Body Mass Index Changes Associated with Smoking are Moderated by Race and Depression

Abstract

Smoking and obesity relate to several leading causes of death in the U.S. and are common within the criminal justice system. Previous studies demonstrate links between smoking, obesity, depression, and race but have not examined all four variables together. The current study evaluated these relationships after a smoking cessation intervention. Participants (N=500) were recruited from community corrections. The Center for Epidemiological Studies Depression Scale (CES-D) measured depression. Self-reported number of cigarettes and weight and height measurements assessed smoking status and body mass index (BMI) at baseline and 12-month follow-up. Depression was associated with increased BMI. Among Blacks without depression, there was a significant relationship between smoking and BMI, such that greater smoking reduction was associated with greater weight gain. This is the first study to assess the interaction between race, smoking, BMI and depression. These findings support tailoring smoking cessation and depression interventions for different races.

Smoking is the leading cause of preventable death in the U.S., accounting for one in five deaths each year.¹ Similarly, obesity is associated with several of the leading causes of death in the United States, including diabetes, heart disease, stroke and cancer.² Smoking is associated with several of the same diseases as obesity, including heart disease and cancer.¹ Almost 40% of adults in the U.S. are obese, with non-Hispanic Black people having the highest age-adjusted rates of obesity. This is especially true for non-Hispanic Black women among whom the prevalence of obesity is about 1.5 times the prevalence among non-Hispanic White women (while that of Black men is only marginally higher—1.08—than that of White men).²

Compared with non-smokers, smokers are less likely to be overweight or obese and weigh less on average than non-smokers. The mechanisms involved in the relationship between smoking and weight are multi-faceted and not completely understood. Smoking primarily affects weight due to nicotine; however, it may also regulate weight when used as a behavioral alternative to eating.³ Nicotine affects weight by increasing metabolic rate and decreasing caloric absorption, resulting in decreased caloric intake. Smoking one cigarette results in a 3% decrease in caloric consumption within 20 minutes.⁴ Additionally, nicotine increases energy expenditure due to its sympathomimetic properties.¹ There is also a well-established relationship between smoking cessation and weight gain, with individuals who quit smoking gaining approximately 10 pounds after one year of abstinence.³ In a study of female prisoners, a net weight gain of 10 pounds was seen at six months post-cessation between smokers and abstainers, although this weight gain attenuated by one year.⁵

Despite this, heavy smokers have been found to have greater body weight than light smokers or non-smokers, likely due to a collection of unhealthy behaviors, such as a sedentary lifestyle, poor diet and smoking. Heavy smokers are also more likely to engage in risk behaviors related to obesity (i.e., decreased physical activity, poor diet, and excess alcohol consumption) compared with light smokers or non-smokers.^6,7 Smoking has also been found to increase insulin resistance and central fat accumulation, which increases the risk for type 2 diabetes and metabolic syndrome.⁴ This risk is especially significant for Blacks, who have significantly greater insulin resistance than Whites.⁸

Smoking is also associated with impaired mental health. Smokers are more likely to experience depressive symptoms than non-smokers, and individuals who are depressed are more likely to smoke than those who are not. Further, depressed smokers are 40% less likely than non-depressed smokers to quit smoking.⁹ Similarly, according to the National Health and Nutrition Examination Surveys data, 43% of adults with depression were found to be obese,¹⁰ and prevalence of depression among adults with obesity has been reported to be up to double the prevalence among normal weight individuals.¹¹

Individuals within the criminal justice system represent more than 12% of all current smokers in the U.S.^12,13 Given that smoking is typically banned in prison settings, there has been a shift in attention to individuals who are in community corrections over prison settings. Individuals in the community corrections setting represent 70% of those in the criminal justice system, and have smoking rates of approximately 70–80%.^14,15 Similarly, obesity is a common problem within the criminal justice system, with one study finding that 77% of individuals within the criminal justice system were either overweight or obese. Furthermore, the two most common chronic medical conditions within the criminal justice system, coronary artery disease and hypertension, are associated with obesity.¹⁶ Most prisoners return to the community with health concerns identified in prison, cardiovascular disease being the most common. In terms of mental health, depression is the most prevalent condition within the criminal justice system, and a large proportion of individuals under community corrections have depressed mood.^16,17 There is also evidence that incarceration exacerbates both physical and mental health problems, and individuals with a history of incarceration are less likely to have insurance and access to health care than their peers.¹⁸ This risk is especially salient among Black people, who have higher proportional rates of probation and parole compared with Whites.¹⁹ Additionally, Blacks—in particular Black women—have higher rates of obesity and chronic depression compared with Whites.^2,20

While some relationships between smoking, obesity, depression and race are well established there is little research that assesses the relationship among all four variables. Further, there has been a recent push from both public health and criminal justice experts to research the relationship between the criminal justice system and health disparities, given that they both disproportionately affect minorities. The aim of the current study was to examine the relationship between smoking, weight, depression and race after a smoking cessation intervention.

