Abstract
Objectives: This study aims to examine the (a) probability of transition between stages of alcohol involvement and (b) influence of tobacco use and nicotine dependence on transitions. Methods: Data came from Waves 1 and 2 of the National Epidemiologic Survey on Alcohol and Related Conditions. Latent transition analysis estimated the probability of transitioning between stages of alcohol involvement across waves and the impact of tobacco use and nicotine dependence at Wave 1 on transitions. Results: Males reporting current tobacco use but no dependence at Wave 1 were more likely to progress from No Problems to Moderate Problems (adjusted odds ratio [aOR] = 1.79; 95% confidence interval [CI] [1.44, 2.22]) and from No Problems to Severe Problems (aOR = 2.44; 95% CI [1.25, 4.77]) than nontobacco users. Females reporting current tobacco use but no dependence were more likely to progress from No Problems to Moderate Problems (aOR = 2.00; 95% CI [1.37, 2.94]) and from No Problems to Severe Problems (aOR = 2.87; 95% CI [1.34, 6.13]). Females reporting current tobacco use and dependence were more likely than females not using tobacco to transition from Moderate to No Problems (aOR = 2.10; 95% CI [1.04, 4.22]). Conclusions: Results suggest that tobacco use is a preceding correlate of progression in alcohol involvement among males and females. Among females, tobacco use and nicotine dependence are also related to alcohol involvement recovery.
| Original language | English (US) |
|---|---|
| Article number | e1789 |
| Journal | International Journal of Methods in Psychiatric Research |
| Volume | 28 |
| Issue number | 3 |
| DOIs | |
| State | Published - Sep 1 2019 |
Keywords
- NESARC
- alcohol use disorder
- latent class analysis
- latent transition analysis
- tobacco use
ASJC Scopus subject areas
- Psychiatry and Mental health
Access to Document
Other files and links
Fingerprint
Dive into the research topics of 'Current tobacco use, nicotine dependence, and transitions across stages of alcohol involvement: A latent transition analysis approach'. Together they form a unique fingerprint.Cite this
- APA
- Standard
- Harvard
- Vancouver
- Author
- BIBTEX
- RIS
In: International Journal of Methods in Psychiatric Research, Vol. 28, No. 3, e1789, 01.09.2019.
Research output: Contribution to journal › Article › peer-review
}
TY - JOUR
T1 - Current tobacco use, nicotine dependence, and transitions across stages of alcohol involvement
T2 - A latent transition analysis approach
AU - Pacek, Lauren R.
AU - Reboussin, Beth A.
AU - Green, Kerry M.
AU - LaFlair, Lareina N.
AU - Storr, Carla L.
AU - Alvanzo, Anika A.H.
AU - Mojtabai, Ramin
AU - Cullen, Bernadette
AU - Young, Andrea S.
AU - Tormohen, Kayla
AU - Riehm, Kira
AU - Crum, Rosa M.
