Body Mass Index and Functional Impairment: The Explanatory Role of Anxiety Sensitivity Among Treatment-seeking Smokers

Abstract

Obesity and smoking are highly prevalent public health concerns in the United States. Data indicate that elevated BMI is related to functional impairment. However, there is limited understanding of mechanisms that may explain their comorbidity among smokers. The current study sought to test whether anxiety sensitivity explained the relation between BMI and functional impairment among 420 (46.9% females; M_age = 38 years, SD = 13.42) treatment-seeking, adult smokers. Results indicated that BMI yielded a significant indirect effect through anxiety sensitivity for functional impairment, b = 0.01, SE = .01, 95 % CI = [.002, .021]. These findings remained significant after controlling for participant sex, negative affectivity, tobacco dependence, psychopathology, and medical conditions (i.e., hypertension, heart problems, respiratory disease, asthma). Such data provide novel empirical evidence that, among smokers, BMI may be a risk factor for functional impairment indirectly through anxiety sensitivity. Overall, such findings could potentially inform the development of personalized interventions among this particularly vulnerable segment of the smoking population.

Introduction

Smoking is one of the leading causes of preventable death and disability in the United States (U.S.), contributing to approximately 480,000 premature deaths annually (Jamal et al., 2015; USDHHS, 2014). Obesity is another major public health concern (Lim et al., 2012) related to several negative health risks, including heart disease, stroke, type 2 diabetes, and certain types of cancer (Centers for Disease Control and Prevention, 2016; Ma & Xiao, 2010). Many smokers may be at risk for weight gain and obesity as a result of their tendency to engage in comorbid problematic health behaviors (e.g., physical inactivity, poor diet; Chiolero, Jacot-Sadowski, Faeh, Paccaud, & Cornuz, 2007). Indeed, existing work found that over 60% of smokers calling into a quit-line are either overweight or obese (Bush et al., 2012). Furthermore, the combined impact of obesity and smoking is associated with a 3.5 to 5.2 fold increased risk of mortality relative to normal weight non-smokers (Freedman et al., 2006; Whitlock et al., 2009). As such, weight loss and tobacco cessation have been a focus of public health efforts to decrease common negative health consequences associated with these risk factors (Centers for Disease Control and Prevention, 2016; Ma & Xiao, 2010; USDHHS, 2014). Despite the well-documented health consequences of comorbid obesity and smoking, limited work has evaluated how the co-occurrence of these risk factors may impact other clinically important outcomes.

Functional impairment has received little scientific attention in the weight or smoking literature despite the fact that it is the leading reason individuals seek treatment (Dobson & Beshai, 2013; Hunt & McKenna, 1993). Functional impairment reflects difficulties in key psychosocial domains of living, including deficits in work/school work, social life/leisure activities, and family life/home responsibilities (Sheehan, 2008). Initial evidence suggests that obesity and overweight status are related to greater functional impairment (Wei & Wu, 2014). Specifically, obese individuals score within the clinical range of impairment at a rate that is more than three times that of underweight and normal-weight individuals (Rø, Bang, Reas, & Rosenvinge, 2012). Independent of this work, smoking has been implicated as a risk factor of functional impairment (Zvolensky, Schmidt, & McCreary, 2003). Theoretically, smokers with higher Body Mass Index (a measure of overweight and obesity; [BMI]; USDHHS, 2012b) may experience greater functional impairment as a result of the additive negative consequences of smoking. Yet, no scientific work has examined the relation between BMI and functional impairment among smokers.

Beyond evaluating the relation between BMI and functional impairment among smokers, it is clinically important to evaluate mechanisms that underpin their association. Anxiety sensitivity is one plausible transdiagnostic factor that may serve to explicate the relation between BMI and functional impairment among smokers. Anxiety sensitivity reflects the tendency to fear anxiety-related sensations (Reiss & McNally, 1985) and is involved in a wide variety of health behaviors known to be associated with weight status (e.g., inactivity, maladaptive eating; Otto et al., 2016). Extant work has suggested anxiety sensitivity may further perpetuate negative health behaviors by lowering persistence to behavioral change (Otto et al., 2016). Some research has found high BMI (≥ 30) as a precursor to the development of common mental health disorders (e.g., depression; Gatineau & Dent, 2011). Furthermore, given that past work has documented the role of anxiety sensitivity in contributing to functional impairment (Korte, Brown, & Schmidt, 2013; McLeish, Zvolensky, Smits, Bonn-Miller, & Gregor, 2007), it would be useful to examine whether anxiety sensitivity accounted for the association between BMI and functional impairment.

