Addictive Behaviors 41 (2015) 238–246

Contents lists available at ScienceDirect

Addictive Behaviors

Tobacco use trajectories among a large cohort of treated smokers with posttraumatic stress disorder Carol A. Malte a,⁎, Paul A. Dennis b, Andrew J. Saxon a,c, Miles McFall a,c, Timothy P. Carmody d,e, William Unger f, Jean C. Beckham g,h a

Veterans Affairs Puget Sound Health Care System, Seattle, WA, United States Research and Development Service, Durham Veterans Affairs Medical Center, Durham, NC, United States Department of Psychiatry and Behavioral Sciences, University of Washington, Seattle, WA, United States d Veterans Affairs Medical Center, San Francisco, CA, United States e University of California, San Francisco, United States f Providence Veterans Affairs Medical Center, Providence, RI, United States g Veterans Affairs Mid-Atlantic Region Mental Illness Research, Education, and Clinical Center, Durham, NC, United States h Department of Psychiatry and Behavioral Sciences, Duke University Medical Center, Durham, NC, United States b c

H I G H L I G H T S • • • • •

We identified four unique tobacco use trajectories among treated smokers with PTSD. Trajectory groups were early sustained reduction, late sustained reduction, temporary reduction and no reduction. Early sustained and temporary reduction groups received similar cessation treatment. Temporary reducers had greater mental health symptoms than sustained reducers. Findings can aid development of targeted interventions for smokers with PTSD.

a r t i c l e

i n f o

Available online 4 November 2014 Keywords: Smoking cessation Posttraumatic stress disorder Veterans Major depressive disorder Relapse

a b s t r a c t Introduction: This study identified distinct tobacco use trajectories across 18 months in 943 veteran smokers with posttraumatic stress disorder (PTSD) in order to describe quit and relapse patterns, examine associations between trajectory groups on baseline characteristics and cessation service utilization, and explore group differences in mental health outcomes. Methods: Veterans who participated in a multisite, randomized trial of integrated smoking cessation care were grouped using k-means clustering based on reported daily tobacco use between baseline and 18 months. Four trajectory clusters were identified: no reduction (62%), temporary reduction (11%), late sustained reduction (9%) and early sustained reduction (18%). Results: Median quit times in the early, late, temporary, and no reduction groups were 451, 141.5, 97, and 2 days, respectively. Compared to the early reduction group, the temporary reduction group exhibited higher baseline depression (p b 0.01) and anxiety (p b 0.01), but did not differ in treatment received, with both groups attending significantly more cessation visits (p b 0.001) and more likely to receive recommended pharmacotherapy (p b 0.001) than the no reduction group between baseline and 6 months. The early reduction group exhibited lower depression relative to the no reduction (p b 0.01) and temporary reduction (p b 0.01) groups across all assessments between baseline and 18 months. Differences were not observed between groups in depressive or PTSD symptom change over time between baseline and 18 months. Conclusions: Tobacco use trajectories among treated smokers with PTSD vary distinctly. Characteristics of identified subgroups may lead to targeted interventions among smokers with PTSD and potentially other psychiatric disorders. Published by Elsevier Ltd.

⁎ Corresponding author at: Center of Excellence in Substance Abuse Treatment and Education, VA Puget Sound Health Care System, 1660 S. Columbian Way, S-116 ATC, Seattle, WA 98108, United States. Tel.: +1 206 277 3780; fax: +1 206 764 2293. E-mail address: [email protected] (C.A. Malte).

http://dx.doi.org/10.1016/j.addbeh.2014.10.034 0306-4603/Published by Elsevier Ltd.

1. Introduction Posttraumatic stress disorder (PTSD), one of the most commonly occurring mental health disorders (Kessler, Sonnega, Bromet, Hughes, &