N1 - Funding Information: The analyses and preparation of this project were supported by a grant from the National Institute on Alcohol Abuse and Alcoholism (R01AA016346). Dr. Pacek (K01DA043413), Dr. Young (K23DA044288), and Ms. Tormohen (T32DA007292) were supported by grants from the National Institute on Drug Abuse. Dr. Young was also supported by a NARSAD Young Investigator Award from the Brain and Behavior Research Foundation. The sample was drawn from Waves 1 (2001–2002) and 2 (2004–2005) of the National Epidemiologic Survey on Alcohol and Related Conditions (NESARC), a U.S. nationally representative survey conducted by the National Institute on Alcohol Abuse and Alcoholism (NIAAA). Wave 1 included 43,093 civilian participants age 18 and older. Of the 39,959 participants eligible for Wave 2 interviews, a total of 34,653 were re-interviewed in 2004–2005. Our sample included a total of 14,564 male and 20,089 female participants who participated in both the Wave 1 and Wave 2 assessment. Data were collected via computer-assisted personal interviews. Response rates for Waves 1 and 2 were 81% and 87%, respectively. Black, Hispanic, and young adult (age 18–24) participants were intentionally oversampled. Data were weighted based on demographic information from the 2000 Census in order to be representative of the U.S. noninstitutionalized, civilian population. Additional sampling procedures for the NESARC are described elsewhere (Grant et al.,). Analyses were based on de-identified publicly available data that are exempt from review by the Institutional Review Board. The sample was drawn from Waves 1 (2001–2002) and 2 (2004–2005) of the National Epidemiologic Survey on Alcohol and Related Conditions (NESARC), a U.S. nationally representative survey conducted by the National Institute on Alcohol Abuse and Alcoholism (NIAAA). Wave 1 included 43,093 civilian participants age 18 and older. Of the 39,959 participants eligible for Wave 2 interviews, a total of 34,653 were re-interviewed in 2004–2005. Our sample included a total of 14,564 male and 20,089 female participants who participated in both the Wave 1 and Wave 2 assessment. Data were collected via computer-assisted personal interviews. Response rates for Waves 1 and 2 were 81% and 87%, respectively. Black, Hispanic, and young adult (age 18–24) participants were intentionally oversampled. Data were weighted based on demographic information from the 2000 Census in order to be representative of the U.S. noninstitutionalized, civilian population. Additional sampling procedures for the NESARC are described elsewhere (Grant et al.,). Analyses were based on de-identified publicly available data that are exempt from review by the Institutional Review Board. Alcohol abuse and dependence criteria were assessed using the Alcohol Use Disorders and Associated Disabilities Interview Schedule (AUDADIS-IV; Grant et al.,), a structured diagnostic interview designed to assess alcohol, drug, and mental health disorders according to DSM-IV diagnostic criteria (American Psychiatric Association,). Sets of past-year symptom questions were combined to form dichotomous indicators (yes/no) of the various AUD criteria that included four alcohol abuse criteria, (a) recurrent drinking resulting in a failure to fulfill major role obligations, (b) recurrent drinking in hazardous situations, (c) recurrent drinking-related legal problems, and (d) continued drinking despite recurrent interpersonal problems caused or exacerbated by drinking, and seven alcohol dependence criteria, (a) tolerance, (b) having two or more withdrawal symptoms, (c) drinking larger amounts/for a longer period than intended, (d) having a persistent desire/unsuccessful attempts to cut down on drinking, (e) spending a great deal of time obtaining alcohol, drinking, or recovering from drinking's effects, (f) giving up important social, occupational, or recreational activities in order to drink, and (g) continued drinking despite physical or psychological problems caused by drinking. Consistent with the approach adopted in DSM-5, abuse and dependence criteria were not separated. These alcohol abuse and dependence criteria, as well as this approach, have been used successfully in prior analyses using latent class analysis (LCA) and latent transition analysis (LTA) among women in the NESARC (La Flair et al.,), as well as analyses that included both men and women from the NESARC (Crum et al.,; Cullen et al.,). At Wave 1, the tobacco use section includes separate questions about the following five different types of tobacco: (a) cigarettes, (b) cigars, (c) pipe, (d) snuff, and (e) chewing tobacco. A variable exists in the data set that summarizes the overall tobacco use status of each person in the NESARC. The three values for this variable are (a) current user (past 12 months) of one or more types of tobacco; (b) ex-user (not in past 12 months) of one or more types of tobacco; and (c) lifetime nonuser of any type of tobacco. This variable was collapsed to form a binary variable in which 1 = current user and 0 = current nonuser (i.e., former user or lifetime abstainer). Past-year nicotine dependence diagnoses were made according to the DSM-IV criteria at Wave 1. Nicotine dependence was identified as having three out of the seven following criteria: (a) the need for more nicotine to achieve desired effect; (b) meeting the criteria for nicotine withdrawal syndromes; (c) using tobacco more frequently or in larger amount than initially