Theoretically, smokers with higher BMI may be more apt to misinterpret internal bodily sensations because of poorer perceived or objective health (Gatineau & Dent, 2011; McLeish, Zvolensky, Bonn-Miller, & Bernstein, 2006; McLeish, Zvolensky, Marshall, & Leyro, 2009). As such, they may respond to internal sensations in a less adaptive fashion resulting in the fear of the expected negative consequences of anxiety and bodily sensations. For example, a smoker with a higher BMI may be more fearful of having a heart attack and misattribute a racing heart to a serious physical health issue when, in fact, they may be exhibiting symptoms of anxiety, nicotine withdrawal, or life stress. Thus, these smokers experience increased anxiety sensitivity, which may then contribute to difficulties in several areas of their lives, including work/school work, social life/leisure activities, and family life/home responsibilities.

Together, the current study tested the hypothesis that BMI would exert an indirect effect on functional impairment via anxiety sensitivity. Specifically, BMI was expected to be positively associated with anxiety sensitivity, which, in turn, would be associated with functional impairment. It was hypothesized that an effect of BMI on functional impairment via anxiety sensitivity would be evident over and above variance accounted for by theoretically-relevant covariates, including participant sex, negative affectivity, tobacco dependence, psychopathology, and medical conditions (i.e., hypertension, heart problems, respiratory disease, asthma; Kessler, Greenberg, Mickelson, Meneades, & Wang, 2001; McCabe, Lansing, Garland, & Hough, 2002; Norbert Schmitz, Johannes Kruse, & Joachim Kugler, 2003).

Method

Participants

Data from the current study includes 420 (46.9% females; M_age = 38 years, SD = 13.42) treatment-seeking, adult smokers recruited as part of a larger study designed to evaluate the efficacy of two smoking cessation interventions (Schmidt, Raines, Allan, & Zvolensky, 2016). Eligibility criteria for the current study included being between ages 18–65 and smoking at least 8 cigarettes per day for the past year. In the current sample, the average BMI was 26.9 and 73.1% fell within the “obese” range defined as a BMI of 30 or greater (Centers for Disease Control and Prevention, 2016). Exclusion criteria for the current study included, current use of any psychotherapy or pharmacotherapy for smoking cessation not provided by the researchers, current use of other tobacco products, current suicidality warranting immediate intervention and psychosis.

Measures

Demographics Questionnaire

A demographic form was used to collect data on participant sex (coded: 0 = male and 1 = female), age, race, education, and marital status for descriptive purposes. Participant sex was also used as a covariate.

Structured Clinical Interview for DSM-IV Diagnosis of Axis I Disorders Non-Patient Version (SCID-NP; First, Spitzer, Gibbon, & Williams, 2007)

The SCID-NP was used to describe the presence of current psychopathology of the sample. Presence/absence of current (past year) psychopathology was coded yes (1) or no (0). The interviews were administered by advanced doctoral level therapists and supervised by a licensed clinical psychologist. All interviews were audio-taped and reliability of 12.5% of interviews were checked for accuracy; no cases of disagreement were noted.

Positive and Negative Affect Schedule (PANAS; Watson, Clark, & Tellegen, 1988)

The PANAS is a self-report measure used to assess the degree to which respondents experience 20 different emotions and feelings (e.g., excited, distressed). Items are rated on a Likert-type scale ranging from 1 (very slightly or not at all) to 5 (extremely). The measure yields two factors, positive and negative affect, which have demonstrated strong psychometric properties (Watson et al., 1988). In the present study, the 10-item Negative Affect subscale was included as a covariate. Internal consistency for the Negative Affect subscale items was good (Cronbach’s α = .91).