C.A. Malte et al. / Addictive Behaviors 41 (2015) 238–246

Nelson, 1995), is highly associated with smoking. An estimated one in ten current U.S. smokers have lifetime PTSD (Lasser et al., 2000). In addition to smoking at disproportionally high rates, individuals with PTSD are less likely to quit successfully (Hapke et al., 2005) and are more likely to relapse quickly (Beckham, Calhoun, Dennis, Wilson, & Dedert, 2013; Zvolensky et al., 2008) than those without PTSD. Treatments such as integrated smoking cessation care (IC), which integrates guideline-based cessation treatment into PTSD care, have been shown to double one-year prolonged abstinence rates in veterans with PTSD compared to specialized treatment in Veterans Affairs (VA) smoking cessation clinics in a multisite, randomized trial (McFall et al., 2010). Nevertheless, a large majority of trial participants who made a quit attempt relapsed prior to achieving long-term abstinence. Increased understanding of tobacco use patterns and the factors associated with those patterns in this high-risk population could allow for development of more targeted interventions to promote prolonged smoking cessation. Prior longitudinal studies in the general population have identified smoking-related factors associated with relapse including severity of nicotine dependence (Hyland et al., 2006; Zhou et al., 2009), length of longest prior quit attempt (Hyland et al., 2006), and age of starting smoking (West, McEwen, Bolling, & Owen, 2001), as well as dynamic factors such as lack of cessation aids (Zhou et al., 2009). How such factors relate to lapse/relapse patterns over time in smokers with PTSD is not known, but may be particularly relevant given that individuals with PTSD are more likely to smoke heavily (Beckham et al., 1997) relative to those without PTSD. Mental health factors associated with relapse include negative affect (Carmody, Vieten, & Astin, 2007) and low positive affect (Cook, Spring, McChargue, & Doran, 2010; Leventhal, Ramsey, Brown, LaChance, & Kahler, 2008). Among those with PTSD, positive and negative affect, PTSD symptoms, and trauma cues are strongly related to smoking (Beckham et al., 2005, 2007), and negative affect and trauma reminders are related to relapse (Beckham et al., 2013). Cook, McFall, Calhoun, and Beckham (2007) posit that the chronic nature of PTSD may promote vulnerability to relapse even after significant periods of abstinence. However, much of the data on affect- and PTSD-related factors associated with relapse in this population are derived from short-term diarybased studies; little information is available on the long-term impact of these factors on lapse/relapse. The current study applied k-means clustering to data from a large multisite trial of IC (McFall et al., 2010) to identify distinct tobacco use trajectories across an 18-month period among smokers with PTSD randomized to IC or smoking cessation clinic (SCC). Trajectory groups were compared in order to: 1) describe unique patterns of quitting and relapse, as well as cessation treatment utilization; 2) examine prospective associations between baseline factors, including nicotine dependence, smoking behavior, and mental health, and group membership; and 3) explore differences between groups in PTSD and depressive symptom change over the study course. We hypothesized that trajectory groups with sustained reductions in tobacco use would demonstrate less severity with respect to both smoking-related and mental health factors and higher treatment utilization relative to those groups with no or minimal sustained reductions in use. 2. Methods 2.1. Study participants Between November 2004 and December 2007, veterans (N = 943) engaged in outpatient PTSD treatment at 10 VA medical centers enrolled in a randomized, controlled trial comparing IC to SCC. Eligibility criteria included: 1) military-related PTSD; 2) smoking N10 cigarettes on at least 15 of 30 days before screening; and 3) consent to receive cessation interventions. Veterans who used non-cigarette tobacco, met DSM-IV criteria for current psychotic, bipolar, or substance dependence