intended; (d) the persistent desire or unsuccessful efforts to cut down on nicotine use; (e) a great deal of time spent using tobacco (e.g., chain smoking); (f) the necessity to give up activities in favor of nicotine use; and (g) continued use despite recurrent physical or psychological problems likely to have been caused by nicotine use. Nicotine withdrawal was assessed as a syndrome as described by the DSM-IV based on four symptoms: (a) the use of nicotine upon waking; (b) the use of nicotine after being in a situation in which nicotine was restricted; (c) the use of nicotine to relieve or avoid withdrawal symptoms; and (d) the need to wake up in the middle of the night to use nicotine. Based on the current tobacco use and nicotine dependence variables, a variable with three levels was generated: current nontobacco user; current non-nicotine dependent tobacco user; and current nicotine-dependent tobacco user. Other variables included in these analyses were tertiled age (categorized into 18–35, 36–49, and 50+ years), race (non-Hispanic White, non-Hispanic Black, Hispanic, and Other), education (< high school vs. ≥ high school graduate), lifetime mood disorder comorbidity (major depressive disorder, dysthymia, mania, and hypomania), lifetime anxiety disorder comorbidity (generalized anxiety disorder, panic disorder, social phobia, and specific phobia), history illicit drug use disorder, and family history of AUDs (any first degree relative). LCA was applied to the 11 DSM-IV AUD criteria to generate latent classes of alcohol involvement. LCA is a data-driven approach that allowed us to characterize patterns of clinical features of AUD observed in our data that may manifest prior to meeting full criteria for AUD or may represent qualitatively different subtypes (e.g., stages). Procedures and model fit statistics are provided in greater detail below and elsewhere (Crum et al.,). Based on previous findings of sex differences for AUDs (Grucza, Bucholz, Rice, & Bierut,; Keyes, Martins, Blanco, & Hasin,), models were fit separately for males and females. Next, we utilized LTA to estimate the probability of transitioning between the latent stages of alcohol involvement across the two waves. LTA is an extension of LCA to the longitudinal framework, which expresses change over time in terms of transition probabilities and models the impact of covariates on transitions using a multinomial regression formulation. Transition probabilities reflect the probability of transitioning from a latent stage at Wave 1 to another latent stage at Wave 2 (Collins & Wugalter,; Reboussin, Liang, & Reboussin,). Using LTA, we assessed our hypothesis that baseline tobacco use—both with and without nicotine dependence—impacted the probability of progressive transition from one alcohol involvement stage to a more advanced stage. Simultaneously, we assessed whether baseline nicotine-dependent and nondependent tobacco use were associated with a reduced probability of remission to a less symptomatic alcohol involvement stage. We utilized the propensity score method of inverse probability of treatment weighting (IPTW) to address potential baseline differences between nicotine-dependent current tobacco users, non-nicotine-dependent current tobacco users, and nontobacco users, which could bias the effect estimates (Curtis, Hammill, Eisenstein, Kramer, & Anstrom,; Rubin,; Stuart,). In this technique, propensity scores (i.e., probability of tobacco use, with and without nicotine dependence) are computed using a multinomial logistic regression model; they were computed via the SAS software (version 9.4; SAS Institute,) TWANG macros, which can support estimation of propensity scores and their associated weights for comparisons involving two or more treatment groups. These scores reflect each participant's likelihood of being a nicotine-dependent current tobacco user, non-nicotine-dependent current tobacco user, or nonuser, given their sociodemographic and clinical characteristics. Propensity score methods are designed to replicate a randomized design where the observed confounders are balanced in the exposed and comparison groups (i.e., the three tobacco groups). Therefore, this approach removes the influence of differences in distributions of observed covariates by creating balance in the tobacco groups. See Stuart () for more information. Next, data are weighted by their inverse probability of being in their observed group (i.e., those who are and are not nicotine-dependent current tobacco users, non-nicotine-dependent current tobacco users, or nonusers). Because the NESARC used a complex sampling design, both LCA and LTA were carried out using Mplus version 7.0 (Muthén & Muthén,) taking into account survey weights, clustering, and stratification. The propensity score weights were multiplied by the survey weights, and the resulting combined weights were used in the analysis of the association of nicotine-dependent and nondependent tobacco use with transitions in alcohol involvement stages (Dugoff, Schuler, & Stuart,). To assess the effectiveness of IPTW in balancing the composition of nicotine-dependent, nondependent, and nontobacco using groups, we compared characteristics of the group before and after applying the weights (Stuart,). Application of IPTW was deemed successful as the groups were similar with respect to the observed characteristics after using the weights. Alcohol and tobacco are two of the most commonly used psychoactive substances in the