Fagerström Test for Cigarette Dependence (FTCD; Fagerstrom, 2012)

The FTCD is a 6-item self-report assessment of levels of tobacco dependence with higher scores indicating higher levels of physiological dependence on tobacco (possible range 0–10). The FTCD items have high test-retest reliability (Pomerleau, Carton, Lutzke, Flessland, & Pomerleau, 1994), acceptable levels of internal consistency, and are closely related to key smoking variables (e.g., saliva cotinine; Heatherton, Kozlowski, Frecker, & Fagerstrom, 1991). Internal consistency for FTCD items in the current sample was low (Cronbach’s α = .61), which is not uncommon for this measure (Korte, Capron, Zvolensky, & Schmidt, 2013) and scales containing fewer than 10 items (DeVellis, 2003).

Medical History Form

The medical history form was utilized to assess for current and lifetime medical problems. Presence of well-established health risks that are associated with BMI, including hypertension (Colin Bell, Adair, & Popkin, 2002), heart problems (Joshi, Day, Lubowski, & Ambegaonkar, 2005), respiratory disease, and asthma (all coded: 0 = absent and 1 = present); Liu et al., 2015) were included as covariates.

Body Mass Index (BMI)

Height and weight were collected from participants via self-report. BMI was calculated per World Health Organization recommendations based on measured weight and height ([weight (pounds)]/[height (inches)2 × 703]); (WHO), 2000).

Anxiety Sensitivity Index-3 (ASI-3; Taylor et al., 2007)

The ASI-3 is an 18-item self-report measure, derived in part from the original ASI (Reiss & McNally, 1985), used to assess sensitivity to, and fear of, the potential negative consequences of anxiety-related symptoms and sensations. Respondents were asked to rate on a 5-point Likert-type scale ranging from 0 (very little) to 4 (very much), the extent to which they were concerned about these possible negative consequences. The ASI-3 has been validated among smokers (Farris et al., 2015). The ASI-3 yields three subscales (Physical, Cognitive and Social Concerns) which were utilized to form a latent anxiety sensitivity variable. The three subscales evidenced good internal consistency (Physical: Cronbach’s α = .87; Cognitive: Cronbach’s α = .91; Social Concerns: Cronbach’s α = .84).

Sheehan’s Disability Scale (SDS; Sheehan, 2008)

The SDS is a 3-item self-report measure used to assess impaired functioning in key psychosocial domains of living due to illness. The SDS assessed work/school work, social life/leisure activities, and family life/home responsibilities. Items are rated on a 10-point Likert-type scale ranging from 0 (not at all) to 10 (extremely). A latent dependent variable was created using the three items.

Procedures

Participants were recruited at two sites [(Vermont (n = 193) and Florida (n = 227)]. Interested persons responding to community-based advertisements (e.g., flyers, newspaper ads, radio announcements) contacted the research team and were provided with a detailed description of the study via phone. Participants were then screened for initial eligibility, and if eligible, scheduled for an appointment at the University of Vermont or Florida State University. After providing written informed consent, participants were interviewed using the SCID-I/NP and completed a computerized self-report assessment battery as well as biochemical verification of smoking status. Participants eligible for the larger trial were randomly assigned to one of two smoking cessation treatment: (a) Smoking Cessation Program or (b) Panic-Smoking Prevention Program. Participants were compensated at baseline and all time-points excluding the treatment sessions. The study protocol was approved by the Institutional Review Boards at the University of Vermont and Florida State University (clinicaltrials.gov # NCT01753141); all study procedures and treatment of human subjects were conducted in compliance with ethical standards of the American Psychological Association. The current study is based on secondary analyses of baseline (pre-treatment) data for a sub-set of the sample.

Analytic Strategy

Descriptive statistics and Pearson correlations among study variables were first examined. The primary hypothesis was examined using structural equation modeling. Anxiety sensitivity and functional impairment were represented as latent constructs with three indicators each, while BMI was represented as a manifest variable.

In the structural model, directional paths led from BMI to anxiety sensitivity and functional impairment as well as from anxiety sensitivity to functional impairment. Additionally, directional paths led from covariates (including, participant sex (coded: 0 = male and 1 = female), negative affectivity, tobacco dependence, psychopathology (coded: 0 = no and 1 = yes), and medical conditions (i.e., hypertension, heart problems, respiratory disease, asthma; all coded: 0 = absent and 1 = present) to anxiety sensitivity and functional impairment to adjust for these constructs in the model.