239

disorder other than nicotine, or exhibited severe psychiatric instability or cognitive impairment were excluded. Following baseline assessment, eligible veterans were randomized in a 1:1 ratio to IC or SCC and reassessed every 3 months for 18 months on tobacco use, cessation medication use, PTSD and depressive symptoms, and smoking cessation treatment utilization. Veterans in both conditions typically completed the initial cessation treatment course within 3 months of randomization. Those randomized to IC received 8 weekly individual sessions followed by monthly booster sessions delivered by their PTSD provider. Use of FDA-approved cessation medications was recommended but not required. Veterans randomized to SCC received care in accordance with local SCC policies. Veterans across conditions received ongoing mental health care through the PTSD clinic. The Human Rights Committee of the Palo Alto Cooperative Studies Program Coordinating Center and the Institutional Review Boards at participating sites approved the study. Veterans gave written informed consent prior to enrollment. Please see McFall et al. (2007, 2010) for a detailed description of procedures, interventions, and primary outcomes. 2.2. Outcome measures Self-reported demographics and tobacco use history were obtained at baseline. Nicotine dependence severity was measured by the Fagerström Test for Nicotine Dependence (FTND) (Heatherton, Kozlowski, Frecker, & Fagerstrom, 1991). Baseline health status was measured by the Smoking Cessation Quality of Life Questionnaire (SCQoL), which combines the veterans SF-36 (SF-36V) (Kazis et al., 1999) with 15 additional questions comprising five smoking cessation-targeted scales (Olufade et al., 1999). Baseline PTSD severity was measured by the Clinician Administered PTSD Scale (CAPS) (Weathers, Keane, & Davidson, 2001), with scores between 60 and 79 indicating severe PTSD symptomatology. Additional psychiatric diagnoses were determined using the Structured Clinical Interview for DSM-IV Axis I Disorders (SCID-I/P), a structured interview schedule used for diagnosing DSM-IV Axis I mental disorders (First, Spitzer, Gibbon, & Williams, 2001). To assess fidelity, assessors submitted audio recordings of CAPS and SCID-I/P interviews for outside review. The majority of SCID diagnoses from one participating site were excluded from analyses due to fidelity concerns. Daily self-reported tobacco use data, obtained via timeline followback interview method (Collins, Eck, Torchalla, Schroter, & Batra, 2009), were used to determine number of quit attempts, length of longest quit and days from baseline to longest quit. These outcomes, rather than categorical measures of point prevalence and prolonged abstinence (McFall et al., 2010), were selected to capture the full variation in quit and relapse patterns by cluster. The first 24-hour period after baseline with no tobacco use was considered the initial quit attempt. To avoid attributing multiple quit attempts to consistent non-daily smokers (e.g. using tobacco every other day), subsequent quit attempts were defined as a 24-hour period with no tobacco use following daily use for two weeks or more. Treatment process variables included total cessation treatment visits and type and amount of smoking cessation medications used. Data pertaining to treatment visits were extracted from Veterans Health Administration (VHA) electronic medical records. Participants' use of cessation medications was obtained using the timeline follow-back method. Change in PTSD severity over time was measured by the PTSD Checklist (PCL; range 17–85) (Weathers, Litz, Herman, Huska, & Keane, 1993). The Patient Health Questionnaire 9 (PHQ-9; range 0–27) (Kroenke, Spitzer, & Williams, 2001) measured depression over time. Veterans were determined to have probable major depressive disorder if they endorsed either anhedonia (item 1) or depressed mood (item 2) and endorsed ≥ 5 items per the algorithm in Kroenke et al. (2001).

240

C.A. Malte et al. / Addictive Behaviors 41 (2015) 238–246

2.3. Tobacco use cluster development To identify tobacco use trajectories, participants were clustered based on self-reported percent days of tobacco use, collected every 3 months between baseline and 18 months. Missing observations (7% of observations) were imputed via a single imputation using the imputation-by-chained-equations (ICE) program (Royston, 2004) in Stata version 12.1. The sample was randomly split into 2 subsamples (sample 1 n = 397; sample 2 n = 396; veterans with imputed observations excluded) to determine trajectory stability in independent analyses. Percent days of tobacco use data for each subsample were submitted to k-means clustering, in which the number of clusters is user-defined and initial cluster seeds are estimated. Participants are assigned iteratively to clusters based on the Euclidean distance from the seeds, which are subsequently replaced by cluster means. Assignment and re-assignment proceeds until changes in cluster means shrink below a pre-determined threshold. Initially, we examined all solutions between 2 and 10 clusters, evaluating each on consistency across the 2 subsamples, presence of small clusters (n ≤ 20), and minimum distance between clusters. The 10-, 9-, 8-, 7-, and 6-cluster solutions all yielded inconsistent trajectories and at least 1 cluster representing ≤ 20 participants. Remaining solutions were consistent across subsamples and yielded sufficiently large clusters. In both subsamples, the 4-cluster solution produced the second largest minimum distance between clusters, after the 2-cluster solution. To determine which of the 5-, 4-, 3-, and 2-cluster solutions most efficiently summarized the data, we conducted n (clusters) × 7 (time points) repeated-measures ANOVAs using the entire sample of nonimputed data (n = 793) to estimate the proportion of variance in tobacco use accounted for by each solution. We plotted these estimates—η2, partial η2, ω2, and adjusted R2—against the number of clusters specified to produce an analog to the scree plot. The plot of estimates of variance explained indicated an inflection point at the 4th cluster, further supporting a 4-cluster solution. The resulting clusters generated from the entire sample (N = 943) were labeled no reduction (n = 585), temporary reduction (n = 105), late sustained reduction (n = 82), and early sustained reduction (n = 171) (see Fig. 1). In subsequent text, these groups are referred to as no reduction, temporary reduction, late reduction, and early reduction. The clusters reflect group patterns of percent days of tobacco use over time rather than a particular participant meeting a set criterion for a successful quit (e.g., 7-day point prevalence). 2.4. Data analyses Descriptive statistics compared veteran characteristics, treatment received, and quit patterns in the four clusters. Unadjusted linear, logistic and multinomial logistic models, for continuous, dichotomous and categorical variables, respectively, were used to test for differences. Main analyses compared the no reduction cluster to the three clusters with reduced use, with post-estimation contrasts used to examine differences between the reduced use clusters. Multilevel mixed-effects linear regression models with random effects for VA site and time were used to examine change over time by cluster on PCL and PHQ-9 scores. Multilevel mixed-effects models yield individual change parameters based on available data and are less vulnerable to distortion by missing data at one or more assessments. The models examining change in outcome included assessment time (modeled as a categorical variable), cluster, and a time × cluster interaction term, and were adjusted for treatment (IC vs. SCC), age, baseline Fagerström score, quit attempt ≥1 year at baseline, age started smoking, baseline major depression and PTSD severity. Due to possible family-wise error resulting from multiple comparisons, results with a 2-sided p-value b 0.01 were considered significant. Analyses were conducted with Stata version 12.1. Robust standard error estimates were used whenever possible.

