This study described cigarette smoking behaviors among a diverse sample of middle and high school students and the association of cigarette smoking status with a range of lifestyle behaviors, including physical activity, eating patterns, and dietary intake. Self-reported current (past month) cigarette smoking rates in the sample were similar but somewhat lower than rates reported by national surveys of middle school students (7% versus 9%) and high school students (22% versus 28%) in 1999 (24
). In the present study, smoking was associated with grade level (high school), race and ethnicity (i.e., Native American, white, Hispanic), and with low-middle and middle SES but not with sex or weight status. These findings about the relationship of smoking with sex, race and ethnicity, and grade level are consistent with national surveillance data (3
). However, the lack of a relationship with weight status was not consistent with other findings that have shown higher rates of smoking among overweight adolescents (25
Overall, we found that adolescents who smoke were less likely to have healthful eating and physical activity habits. In agreement with other research, smoking frequency was inversely related to team sport participation (26
), but we did not observe strong associations with moderate or vigorous physical activity. Other research studies have reported strong inverse associations between physical activity and smoking, and it is possible that our findings did not support these studies because of measurement differences (6
). Most other studies have considered frequency rather than total hours of weekly physical activity and have focused only on vigorous physical activity that makes one breathe hard (6
). Our study also found an association suggesting reduced vigorous physical activity in adolescents who smoke more frequently; however, this association was at the cut-off point of statistical significance (P
= .05). Observations indicating a potential inverse association between smoking frequency and vigorous physical activity in adolescents are of particular concern because of the tendency for physical activity to decline as adolescents transition to young adulthood and because research has linked declining activity levels to increases in BMI (27
Because few studies have examined associations between smoking frequency and sedentary behaviors, in this study we considered whether a relationship exists between smoking and hours of watching television or videos. Although we found that the prevalence of exceeding the maximum recommended for hours of television- and video-watching was high in each category of smoking, exceeding the recommendation was unrelated to smoking frequency. These results contradicted our hypothesis, which we based on a prospective study of media use and smoking initiation in young people (29
). Given the observation in previous research of a direct, dose-response relationship between television viewing hours and rates of smoking initiation (29
), we hypothesized that more frequent smoking would be directly related to time spent watching television and videos. Although use of these media may be related to smoking initiation, in this study it did not appear to be related to the frequency of smoking among adolescents.
We observed associations between smoking and eating behaviors with greater consistency than associations between smoking and physical activity. One strong finding among students in our study was a direct association of smoking frequency with irregular meal patterns at breakfast, lunch, and dinner. Irregular meal patterns have been related to poorer intakes of key micronutrients (e.g., calcium, zinc, vitamin C, iron) and greater consumption of sugar (30
). Few other studies have investigated whether smoking frequency is related to fast-food consumption or family meal patterns (31
). The results of our study were in agreement with another study, which indicated that adolescent smoking frequency is directly related to more frequent intake of fast food (32
). In a previous analysis of the Project EAT data, we found associations between participation in family meals and a number of substance-use behaviors, including smoking (31
). Frequent intake of fast-food meals and irregular family meal patterns are of concern as these behaviors have also been related to diets of lower nutritional quality in adolescents and because fast-food intake has been shown to prospectively predict risk of increasing BMI in adolescent females (33
To investigate relationships between smoking and diet quality, we examined whether smoking frequency is associated with food-group servings and intake of selected nutrients among adolescents. In general, smoking frequency was inversely related to healthful food selections and to nutrient intakes adjusted for total energy intake. These findings build on previous studies that have reported similar associations between food choices and smoking behavior in adolescents (6
). The more comprehensive assessment of dietary patterns in this study demonstrated that, compared with the diets of nonsmokers, the diets of adolescents who reported frequent smoking were lower in several key nutrients important for their roles in the prevention of chronic disease (e.g., calcium for the prevention of osteoporosis).
One of the strengths of this study was its large and diverse sample of adolescents, which allowed us to assess for the effect of modification by sex and grade level. Other strengths were the range of physical activity and eating behaviors evaluated within this sample and the comprehensive assessment of dietary intake accomplished using a validated instrument (YAQ) (20
). The response rate of 81.5% was relatively high, and the demographic makeup of the study sample approximated the makeup of the source school populations.
In interpreting the results of this study, certain limitations should also be considered that may have interfered with our ability to observe the total effect of associations between smoking, physical activity, and nutrition behaviors. For example, parents' level of education is a widely used indicator of SES, but the economic benefits of higher education are not uniform, and this indicator may have produced incomplete adjustment for SES (37
). Despite efforts to further adjust associations for sex, grade level, and race and ethnicity, it is also possible that factors other than the covariates included in our models are related to the associations between smoking and patterns of poor diet and physical activity. This also could have influenced the observed pattern of associations.
Finally, using only a single item to assess cigarette use may have led to some misclassification of smoking status. Misclassification would have attenuated associations of smoking with physical and sedentary activity, dietary patterns, and BMI. Additional measures of lifetime smoking habits would have allowed for analyses to consider whether established habitual patterns of smoking behavior are associated with unhealthy lifestyle behaviors. Future research should use stronger measures of smoking behavior and use longitudinal study designs to clarify the temporal order of increases in smoking and decreases in healthful eating and activity behaviors.
Collectively, this study and previous research suggest the importance of early intervention in adolescents on smoking, physical activity, and dietary patterns. It is possible that the observed covariation in health behaviors is due to a greater belief in chance as an influence on health and a lower perceived benefit of engaging in healthful behaviors (38
). Researchers are still learning how adolescents receive, process, and prioritize information about health risks and how they modify their behavior as a result. Therefore, young people who smoke may benefit from comprehensive, coordinated interventions that address self-efficacy to improve or maintain personal health by engaging in regular physical activity, healthful eating behaviors, and other positive lifestyle behaviors in addition to reducing cigarette smoking. CSHPs are one example of how messages may be combined to address and reinforce multiple lifestyle behaviors through classroom instruction and environmental interventions (40
). These programs can involve all areas of a school in promoting good nutritional habits, physical activity, and a nonsmoking lifestyle. Programs should be designed not only to teach adolescents about the benefits of engaging in healthful behaviors but also the behavioral skills necessary to carry them out (40
). In addition, the school environment should be supportive of healthy lifestyle behaviors (e.g., healthy food available in the cafeteria and in vending machines, teachers and staff serving as role models for a physically active, smoke-free lifestyle). To have the greatest impact on reducing risk for chronic disease in adulthood, prevention programs need to target those at highest risk for unhealthy behaviors, engage youth early in adolescence, and focus on specific behaviors so that young people can observe and learn the benefits of adopting healthful lifestyle behaviors.