Models were fit with the maximum likelihood (ML) estimator in Mplus 7.31 (Muthén & Muthén, 1998–2013). Model fit was assessed with root mean square error of approximation (RMSEA), with values of less than .06 indicating excellent fit and values above .10 suggesting poor fit; Comparative Fit Index (CFI), with values between .95 and 1.00 indicated excellent fit and values between .90 and .94 indicated acceptable fit; and standardized root mean square residual (SRMR), with values less than .08 indicating acceptable fit (Hu & Bentler, 1999).

We employed bias-corrected bootstrapped 95% confidence intervals to test the hypothesized indirect effect. Specifically, an indirect effect can be defined as the product of path a (the association between the predictor variable [x] and the proposed explanatory variable [m]) and path b (the association between the proposed explanatory variable [m] and the criterion variable [y] controlling for x). The indirect effect is assumed to be statistically significant if the confidence intervals (CIs) around their product do not include zero (Preacher & Hayes, 2008; Zhao, Lynch, & Chen, 2010). The model of interest was examined with BMI (modeled as an observed variable) as the criterion variable, anxiety sensitivity (modeled as a latent variable) as the proposed explanatory variable, and functional impairment (modeled as a latent variable) as the criterion variable. Confidence intervals were estimated based on 5000 iterations (Preacher & Hayes, 2008). Bias corrected confidence intervals are computed for unstandardized estimates; therefore, unstandardized coefficients and 95% confidence intervals are presented.

Results

Descriptive statistics and relations of study variables are presented in Table 1 and Table 2. Anxiety sensitivity was correlated with both BMI (r = .15, p = .006) and functional impairment (r = .59, p < .001). Additionally, functional impairment was correlated with BMI (r = .12, p = .027). A measurement model in which no directional paths were specified and constructs of interest (i.e., BMI, anxiety sensitivity, and functional impairment) could correlate provided evidence for good fit for the data (χ²[12] = 21.21, p = .047, RMSEA = .04 [90% CI: .005, .072], SRMR = .02, CFI = .99). Thus, results from the measurement model supported us proceeding with the structural model. The structural model (see Figure 1) fit the data well, χ²(44) = 81.14, p < .001, RMSEA = .05 [90% CI: .029, .060], SRMR = .02, CFI = .98. The model explained 58% of the variance in anxiety sensitivity and 41% of the variance in functional impairment.

Table 1.

Descriptive Variables of Study Participants

Variable	N	%
Race
Caucasian	361	86.4%
African American/Non-Hispanic	30	7.2%
African American/Hispanic	2	0.5%
Hispanic	10	2.4%
Asian American	4	1.0%
Other	11	2.6%
Education
Graduate/professional degree	35	8.4%
Partial graduate/professional school	22	5.3%
Four-year college degree	60	14.4%
Two-year college degree	36	8.6%
Partial college	139	33.3%
High school/equivalent degree	100	23.9%
Partial high school	26	6.2%
Marital status
Married/living with partner	152	36.4%
Separated	14	3.3%
Divorced	77	18.4%
Never married	166	39.7%
Widowed	9	2.2%
Medical Conditions
Hypertension	62	14.9%
Asthma	54	12.9%
Heart Problems	20	7.2%
Respiratory Disease	14	3.4%
Psychopathology
Social Anxiety Disorder	44	10.5%
Specific Phobia	21	5.1%
Alcohol Use Disorder	16	3.9%
Major Depressive Disorder	16	3.8%
Posttraumatic Stress Disorder	14	3.3%
Generalized Anxiety Disorder	14	3.3%
Cannabis Use Disorder	10	2.4%
Panic Disorder with/without Agoraphobia	9	2.2%
Anxiety Disorder NOS	7	1.7%
Dysthymia	7	1.7%
Obsessive-Compulsive Disorder	6	1.4%
Depressive Disorder NOS	2	0.5%
Bipolar I or II Disorders	2	0.5%
Agoraphobia	1	0.2%
Anorexia Nervosa	1	0.2%

Table 2.

Descriptive statistics and Pearson correlations among study variables.

