantly associated with FVC after statistically adjusting for education and household income. While socioeconomic status may provide some explanation for the relationship, it does not appear to be a primary explanatory mechanism for the relationship between FVC and smoking.

Research indicates that smokers tend to have a more generally unhealthy lifestyle than people who do not smoke (Boyle, O’Connor, Pronk, & Tan, 2000; Schoenborn & Benson, 1998); however, this is not uniform. For example, in some studies that observed differences in diet among smokers versus nonsmokers, significant differences in physical activity based on smoking status were not observed (de Vries, Kremers, Smeets, & Reubsaet, 2008; Woodward et al., 1994). We adjusted for general health orientation using indicators of physical activity, alcohol use, and illicit drug use, and the relationship between FVC and smoking abstinence remained significant. A related mechanism has been proposed by Woodward et al. (1994), who observed that smokers had poorer dietary knowledge than nonsmokers and suggested a dietary improvement strategy involving increased FVC as a possible precursor to quitting.

Palatability is a mechanism in which foods may influence the taste of smoking or smoking may influence the taste of certain foods, thereby influencing consumption. McClernon et al. (2007) tested this mechanism by asking smokers whether certain foods worsen or enhance the flavor of cigarettes. The researchers found that F&Vs, noncaffeinated beverages, and dairy products worsened the perceived taste of cigarettes whereas meats, caffeinated beverages, and alcohol were perceived as enhancing the taste of cigarettes. This tracks very closely with dietary differences between smokers and nonsmokers. Changing smokers’ diets, for example, by increasing FVC and reducing meat consumption, could possibly facilitate quitting by influencing the perceived taste of cigarettes. The mechanism of palatability may also operate in the other direction where smoking may influence taste. A study by Pavlos et al. (2009) found that smokers had a higher taste threshold and generally fewer and flatter taste buds than nonsmokers.

Central to the neurobiology of both smoking and food consumption are the dopaminergic (Balfour, 2004) and serotonergic systems (Olausson, Engel, & Söderpalm, 2002) where dopamine acts on the appetitive or rewarding end and serotonin on the aversive or inhibitory end of opponency (Boureau & Dayan, 2010). The intake of nicotine via smoking activates nicotinic receptors in the brain, which in turn excite dopamine neurons and increase the release of dopamine. The increased dopamine release mediates the reinforcing properties of nicotine via feelings of pleasure, thus increasing the desire to engage the behavior (De Biasi & Dani, 2010). This is similar to the process in which the dopaminergic reward system works to reinforce intake of food (Balfour, 2004). The dopaminergic reward system might explain the decrease in consumption of sweet-tasting foods following smoking or nicotine administration observed by Grunberg (1982) and why recent ex-smokers tend to have a higher sugar consumption (Bolton-Smith, Woodward, Brown, & Tunstall-Pedoe, 1993). Sweet-tasting substances can serve as an alternative reinforcer to a drug (Campbell & Carroll, 2000). The alternative reinforcer hypothesis is further supported in smokers from a study by Helmers and Young (1998), where abstaining smokers given sucrose (i.e., table sugar comprised of fructose and glucose) experienced less anxiety and other withdrawal symptoms than controls administered placebo. Smokers may use sugars as an alternative reinforcer to nicotine (West, 2001). The differences we observed in fruit consumption between smokers and nonsmokers may be explained by alternative reinforcement. The sugars in fruit and fruit juice may increase dopamine levels and thus reduce the craving or perceived need for a cigarette, resulting in fewer cigarettes smoked each day and less nicotine dependence. While a sugar or glucose substitute may seem more efficient, fruit contains fiber and many other beneficial nutrients, some of which may also interact with dopamine; for example, Deshpande, Dhir, and Kulkarni (2006) found ascorbic acid (vitamin C) to interact with the dopaminergic system.