Methods

Participants.

Data came from a clinical trial evaluating the effects of bupropion on smoking cessation among individuals under community correction supervision.²¹ Par ticipants (N=500) were recruited from a community corrections office and consented to participate in the study between October 2009 and February 2013. (The data for this paper are the result of a secondary data analysis, which was conducted after the primary analyses were completed.)

Demographic characteristics of participants who completed both baseline and 12-month follow up (N = 230) are detailed in Table 1.

Table 1.

DEMOGRAPHIC CHARACTERISTICS AND CHI-SQUARE ANALYSIS COMPARING INDIVIDUALS WHO DID OR DID NOT REDUCE THEIR BMI AT 12-MONTH FOLLOW UP

Variable	Overall Sample (N = 230)	Reduction in BMI (n=92)	No reduction in BMI (n=138)	p-value
Age
Mean	37.81	37.83	37.56	.421
SD	11.60	11.60	11.81
Race
White	23.0%	27.2%	21.0%	.280
Black	77.0%	72.8%	79.0%
Sex
Male	62.6%	64.1%	60.1%	.542
Female	37.4%	35.9%	39.9%
Marital Status
Never Married	57.7%	55.4%	59.9%	.441
Married	9.8%	7.6%	11.7%
Divorced/Separated	29.5%	32.6%	26.3%
Widow	3.0%	4.3%	2.2%
Education Level
<HS	30.3%	30.4%	39.9%	.737
HS/GED	38.0%	34.8%	39.4%
>HS	31.6%	34.8%	30.7%
Depressiona
Yes	57.3%	33.7%	48.9%	.023
No	42.7%	66.3%	51.1%
Reduction in Smoking
Yes	54.3%	54.4%	59.2%	.449
No	45.7%	45.6%	40.8%

Note:

^aDepression was measured with the Center for Epidemiological Studies Depression Scale (CES-D).

BMI = Body Mass Index

SD = Standard Deviation

Study inclusion criteria included being at least 19 years or older and currently under community corrections supervision (e.g., parole, drug court, probation), smoking at least five cigarettes per day, a cotinine level above 200 mg/mL at baseline, living in an environment that allows smoking (e.g., not living in a substance abuse rehabilitation center that prohibited smoking), and willing to take bupropion and receive four sessions of behavioral counseling to quit smoking. Participants were excluded from the study if they had a history of mania or eating disorder, a seizure disorder, current suicidal ideation or a suicide attempt in the past six months, were pregnant or nursing a child, were non-English speaking, had cognitive impairment, or were medically unstable.

Measures.

Demographic information including age, gender, race, highest educational attainment, marital status, and monthly income were obtained at baseline appointment. Current smoking status was assessed through self-reported number of cigarettes smoked yesterday at the baseline appointment and 12-month follow-up. Similarly, body mass index (BMI) was calculated from weight and height measurements at baseline appointment and 12-month follow-up. The Center for Epidemiological Studies Depression Scale (CES-D) assessed current depressive symptoms at the baseline appointment.²² The presence of depression was determined through the CES-D clinical cutoff score of 16 or higher.^22,23

Procedure.

This was part of a larger study assessing the effects of bupropion on smoking outcomes among individuals in the criminal justice system. The protocol is extensively detailed in another publication.²¹ All participants were recruited via flyers posted at the community corrections site. Eligible participants were scheduled for a baseline assessment at the community corrections offices at which time they provided informed consent and completed baseline questionnaires. All participants received 12 weeks of bupropion and were randomized to either four weekly 30-minute smoking cessation counseling sessions or brief physician advice to quit smoking. Participants completed follow-up appointments at six, nine, and 12 months post-baseline to assess smoking outcomes. This study was registered with ClinicalTrials.gov (NCT01257490) and received approval by the University of Alabama at Birmingham Institutional Review Board.

Data analytic approach.