Fig. 1. Percent days of tobacco use by cluster. Error bars represent 95% confidence intervals.

3. Results 3.1. Baseline characteristics No between-cluster differences were detected in demographic variables, including age, gender, race, and education (Table 1). Compared to the no reduction group, veterans in the early reduction group were more likely to have a prior quit ≥ one year and were less likely to be living with smokers at baseline, and those in the temporary reduction group expressed greater confidence in their ability to quit. The temporary reduction group exhibited poorer mental health relative to the early reduction group per the PHQ-9 total score, the Mental Component Summary and anxiety scale from the SCQoL. Further, a greater proportion of the temporary reduction group met criteria for major depressive disorder per the PHQ-9 and the SCID-I/P compared to the no and early reduction groups. No between-cluster differences were detected in baseline PTSD symptoms or mental health conditions other than major depressive disorder. 3.2. Tobacco use patterns Veterans in the early reduction group had significantly longer quits relative to all other groups and achieved their longest quit earlier than the late and no reduction groups (Table 2). For the majority of veterans in the early and temporary reduction groups, their first quit was their longest, although those in the temporary reduction group had significantly more quit attempts (Mdn = 3 vs. 1), suggesting that these veterans continued to make quit attempts following relapse. In contrast, the majority (65.8%) of those in the late reduction group made at least one quit attempt prior to their longest quit. Regardless of cluster membership, the majority of veterans who relapsed prior to 18 months made another quit attempt (70%), with a typical length of time between first and second quits of ≤3 months. 3.3. Treatment differences Compared to the no reduction group, veterans in the early and temporary reduction groups were more likely to be randomized to IC (Table 1). With respect to cessation treatment received (Table 2), both the early and temporary reduction groups attended significantly more cessation visits (Mdn = 6.5 and 5.8, respectively) than the no reduction (3.1) and late reduction (4.0) groups between randomization and 6 months. Further, early and temporary reduction groups were twice as likely as the late and no reduction groups to receive long-term cessation pharmacotherapy (≥90 days) and twice as likely as the no reduction group to receive combination pharmacotherapy [e.g. nicotine

Table 1 Baseline characteristics by cluster.