The serotonergic system has been linked to mood and moderation of the dopaminergic system (Cools, Nakamura, & Daw, 2011). Serotonin mediates the release of dopamine, reducing the reward associated with nicotine consumption (Boureau & Dayan, 2010). Lower levels of serotonin in the brain have been associated with feelings of anxiety and depression (Ressler & Nemeroff, 2000), which are in turn associated with smoking (Bonnet et al., 2005; Lasser et al., 2000). Lower serotonin levels during smoking abstinence may result in a more depressed mood and increased anxiety and cravings for cigarettes (Bowen et al., 1991). Cross-sectional studies have observed that high levels of FVC (and other dietary factors) are associated with lower levels of depression and anxiety (Cook & Benton, 1993; Jacka et al., 2010), and a recent cohort study observed that a whole food diet high in FVC was protective against depression (Akbaraly et al., 2009). Feelings of anxiety associated with both smoking withdrawal and diet may be a barrier to smoking cessation (Bowen et al., 1991), and negative affect is associated with smoking relapse (Shiffman & Waters, 2004). Further research is needed to determine more specifically how FVC (and other dietary factors) may influence the serotonergic system.

The mechanism of satiety may be related to cigarette smoking as cravings for cigarettes and foods are closely linked and often confused with one another (Kos, Hasenfratz, & Bättig, 1997; Pepino, Finkbeiner, & Mennella, 2009; West, 2001). Satiety is controlled physiologically by the hormones ghrelin and leptin. Ghrelin stimulates feelings of hunger when the stomach is empty of food (Levin et al., 2006). Feelings of hunger for food may be confused with cravings for cigarettes where either consumption of food or smoking may relieve these feelings. Leptin is released when the stomach is distended by food. It induces a feeling of satiety by suppressing ghrelin (Cummings & Foster, 2003), thereby resulting in reduced food intake (Huo, Maeng, Bjorbaek, & Grill, 2007). FVC is a major determinant of satiety, primarily due to the high fiber, water, and nutrient content of F&Vs (Rolls, Ello-Martin, & Tohill, 2004). Our findings suggest that a relatively large volume of F&Vs might be needed to adequately expand the stomach to mediate abstinence, given that abstinence was only increased in the fourth quartile of FVC.

Another possible role for FVC on cessation involves constipation. In a study of 514 clinic patients, who were abstinent for 4 weeks, 17% developed constipation, and for 9%, the problem was severe (Hajek, Gillison, & McRobbie, 2003). Constipation has also been linked to anxiety and depression (Cheng, Chan, Hui, & Lam, 2003), which, as discussed earlier, is associated with smoking. Compared with a low-fiber diet, a high-fiber diet from increased FVC reduces bowel transit time and increases the number of bowel movements, while reducing energy and fat intake (Kelsay, Behall, & Prather, 1978). A diet high in F&Vs may reduce the withdrawal symptom of constipation and also moderate weight gain after quitting smoking.

### Limitations

In our study, interviewees were limited to households with landline telephones. Those without a landline are more likely to be of lower income and younger age (Blumberg & Luke, 2007). The baseline response rate of 45.7% is somewhat in the range of this type of survey (e.g., BRFSS 2009 had a median response rate of 52.5% across states (Centers for Disease Control and Prevention, 2011); however, given the smaller sample size, including the loss of one-fourth of the sample at follow-up, it is not necessarily generalizable to the U.S. population. Nevertheless, it has an advantage of not being limited to one geographic region of the United States, and the findings of this study are not dependent on the sample being representative. Although this cohort study improves on previous cross-sectional analyses, the findings remain observational. An experimental study that included a significant change in FVC and multiple follow-up assessments would provide a more accurate understanding of the nature of the relationship. Other dietary factors could also be accounting for the relationship between FVC and smoking, thus, further cohort analysis that includes all of the major dietary components is warranted.

### Future Research Directions

While our findings suggest that FVC might influence smoking cessation, further cohort studies could verify these findings. An experimental study would optimally assess whether increasing FVC leads to reduced cigarette consumption, less nicotine dependence, and increased cessation. Animal studies to evaluate whether isocaloric intake of different types of foods influence nicotine self-administration would also be informative. Not only might FVC influence smoking cessation, low FVC might be a risk factor for smoking initiation. A longitudinal cohort study could best evaluate whether FVC predicts smoking initiation, adjusting for socioeconomic status, alcohol consumption, and other important variables. Future studies should include measures that would help evaluate the biological mechanisms discussed in this article. Studies that assess the possible role of other dietary factors should also be conducted.