Variable	Mean/n (SD/%)	1	2	3	4	5	6	7	8	9
1. ASI3-Physical	26.86/416 (5.94)	-
2. ASI3-Cognitive	4.72/417 (4.74)	.65***	-
3. ASI3-Social	3.17/417 (4.34)	.61***	.68***	-
4. SDS-Work/School	7.27/417 (5.29)	.32***	.41***	.37***	-
5. SDS-Family/Home	1.47/399 (2.28)	.34***	.45***	.41***	.63***	-
6. SDS-Social/Leisure	1.87/399 (2.49)	.33***	.41***	.44***	.66***	.70***	-
7. BMI	2.06/399 (2.62)	.19***	.11*	.08	.06	.12*	.11*	-
8. Anxiety Sensitivity	4.01/417 (1.55)							.15**	-
9. Functional Impairment	2.81/399 (1.3							.12*	.59***	-

Note.

^***p < .001,

^**p < .01,

^*p < .05.

N = 399–417; ASI3-Physical = Anxiety Sensitivity Index-3-Physical Subscale (Taylor et al., 2007); ASI3-Cognitive = Anxiety Sensitivity Index-3-Cognitive Subscale (Taylor et al., 2007); ASI3-Social = Anxiety Sensitivity Index-3-Social Subscale (Taylor et al., 2007); SDS-Work/School = Sheehan’s Disability Scale’s-Work/School Work Subscale (Sheehan, 2008); SDS-Family/Home = Sheehan’s Disability Scale’s-Family Life/Home Responsibilities Subscale (Sheehan, 2008); SDS-Social/Leisure = Sheehan’s Disability Scale’s-Social Life/Leisure Activities Subscale (Sheehan, 2008); Anxiety Sensitivity = Anxiety Sensitivity Index-3 (Taylor et al., 2007) three subscales (Physical, Cognitive and Social Concerns) were utilized to form a latent anxiety sensitivity variable; Functional Impairment = Sheehan’s Disability Scale’s (Sheehan, 2008) 3 items were used to form a latent functional impairment variable; BMI = Body Mass Index.

Figure 1.

Structural equation model of the indirect and direct effects of BMI on functional impairment via anxiety sensitivity.

Note. Unstandardized coefficients, standard rror and 95% confidence intervals are presented. BMI = Body Mass Index; Anxiety Sensitivity = Anxiety Sensitivity Index-3’s (Taylor et al., 2007) three subscales (Physical, Cognitive and Social Concerns) were utilized to form a latent anxiety sensitivity variable; Functional Impairment = Sheehan’s Disability Scale’s (Sheehan, 2008) 3 items were used to form a latent functional impairment variable. Covariates included participant sex (coded: 0 = male and 1 = female), negative affectivity, tobacco dependence, psychopathology (coded: 0 = no and 1 = yes), and medical conditions (i.e., hypertension, heart problems, respiratory disease, asthma; all coded: 0 = absent and 1 = present).

Findings from the hypothesized structural model indicated a significant association between BMI and anxiety sensitivity after controlling for covariates (b = 0.03, SE = .01, 95 % CI = [.008, .049]). Anxiety sensitivity had a significant association with functional impairment (b = 0.33, SE = .10, 95 % CI = .133 .520], but the direct effect from BMI to functional impairment after controlling for covariates was not significant, b = 0.01, SE = .01, 95 % CI = [−.008, .036]. Lastly, there was a medium (k² = 0.05) indirect effect of BMI on functional impairment through anxiety sensitivity, b = 0.01, SE = .01, 95 % CI = [.002, .021]. Higher BMI was associated with greater anxiety sensitivity; greater anxiety sensitivity, in turn, was associated with greater functional impairment.

To further strengthen the interpretation of the results, an alternative model was tested by reversing the proposed explanatory variable for the model (Preacher & Hayes, 2004); specifically, anxiety sensitivity was the predictor, BMI was the indirect variable, and functional impairment remained as the criterion variable. The indirect effect of the alternate model was non-significant, b = 0.02, SE = .02, 95 % CI = [−.008, .062].