A total of 500 participants were randomized and enrolled in the original trial. A total of 230 participants completed both baseline and 12-month follow-up appointments and were included in subsequent analyses. A log transformation of BMI was conducted to correct for violated normality. Changes in BMI and smoking were calculated from the difference between their 12-month assessment and baseline appointment. Chi-square analysis was used to examine differences in demographic characteristics among individuals with reductions in BMI compared with individuals without reductions in BMI at 12-months post-baseline. Regression analysis was used to predict change in BMI by change in smoking, depression (no depression v. depression), and race (White v. Black). PROCESS macro for SPSS²⁴ was used to examine the three-way interaction of race (White v. Black) moderating the moderation of depression (no depression v. depression) on the relationship between changes in smoking and BMI (see Figure 1).

Figure 1.

Three-way interaction of race (White v. Black) moderating the moderation of depression (no depression v. depression) on the relationship between changes in smoking and BMI

Results

Chi-square analysis was conducted to examine differences in demographic characteristics among individuals with or without reductions in BMI at 12-month follow-up (Table 1). Results revealed that individuals whose BMI increased were more likely to have depression (48.9%) compared with those whose BMI did decrease (33.7%) (p = .023). A linear regression analysis examining the associated change in BMI from change in smoking, depression and race was conducted. The overall model was significant (F(3,228)=3.05, p<.05 Adj. R²=0.026), and depression was an independent significant predictor of change in BMI (B = .011, t=2.25, p<.05), such that having depression leads to increases in BMI.

A three-way interaction was conducted to examine the moderated moderation of change in smoking’s effect on change in BMI by depression, which is itself moderated by race. The overall model revealed significant effects, F(7, 224) = 2.13, p = .042, R² = .062. Examination of the three-way interaction between change in smoking, depression, and race revealed no significant overall effects (p = .12). However, probing the interaction revealed significant conditional effects of changes in smoking on changes in BMI at specific values of the moderators (Figure 2). Specifically, among Blacks without depression there is a significant negative relationship between changes in smoking and BMI, such that greater smoking reduction was associated with greater weight gain. There was no significant interaction among Blacks with depression and Whites with or without depression.

Figure 2.

The interaction between depression and changes in smoking on changes in BMI for Blacks (12 month follow-up compared to baseline).

Note: BMI=Body Mass Index

Discussion

Results showed that depression, race and change in smoking significantly predict change in BMI over 12 months, suggesting the relationship between BMI and smoking may be moderated by other variables. Results of the three-way interaction indicated depression significantly moderates this relationship for Blacks. Blacks without depression responded to changes in smoking as expected: their BMI increased as smoking decreased. However, the interaction was not significant for Blacks with depression whose BMI seemed to increase at a consistent rate (around 0.5 over 12 months), regardless of change in smoking. However, there was no significant interaction effect between depression, change in BMI and change in smoking among Whites.

The difference in outcomes could be due to differing metabolic response to insulin resistance between Blacks and Whites.⁹ Since heavy smoking increases insulin resistance and promotes central fat accumulation, the BMI of non-depressed Blacks may directly respond to decreases or increases in smoking due to changes in insulin resistance. However, for Blacks with depression, the increased insulin resistance associated with depression may diminish any benefit in insulin resistance they could gain from decreasing smoking.²⁵ Future research should confirm and further test the associations between race, BMI, depressive symptoms and smoking cessation using metabolic bio-markers to better understand the underlying mechanisms.

To our knowledge, this is the first study to assess the interaction between race, smoking, BMI and depression, and our results provide a foundation for the develop ment of tailored interventions to address smoking cessation and depression in Blacks. A relationship between stress, depressive symptoms, and concerns about weight gain was associated with greater smoking dependence for Black but not White women in a previous study.²⁶ This suggests that smoking may be a way to regulate affect and negative emotions, particularly for Black women, which may further affect weight. In addition, in a sample of homeless adults, higher BMI and depressive symptoms were associated with greater weight concerns with smoking cessation.²⁷ These results support the need for consideration of depression and race when counseling patients on smoking cessation and considering the effect it may have on their weight.

Limitations.

This study has several limitations. First, cigarette smoking was assessed as the number of cigarettes smoked the day before, which might not accurately represent the fluctuations of smoking across time. While the findings indicate an association between depression, race, smoking reduction and BMI, the underlying mechanisms among Blacks remain unknown. Insulin may be an important factor, but was not collected in this study, and thus, cannot be tested or confirmed. Finally, these smokers were drawn from a sample of individuals under criminal justice supervision in the community and may not be generalizable to other higher-income smokers or other general-population smokers.