## Conclusions

Our study extends the findings of multiple cross-sectional studies by observing that high FVC predicts smoking cessation. We also extend previous work by demonstrating that FVC is inversely related to indicators of nicotine dependence. Better educating smokers about healthier eating may help some to improve their diet, promoting cessation. Further research may identify dietary improvement as another component of the smoking cessation tool chest of practitioners and public health professionals.

## Funding

This work was supported by the Robert Wood Johnson Foundation's Innovators Combating Substance Abuse Program and the American Legacy Foundation.

## Declaration of Interests

None declared.

Buffalo State College, State University of New York was contracted to complete the interview portion of the study. The authors thank Sarah Klein, M.P.H., for prior data analysis.

## References

Abood
D
Conway
T
Smoking status, body composition, exercise, and diet among navy men
Health Values
,
1994
, vol.
18
(pg.
51
-
62
)
Akbaraly
TN
Brunner
EJ
Ferrie
JE
Marmot
MG
Kivimaki
M
Singh-Manoux
A
Dietary pattern and depressive symptoms in middle age
British Journal of Psychiatry
,
2009
, vol.
195
(pg.
408
-
413
doi:10.1192/bjp.bp.108.058925
American Association for Public Opinion Research
Standard definitions: Final dispositions of case codes and outcome rates for surveys
,
2011
7th ed
Deerfield, IL
Author
Baker
TB
Piper
ME
McCarthy
DE
Bolt
DM
Smith
SS
Kim
S.-Y.
, et al.  .
Time to first cigarette in the morning as an index of ability to quit smoking: Implications for nicotine dependence
Nicotine & Tobacco Research
,
2007
, vol.
9

Suppl. 4
(pg.
S555
-
S570
doi:10.1080/14622200701673480
Balfour
DJK
The neurobiology of tobacco dependence: A preclinical perspective on the role of the dopamine projections to the nucleus
Nicotine & Tobacco Research
,
2004
, vol.
6
(pg.
899
-
912
doi:10.1080/14622200412331324965
Birkett
NJ
Intake of fruits and vegetables in smokers
Public Health Nutrition
,
1999
, vol.
2
(pg.
217
-
222
doi:10.1017/S1368980099000270
Blumberg
SJ
Luke
JV
Public Opinion Quarterly
,
2007
, vol.
71
(pg.
734
-
749
doi:10.1093/poq/nfm047
Bolton-Smith
C
Casey
CE
Gey
KF
Smith
WC
Tunstall-Pedoe
H
Antioxidant vitamin intakes assessed using a food-frequency questionnaire: Correlation with biochemical status in smokers and non-smokers
British Journal of Nutrition
,
1991
, vol.
65
(pg.
337
-
346
doi:10.1079/BJN19910094
Bolton-Smith
C
Woodward
M
Brown
CA
Tunstall-Pedoe
H
Nutrient intake by duration of ex-smoking in the Scottish Heart Health Study
British Journal of Nutrition
,
1993
, vol.
69
(pg.
315
-
332
doi:10.1079/BJN19930036
Bonnet
F
Irving
K
Terra
J.-L.
Nony
P
Berthezène
F
Moulin
P
Anxiety and depression are associated with unhealthy lifestyle in patients at risk of cardiovascular disease
Atherosclerosis
,
2005
, vol.
178
(pg.
339
-
344
doi:10.1016/j.atherosclerosis.2004.08.035
Boureau
YL
Dayan
P
Opponency revisited: Competition and cooperation between dopamine and serotonin
Neuropsychopharmacology
,
2010
, vol.
36
(pg.
74
-
97
doi:10.1038/npp.2010.151
Bowen
DJ
Spring
B
Fox
E
Tryptophan and high-carbohydrate diets as adjuncts to smoking cessation therapy
Journal of Behavioral Medicine
,
1991
, vol.
14
(pg.
97
-
110
doi:10.1007/BF00846173
Boyle
RG
O’Connor
P
Pronk
N
Tan
A
Health behaviors of smokers, ex-Smokers, and never smokers in an HMO
Preventive Medicine
,
2000
, vol.
31
(pg.
177
-
182
doi:10.1006/pmed.2000.0699
Campbell
UC
Carroll
ME
Reduction of drug self-administration by an alternative non-drug reinforcer in rhesus monkeys: Magnitude and temporal effects
Psychopharmacology (Berlin)
,
2000
, vol.
147
(pg.
418
-
425
doi:10.1007/s002130050011
Centers for Disease Control and Prevention
Behavioral risk factor surveillance system 2009 summary data quality report (Version #1 Revised: 02/18/2011)
,
2011
Atlanta, GA