Treatment condition Demographics

temporary reduction (TR) (n=105) n(%) or M(SD)

331 (56.6) 254 (43.4) 54.4 (8.9) 37 (6.3) 18 (3.1)

42 (40.0) 63 (60.0) 54.3 (8.5) 3 (2.9) 8 (7.6)

329 (56.2) 224 (38.3) 32 (5.5)

TR vs NR

late sustained reduction (LR) (n=82) n(%) or M(SD)

LR vs NR

LR vs TR

early sustained reduction (ER) (n=171) n(%) or M(SD)

39 (47.6) 43 (52.4) 54.4 (8.8) 6 (7.3) 5 (6.1)

59 (34.5) 112 (65.5) 55.6 (7.9) 14 (8.2) 8 (4.7)

66 (62.9) 31 (29.5) 8 (7.6)

54 (65.9) 25 (30.5) 3 (3.7)

98 (57.3) 65 (38.0) 8 (4.7)

250 (42.7) 248 (42.4) 61 (10.4) 26 (4.4) 380 (65.2)

44 (41.9) 44 (41.9) 13 (12.4) 4 (3.8) 75 (71.4)

46 (56.1) 26 (31.7) 8 (9.8) 2 (2.4) 52 (63.4)

89 (52.1) 63 (36.8) 13 (7.6) 6 (3.5) 127 (74.3)

124 (21.2) 18 (3.1) 443 (75.7) 423 (72.3)

18 (17.1) 8 (7.6) 79 (75.2) 86 (81.9)

12 (14.6) 4 (4.9) 66 (80.5) 60 (73.2)

28 (16.4) 2 (1.2) 141 (82.5) 126 (73.7)

447 (76.4) 82 (14.0) 67 (11.5) 62 (10.6) 185 (31.6)

85 (81.0) 10 (9.5) 8 (7.6) 8 (7.6) 29 (27.6)

60 (73.2) 17 (20.7) 11 (13.4) 3 (3.7) 25 (30.5)

139 (81.3) 22 (12.9) 12 (7.0) 14 (8.2) 44 (25.7)

167 (28.6)

43 (41.0)

25 (30.5)

68 (39.8)

*

ER vs NR

ER vs TR

ER vs LR

**

C.A. Malte et al. / Addictive Behaviors 41 (2015) 238–246

Smoking characteristics

Usual Care Integrated Care Age, M(SD) Female Hispanic Race White Black Other Marital Status Married Divorced/Separated Never married Widowed Education 12 or more years Employment Status Steady work Temporary work Unemployed Service Connected Disability Period of Servicea Vietnam Post-Vietnam Persian Gulf OEF/OIFb Smoked 25+ cigarettes/ day Past quit of 1+ year

no reduction (NR) (n=585)n(%) or M(SD)

* (continued on next page)

241

242

Table 1 (continued)

Quality of life (QOL) characteristics

temporary reduction (TR) (n=105) n(%) or M(SD)

Age Started Smoking, M(SD) Confidence ability to quit, 1-10, M(SD) Living with smokers, past 90 days Fagerström score, M(SD) CAPSc Total Score, M(SD)

16.9 (4.6)

18.7 (7.9)

6.9 (2.4)

7.6 (2.1)

251 (42.9)

Re-experiencing, M(SD) Avoidance, M(SD) Numbing, M(SD) Hypervigilance, M(SD) Distress, M(SD) PHQ-9d Total Score, M(SD) Met criteria for MDDe Current SCID Diagnosesf Depressive Disorder MDDe Anxiety Disorderg Smoking Cessation QOL, M(SD) PCSh MCSi Social Interactions Self-control Sleep Cognitive Functioning Anxiety

TR vs NR

late sustained reduction (LR) (n=82) n(%) or M(SD)

LR vs NR

LR vs TR

early sustained reduction (ER) (n=171) n(%) or M(SD)

ER vs NR

ER vs TR

17.1 (3.1)

18.2 (7.0)

7.3 (2.3)

7.3 (2.3)

42 (40.0)

33 (40.2)

53 (31.0)

5.7 (2.0) 75.6 (18.1)

5.7 (2.2) 77.0 (20.1)

5.1 (2.2) 73.4 (19.7)

5.5 (2.3) 73.9 (17.9)

20.2 (7.3) 9.9 (3.7) 18.3 (6.6) 24.9 (6.1) 7.5 (2.1) 13.5 (6.3)

20.9 (7.7) 10.2 (3.7) 18.3 (7.1) 25.2 (6.6) 7.9 (2.1) 15.0 (6.2)

19.7 (7.9) 9.7 (3.8) 17.8 (6.8) 24.0 (6.0) 7.8 (2.1) 13.0 (6.8)

19.8 (7.5) 9.9 (3.5) 17.5 (7.0) 24.3 (6.0) 7.4 (2.1) 13.0 (6.2)