United States; recent data estimate the prevalence of current tobacco and alcohol use to be 21.9% and 56.0%, respectively, among adults (National Institute on Alcohol Abuse and Alcoholism,; World Health Organization,). Prior research has consistently established an association between tobacco use and alcohol use disorders (AUD), with the two substances often used concurrently (Kalman, Morissette, & George,; Strine et al.,). For instance, a recent analysis of a dataset representative of the U.S. general population indicated that the prevalence of past-month cigarette smoking was greater among individuals with AUD (38%) and heavy alcohol use (49%) as compared with those without these conditions (18% and 19%, respectively; Weinberger, Gbedemah, & Goodwin,). Alcohol and tobacco use are each associated with significant morbidity and mortality. Despite decades of research and public health action, cigarette smoking remains the leading cause of preventable death in the United States, resulting in approximately 480,000 deaths annually (Centers for Disease Control and Prevention,), whereas excessive alcohol use remains the third-leading lifestyle-related cause of death, accounting for approximately 88,000 deaths annually (Centers for Disease Control and Prevention,). Moreover, combined exposure to alcohol and tobacco appears to have a synergistic effect, resulting in significant health consequences, including increased risk for certain head and neck cancers (Vaillant, Schnurr, Baron, & Gerber,; Znaor et al.,). What is less well understood is the role of tobacco in alcohol progression and recovery, though some work suggests that the use of both substances may be causally linked (Dermody & Donny,; McKee & Weinberger,). Specifically, in two German population-based studies, the number of cigarettes smoked, duration of daily smoking, nicotine dependence, and the number of nicotine dependence symptoms were associated with the severity of alcohol dependence (John et al.,), and lifetime heavier smoking, earlier onset smoking, and nicotine dependence were associated with past-year alcohol dependence (John, Meyer, Rumpf, & Hapke,). Similarly, among a sample of outpatients who met Diagnostic and Statistical Manual of Mental Disorders (DSM-IV) criteria for alcohol dependence, the amount of tobacco smoked was associated with the amount of alcohol consumed and the severity of alcohol dependence. Correlations were also observed between the severity of alcohol dependence and nicotine dependence (Batel, Pessione, Maître, & Rueff,). One notable limitation of these existing studies is that they have focused on examining cross-sectional associations. Explorations of prospective associations between alcohol and tobacco use are lacking. Research on gender differences concerning the relationship between tobacco and alcohol consumption has yielded mixed results. In a pooled series of Danish population studies, Jensen and colleagues () reported similar findings for both men and women, with the level of tobacco consumption predicting an increased risk of becoming a heavier or excessive drinker in a dose-dependent manner. Conversely, Acheson, Mahler, Chi, and de Wit () observed differential effects of nicotine on alcohol consumption among men and women: Nicotine administration increased alcohol consumption in males but decreased alcohol consumption among females. Reasons for the highly prevalent comorbid use of tobacco and alcohol are not entirely clear. Though certainly not an exhaustive list, potential explanations include environmental risk factors that promote the use of both substances, preexisting risk factors in individuals using both tobacco and alcohol, or pharmacological interactions between tobacco and alcohol, so that use of one facilitates use of the other. Evidence does exist to support the hypothesis that nicotine also increases the reinforcing properties of alcohol (Griffiths, Bigelow, & Liebson,; Rose et al.,). Thus, it is possible that in addition to acutely increasing the reinforcing effects of alcohol and alcohol consumption, nicotine and tobacco use may play a substantial role in altering alcohol consumption in the long term. The vast majority of research concerning the association between tobacco and alcohol use has focused on drinking status at the time of interview, or on lifetime drinking status, without accounting for changes in the course of alcohol involvement. This limitation is significant, considering that AUDs are often described as stage-sequential processes (Graham, Collins, Wugalter, Chung, & Hansen,; Guo, Collins, Hill, & Hawkins,). To our knowledge, few studies have explored the influence of tobacco use or nicotine dependence on stages of and transitions between stages of alcohol involvement. One exception showed that non-daily cigarette smoking was predictive of increased, problematic alcohol use among young adults (Harrison & McKee,). In order to explore the relationship between tobacco use and nicotine dependence with AUDs, the present study utilizes latent transition analysis to examine the potential influence of current tobacco use at Wave 1 on the probabilities of transitioning between stages of alcohol involvement, for men and women separately, among a U.S. population-based sample followed up after a 3-year period. The aim of the present study was to investigate the influence of current tobacco use on these transitions, for males and females separately. Publisher Copyright: © 2019 John Wiley & Sons, Ltd.