Cheng
C
Chan
AOO
Hui
WM
Lam
SK
Coping strategies, illness perception, anxiety and depression of patients with idiopathic constipation: A population-based study
Alimentary Pharmacology & Therapeutics
,
2003
, vol.
18
(pg.
319
-
326
doi:10.1046/j.1365-2036.2003.01663.x
Cook
R
Benton
D
The relationship between diet and mental health
Personality and Individual Differences
,
1993
, vol.
14
(pg.
397
-
403
doi:10.1016/0191-8869(93)90308-p
Cools
R
Nakamura
K
Daw
ND
Serotonin and dopamine: unifying affective, activational, and decision functions
Neuropsychopharmacology
,
2011
, vol.
36
(pg.
98
-
113
doi:npp2010121
Cummings
DE
Foster
KE
Ghrelin-leptin tango in body-weight regulation
Gastroenterology
,
2003
, vol.
124
(pg.
1532
-
1535
doi:10.1016/s0016-5085(03)00350-0
Dallongeville
J
Marécaux
N
Fruchart
J.-C.
Amouyel
P
Cigarette smoking is associated with unhealthy patterns of nutrient intake: A meta-analysis
Journal of Nutrition
,
1998
, vol.
128
(pg.
1450
-
1457
)
Darmon
N
Drewnowski
A
Does social class predict diet quality?
American Journal of Clinical Nutrition
,
2008
, vol.
87
(pg.
1107
-
1117
)
De Biasi
M
Dani
JA
Annual Review of Neuroscience
,
2010
, vol.
34
(pg.
105
-
130
doi:10.1146/annurev-neuro-061010-113734
Deshpande
C
Dhir
A
Kulkarni
SK
Antagonistic activity of ascorbic acid (vitamin C) on dopaminergic modulation: Apomorphine-induced stereotypic behavior in mice
Pharmacology
,
2006
, vol.
77
(pg.
38
-
45
doi:10.1159/000092409
de Vries
H
Kremers
S
Smeets
T
Reubsaet
A
Clustering of diet, physical activity and smoking and a general willingness to change
Psychology & Health
,
2008
, vol.
23
(pg.
265
-
278
doi:10.1080/14768320701349107
Dube
S
McClave
A
James
C
Caraballo
R
Kaufmann
R
Pechacek
T
Vital signs: Current cigarette smoking among adults aged ≥ 18 years—United States, 2009
,
2010
Atlanta, GA
U.S. CDC
Giovino
GA
Epidemiology of tobacco use in the United States
Oncogene
,
2002
, vol.
21
(pg.
7326
-
7340
doi:10.1038/sj.onc.1205808
Giovino
GA
Henningfield
JE
Santora
PB
Bickel
WK
Could nutritional factors influence the development and maintenance of addiction to nicotine?
Addiction treatment: Science and policy for the twenty-first century
,
2007
Baltimore, MD
Johns Hopkins University Press
(pg.
77
-
83
)
Goldbourt
U
Medalie
JH
Characteristics of smokers, nonsmokers and ex-smokers among 10,000 adult males in Israel. I. Distribution of selected sociodemographic and behavioral variables and the prevalence of disease
Israel Journal of Medical Sciences
,
1975
, vol.
11
(pg.
1079
-
1101
)
Grunberg
NE
The effects of nicotine and cigarette smoking on food consumption and taste preferences
,
1982
, vol.
7
(pg.
317
-
331
doi:10.1016/0306-4603(82)90001-6
Hajek
P
Gillison
F
McRobbie
H
Stopping smoking can cause constipation
,
2003
, vol.
98
(pg.
1563
-
1567
doi:10.1046/j.1360-0443.2003.00497.x
Helmers
KF
Young
SN
The effect of sucrose on acute tobacco withdrawal in women
Psychopharmacology
,
1998
, vol.
139
(pg.
217
-
221
doi:10.1007/s002130050707
Huo
L
Maeng
L
Bjorbaek
C
Grill
HJ
Leptin and the control of food intake: Neurons in the nucleus of the solitary tract are activated by both gastric distension and leptin
Endocrinology
,
2007
, vol.
148
(pg.
2189
-
2197
doi:10.1210/en.2006-1572
Jacka
FN
Pasco
JA
Mykletun
A
Williams
LJ
Hodge
AM
O’Reilly
SL
, et al.  .
Association of Western and traditional diets with depression and anxiety in women
American Journal of Psychiatry
,
2010
, vol.
167
(pg.
305
-
311
doi:10.1176/appi.ajp.2009.09060881
Kelsay
J
Behall
K
Prather
E
Effect of fiber from fruits and vegetables on metabolic responses of human subjects I. Bowel transit time, number of defecations, fecal weight, urinary excretions of energy and nitrogen and apparent digestibilities of energy, nitrogen, and fat
American Journal of Clinical Nutrition
,
1978
, vol.
31
(pg.
1149
-
1153
)
Kos
J
Hasenfratz
M
Bättig
K
Effects of a 2-day abstinence from smoking on dietary, cognitive, subjective, and physiologic parameters among younger and older female smokers
Physiology & Behavior
,
1997
, vol.
61
(pg.
671
-
678
doi:10.1016/s0031-9384(96)00518-5
Lasser
K
Boyd
JW
Woolhandler
S
Himmelstein
DU
McCormick
D
Bor
DH
Smoking and mental illness: A population-based prevalence study
Journal of the American Medical Association
,
2000
, vol.
284
(pg.
2606
-
2610
doi:10.1001/jama.284.20.2606
Levin
F
Edholm
T
Schmidt
PT
Gryback
P
Jacobsson
H
M
, et al.  .
Ghrelin stimulates gastric emptying and hunger in normal-weight humans
Journal of Clinical Endocrinology & Metabolism
,
2006
, vol.
91
(pg.
3296
-
3302
doi:10.1210/jc.2005-2638
McClernon
FJ
Westman
EC
Rose
JE
Lutz
AM
The effects of foods, beverages, and other factors on cigarette palatability
Nicotine & Tobacco Research
,
2007
, vol.
9
(pg.
505
-
510
doi:10.1080/14622200701243177
Morabia
A
Wynder
E
Dietary habits of smokers, people who never smoked, and exsmokers
American Journal of Clinical Nutrition
,
1990
, vol.
52
(pg.
933
-
937
)
National Cancer Institute
Tobacco use supplement to the Current Population Survey 2003
,
2003