Discussion

The current study tested anxiety sensitivity as an underlying mechanism in the relation between BMI and functional impairment among treatment-seeking smokers. Results indicated that BMI yielded a significant indirect effect through anxiety sensitivity for functional impairment with a medium effect size. Notably, the observed indirect effect was evident after adjusting for participant sex, negative affectivity, tobacco dependence, psychopathology, and medical conditions (i.e., hypertension, heart problems, respiratory disease, asthma). Importantly, competing models where non-significant, which provide further support and evidence for the proposed model. As hypothesized, higher BMI was related to increased anxiety sensitivity, which in turn, was related to increased functional impairment. Such results are broadly consistent with past work indicating that BMI and anxiety sensitivity are related to increased functional impairment (Korte, Brown, et al., 2013; Wei & Wu, 2014) and uniquely extends this work to indicate that anxiety sensitivity may serve as a mechanistic construct in the relation between BMI and functional impairment.

It is important to highlight the direct relation between BMI and anxiety sensitivity. To our knowledge, this is the only work to elucidate BMI as a predictor of anxiety sensitivity. Indeed, extant work has largely focused on anxiety sensitivity as a precursor to elevated BMI (Hearon, Utschig, Smits, Moshier, & Otto, 2013; Smits, Tart, Presnell, Rosenfield, & Otto, 2010). Yet, drawing the broader negative affective literature, BMI has been implicated as a predictor of other emotional vulnerabilities (e.g., depression; Gatineau & Dent, 2011). Thus, these data provide initial evidence that the relationship between BMI and anxiety sensitivity may be bi-directional. Additionally, the alternative model tested (testing the indirect effect of anxiety sensitivity via BMI) showed no significant indirect effect, suggesting the effect was specific to anxiety sensitivity. Accordingly, such data may provide useful in identifying a risk factor (e.g., elevated BMI) that may contribute to emotional vulnerabilities and subsequently, impairments in functioning among smokers.

Clinically, the present investigation may serve to conceptually inform the development of specialized intervention strategies for smokers with elevated BMI. Specifically, it may be advisable to implement weight loss management interventions to reduce BMI. Extant work has suggested that weight loss management may be achieved through psychological rather than physiological pathways (Annesi, 2011). Thus, among this population, it may be advisable to understand and clinically address the psychological component (i.e., anxiety sensitivity) to address weight loss and enhance functioning in key psychosocial domains. One therapeutic tactic that may target anxiety sensitivity among smokers is exercise. Indeed, existing work has found exercise to be an effective therapeutic technique in contributing to reductions in anxiety sensitivity among smokers through the exposure to physiological sensations (Smits et al., 2016). Furthermore, anxiety sensitivity reduction treatments may also have positive impacts on weight loss management, however, to our knowledge, this relation has not been examined among smokers. Future work may benefit from exploring therapeutic tactics that may reduce anxiety sensitivity and facilitate improvements in functional impairment among smokers with elevated BMI.

There are several study caveats. First, the current data were cross-sectional in nature. Thus, the directionality of the observed effects cannot be fully explicated despite the analytic modeling employed. Second, 73.1% of the sample was in the obese range. To increase generalizability, future studies may benefit by exploring smokers across a greater range of the BMI continuum. Third, the sample was largely comprised of White smokers. It will be important for future studies to replicate findings among a more ethnically/racially diverse sample of smokers. Fourth, although there was a significant indirect effect of BMI on functional impairment through anxiety sensitivity, the current study investigated a single yet novel underlying mechanism. Future work should focus on additional explanatory variables that may underlie this association, including other transdiagnostic factors (e.g., emotion dysregulation, distress intolerance, anhedonia). Finally, we employed BMI as the primary composite of weight status. BMI is by far the most commonly employed biomedical index of obesity (National Obesity Observatory, 2009; USDHHS, 2012a). However, BMI has some limitations, including inability to distinguish between muscle and adipose tissue (Nevill, Stewart, Olds, & Holder, 2006). Future work could therefore further explore the linkages between obesity using multiple methods of measuring weight status.

Overall, the present study provides initial evidence in the association between BMI, anxiety sensitivity, and functional impairment among treatment-seeking smokers. It may be possible that among high BMI smokers, targeting anxiety sensitivity reduction may improve outcomes in functional impairment. Future work is needed to replicate and validate the hypothesized model. Furthermore, future work may benefit from examining the nature of the observed variables overtime and explicate the potential impact of age.