PY - 2019/9/1
Y1 - 2019/9/1
N2 - Objectives: This study aims to examine the (a) probability of transition between stages of alcohol involvement and (b) influence of tobacco use and nicotine dependence on transitions. Methods: Data came from Waves 1 and 2 of the National Epidemiologic Survey on Alcohol and Related Conditions. Latent transition analysis estimated the probability of transitioning between stages of alcohol involvement across waves and the impact of tobacco use and nicotine dependence at Wave 1 on transitions. Results: Males reporting current tobacco use but no dependence at Wave 1 were more likely to progress from No Problems to Moderate Problems (adjusted odds ratio [aOR] = 1.79; 95% confidence interval [CI] [1.44, 2.22]) and from No Problems to Severe Problems (aOR = 2.44; 95% CI [1.25, 4.77]) than nontobacco users. Females reporting current tobacco use but no dependence were more likely to progress from No Problems to Moderate Problems (aOR = 2.00; 95% CI [1.37, 2.94]) and from No Problems to Severe Problems (aOR = 2.87; 95% CI [1.34, 6.13]). Females reporting current tobacco use and dependence were more likely than females not using tobacco to transition from Moderate to No Problems (aOR = 2.10; 95% CI [1.04, 4.22]). Conclusions: Results suggest that tobacco use is a preceding correlate of progression in alcohol involvement among males and females. Among females, tobacco use and nicotine dependence are also related to alcohol involvement recovery.
AB - Objectives: This study aims to examine the (a) probability of transition between stages of alcohol involvement and (b) influence of tobacco use and nicotine dependence on transitions. Methods: Data came from Waves 1 and 2 of the National Epidemiologic Survey on Alcohol and Related Conditions. Latent transition analysis estimated the probability of transitioning between stages of alcohol involvement across waves and the impact of tobacco use and nicotine dependence at Wave 1 on transitions. Results: Males reporting current tobacco use but no dependence at Wave 1 were more likely to progress from No Problems to Moderate Problems (adjusted odds ratio [aOR] = 1.79; 95% confidence interval [CI] [1.44, 2.22]) and from No Problems to Severe Problems (aOR = 2.44; 95% CI [1.25, 4.77]) than nontobacco users. Females reporting current tobacco use but no dependence were more likely to progress from No Problems to Moderate Problems (aOR = 2.00; 95% CI [1.37, 2.94]) and from No Problems to Severe Problems (aOR = 2.87; 95% CI [1.34, 6.13]). Females reporting current tobacco use and dependence were more likely than females not using tobacco to transition from Moderate to No Problems (aOR = 2.10; 95% CI [1.04, 4.22]). Conclusions: Results suggest that tobacco use is a preceding correlate of progression in alcohol involvement among males and females. Among females, tobacco use and nicotine dependence are also related to alcohol involvement recovery.
KW - NESARC
KW - alcohol use disorder
KW - latent class analysis
KW - latent transition analysis
KW - tobacco use
UR - http://www.scopus.com/inward/record.url?scp=85066480508&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85066480508&partnerID=8YFLogxK
U2 - 10.1002/mpr.1789
DO - 10.1002/mpr.1789
M3 - Article
C2 - 31141253
AN - SCOPUS:85066480508
SN - 1049-8931
VL - 28
JO - International Journal of Methods in Psychiatric Research
JF - International Journal of Methods in Psychiatric Research
IS - 3
M1 - e1789
ER -