Nelson
DE
Holtzman
D
Bolen
J
Stanwyck
CA
Mack
KA
Reliability and validity of measures from the Behavioral Risk Factor Surveillance System (BRFSS)
Sozial und Präventivmedizin
,
2001
, vol.
46

Suppl. 1
(pg.
S3
-
42
)
Nuttens
MC
Romon
M
Ruidavets
JB
Arveiler
D
Ducimetiere
P
Lecerf
JM
, et al.  .
Relationship between smoking and diet: The MONICA-France project
Journal of Internal Medicine
,
1992
, vol.
231
(pg.
349
-
356
doi:10.1111/j.1365-2796.1992.tb00943.x
Olausson
P
Engel
JA
Söderpalm
B
Involvement of serotonin in nicotine dependence: Processes relevant to positive and negative regulation of drug intake
Pharmacology Biochemistry and Behavior
,
2002
, vol.
71
(pg.
757
-
771
doi:10.1016/s0091-3057(01)00673-6
Palaniappan
U
Starkey
LJ
O’Loughlin
J
Gray-Donald
K
Fruit and vegetable consumption is lower and saturated fat intake is higher among Canadians reporting smoking
Journal of Nutrition
,
2001
, vol.
131
(pg.
1952
-
1958
)
Pavlos
P
Vasilios
N
Antonia
A
Dimitrios
K
Georgios
K
Georgios
A
Evaluation of young smokers and non-smokers with electrogustometry and contact endoscopy
BMC Ear, Nose and Throat Disorders
,
2009
, vol.
9
pg.
9