*

296 (50.7)

68 (64.8)

44 (53.7)

83 (48.5)

*

293 (54.4) 256 (47.1) 122 (23.4)

51 (57.3) 47 (52.8) 22 (25.9)

44 (60.3) 40 (54.1) 18 (26.1)

71 (46.7) 55 (35.3) 31 (21.2)

38.2 (10.6) 34.4 (11.2) 54.0 (18.3) 40.6 (18.0) 33.2 (22.1) 46.5 (22.4) 43.4 (23.1)

38.4 (10.6) 31.5 (9.0) 51.1 (17.8) 41.1 (19.7) 30.8 (19.0) 44.0 (22.2) 38.0 (21.3)

37.9 (11.0) 34.2 (12.5) 51.8 (18.7) 42.7 (19.8) 33.7 (24.0) 46.8 (23.3) 43.1 (22.5)

38.3 (11.7) 34.9 (11.3) 54.2 (17.8) 39.4 (17.8) 34.5 (23.8) 48.1 (20.3) 46.1 (22.3)

*b0.01; **b0.001. a Percentages add up to N100% because multiple choices could be selected b Operation Enduring Freedom/Operation Iraqi Freedom c Clinician Administered PTSD Scale d Patient Health Questionnaire 9 e Major Depressive Disorder f SCID Diagnoses were not available on all subjects due to administrative issues at one site. g Excluded PTSD h Physical Component Summary i Mental Component Summary

*

*

*

ER vs LR

*

*

*

*

*

*

C.A. Malte et al. / Addictive Behaviors 41 (2015) 238–246

Mental health characteristics

no reduction (NR) (n=585)n(%) or M(SD)

C.A. Malte et al. / Addictive Behaviors 41 (2015) 238–246

243

Table 2 Cessation treatment received and tobacco use patterns between randomization and 18 months by cluster.

Treatment received Mean Treatment Visits, M(SD) 0 to 6 months 6 to 12 month 12 to 18 months Any Combination Medication Use, n(%) 0 to 6 months 6 to 12 month 12 to 18 months Any Medications ≥90 Days, n(%) 0 to 6 months 6 to 12 month 12 to 18 months 0 to 18 months Tobacco use patterns Longest quit (days), Mdn (IQR) Days to longest quit, Mdn (IQR)a Number of quit attempts, Mdn (IQR) Number of quit attempts, n (%) 0 1 2-3 4-5 ≥6 Longest quit attempt, n(%) 1 2-3 4-5 ≥6 ≥2 quit attempts if relapse prior to 18 months, n(%)a Days between 1st and 2nd quit attempt, Mdn (IQR)b Percent Living with Smoker, n(%) 0 - 6 months 6 - 12 months 12 - 18 months

no reduction (NR) (n=585)

temporary reduction (TR) (n=105)

TR vs NR

late sustained reduction (LR) (n=82)

LR vs NR

3.1 (3.6) 0.8 (2.1) 0.4 (1.6)

5.8 (4.0) 1.1 (2.0) 0.7 (1.5)

**

4.0 (4.0) 1.8 (2.9) 0.9 (2.5)

*

105 (18.6) 62 (11.3) 49 (9.1)

40 (38.1) 16 (15.8) 14 (14.1)

**

21 (26.3) 16 (20.3) 12 (15.6)

129 (23.3) 134 (24.8) 137 (26.1) 219 (40.0)

44 (44.0) 35 (35.4) 31 (32.0) 56 (56.0)

**

*

18 (22.8) 23 (29.9) 22 (29.3) 42 (53.2)

2 (0-13) 200 (79-381) 1 (0-2)

97 (60-143) 75 (51-138) 3 (2-5)

** ** **

141 (92-250) 302 (236-407) 2.5 (1-4)

260 (44.4) 128 (21.9) 143 (24.4) 27 (4.6) 27 (4.6)

0 (0.0) 25 (23.8) 41 (39.1) 22 (21.0) 17 (16.2)

**

3 (3.7) 21 (25.6) 29 (35.4) 18 (22.0) 11 (13.4)

205 (63.1) 93 (28.6) 18 (5.5) 9 (2.8) 197 (62.2)

70 (66.7) 29 (27.6) 6 (5.7) 0 (0.0) 80 (76.2)