doi:10.1186/1472-6815-9-9
Pepino
MY
Finkbeiner
S
Mennella
JA
Similarities in food cravings and mood states between obese women and women who smoke tobacco
Obesity (Silver Spring, Md.)
,
2009
, vol.
17
(pg.
1158
-
1163
doi:10.1038/oby.2009.46
Ressler
KJ
Nemeroff
CB
Role of serotonergic and noradrenergic systems in the pathophysiology of depression and anxiety disorders
Depression and Anxiety
,
2000
, vol.
12

Suppl. 1
(pg.
2
-
19
doi:10.1002/1520-6394(2000)12:1+<2::AID-DA2>3.0.CO;2–4
Rolls
BJ
Ello-Martin
JA
Tohill
BC
What can intervention studies tell us about the relationship between fruit and vegetable consumption and weight management?
Nutrition Reviews
,
2004
, vol.
62
(pg.
1
-
17
doi:10.1111/j.1753-4887.2004.tb00001.x
Schoenborn
C
Benson
V
Relationship between smoking and other unhealthy habits: United States, 1985 (No. 154. DHHS Publication No. (PHS) 88–1250)
,
1998
Hyattsville, MD
Public Health Service
Serdula
MK
Byers
T
AH
Simoes
E
Mendlein
JM
Coates
RJ
The association between fruit and vegetable intake and chronic disease risk factors
Epidemiology
,
1996
, vol.
7
(pg.
161
-
165
doi:10.1097/00001648-199603000-00010
Shiffman
S
Waters
A
Negative affect and smoking lapses: A prospective analysis
Journal of Consulting and Clinical Psychology
,
2004
, vol.
72
(pg.
192
-
201
doi:10.1037/0022-006x.72.2.192
Shiffman
S
Waters
A
Hickcox
M
The nicotine dependence syndrome scale: A multidimensional measure of nicotine dependence
Nicotine & Tobacco Research
,
2004
, vol.
6
(pg.
327
-
348
doi:10.1080/1462220042000202481
Subar
AF
Harlan
LC
Mattson
ME
Food and nutrient intake differences between smokers and non-smokers in the US
American Journal of Public Health
,
1990
, vol.
80
(pg.
1323
-
1329
doi:10.2105/ajph.80.11.1323
West
R
Glucose for smoking cessation: Does it have a role?
CNS Drugs
,
2001
, vol.
15
(pg.
261
-
265
doi:10.2165/00023210-200115040-00001
Whichelow
MJ
Golding
JF
Treasure
EP
Comparison of some dietary habits of smokers and non-smokers
,
1988
, vol.
83
(pg.
295
-
304
doi:10.1111/j.1360-0443.1988.tb00471.x
Willett
W
Nutritional epidemiology
,
1998
2nd ed.
New York, NY
Oxford University Press

doi:10.1093/acprof:oso/9780195122978.001.0001
Woodward
M
Bolton-Smith
C
Tunstall-Pedoe
H
Deficient health knowledge, diet, and other lifestyles in smokers: Is a multifactorial approach required?
Preventive Medicine
,
1994
, vol.
23
(pg.
354
-
361
doi:10.1006/pmed.1994.1049
Yang
J
Giovino
GA
Smoking status and body mass index in the United States: 1996–2000
,
2007

Unpublished manuscript
Zondervan
KT
Ocke
MC
Smit
HA
Seidell
JC
Do dietary and supplementary intakes of antioxidants differ with smoking status?
International Journal of Epidemiology
,
1996
, vol.
25
(pg.
70
-
79
doi:10.1093/ije/25.1.70