*

27 (34.2) 36 (45.6) 10 (12.7) 6 (7.6) 58 (89.2)

91 (33-216)

48 (14-136)

*

273 (48.5) 265 (48.9) 260 (49.2)

43 (41.0) 39 (38.6) 39 (39.8)

LR vs early sustained reduction ER vs TR (ER) (n=171) NR

*

*

** ** **

** **

6.5 (4.1) 1.3 (2.1) 0.6 (1.5)

** *

64 (39.5) 25 (15.8) 12 (8.1)

**

72 (45.3) 63 (41.2) 48 (32.7) 87 (55.1)

** **

451 (262-502) 66 (45-131) 1 (1-2)

** **

**

*

*

**

10 (5.9) 102 (59.7) 35 (20.5) 14 (8.2) 10 (5.9)

**

** **

115 (71.4) 33 (20.5) 11 (6.8) 2 (1.2) 59 (80.8)

*

91 (46-174)

29 (12-64)

**

37 (46.8) 30 (39.0) 25 (33.8)

64 (39.5) 52 (32.9) 55 (37.2)

** *

**

ER vs ER vs TR LR

** **

** ** **

**

**

**

*

*pb0.01; **pb0.001. a Excludes those with no quit attempts. b Excludes those with b2 quit attempts.

patch plus short-acting (gum, lozenge) nicotine replacement therapy (NRT), bupropion plus NRT]. Fewer differences were detected after 6 months post-randomization, although the early and late reduction groups attended significantly more visits (Mdn = 1.3 and 1.8, respectively) than the no reduction group (0.8) between 6 and 12 months. 3.4. Changes in PTSD and depressive symptoms As shown in Fig. 2, there were significant main effects by group with the early reduction group exhibiting lower scores PCL relative to the no reduction group (β = −2.24; 95% CI = −3.91, −0.56; p = 0.009) and lower PHQ-9 scores relative to the no reduction (β = −1.07; 95% CI = −1.85, −0.30; p = 0.007) and temporary reduction (β = −1.46; 95% CI = −2.53, −0.39; p = 0.008) groups across all assessments. No significant differences were observed between groups in depressive or PTSD symptom change over time between baseline and 18 months. 4. Discussion Evaluation of tobacco use trajectories among 943 veterans with PTSD revealed four unique patterns of tobacco use reduction and relapse to increased use. To our knowledge, no prior study has examined tobacco use trajectories in a sample of adult smokers participating in a

cessation trial. Clusters differed considerably from one another with respect to length of longest quit, number of quit attempts, and cessation treatment utilization. The largest cluster (62% of the sample) smoked 95–99% of days across the study period and was distinguished by low treatment engagement, especially during the first 6 months. Similarities between clusters also were of note, with no differences in demographics and few differences in smoking characteristics detected at baseline, which contrasts with prior findings that severity of nicotine dependence (West et al., 2001; Zhou et al., 2009) and age of starting smoking (West et al., 2001) are associated with relapse. Whereas temporary and early sustained reduction groups mirrored one another with respect to early success in lowering tobacco use and high treatment utilization, the temporary reduction group reported higher levels of baseline depression and anxiety. This may in part explain their divergent tobacco use trajectories after month six. Median quit times in the early (451 days), late (141.5 days) and temporary (97 days) reduction groups indicated that veterans with chronic, severe PTSD can make meaningful quit attempts. Although quit times were short (Mdn = 2 days) in the no reduction group, the majority (65.5%) attempted to quit at least once, and one-third made multiple attempts. In fact, 70% of those who made a quit attempt and subsequently relapsed made additional quit attempts across the study. Those making a second quit attempt typically did so within 3 months

244

C.A. Malte et al. / Addictive Behaviors 41 (2015) 238–246

Fig. 2. Mental health symptoms over time by cluster. Error bars represent 95% confidence intervals. PCL = PTSD Checklist; PHQ-9 = Patient Health Questionnaire 9. The early reduction group exhibited lower PCL scores relative to the no reduction group (β = −2.24; 95% CI = −3.91, −0.56; p = 0.009) and lower PHQ-9 scores relative to the no reduction (β = −1.07; 95% CI = −1.85, −0.30; p = 0.007) and temporary reduction (β = −1.46; 95% CI = −2.53, −0.39; p = 0.008) groups across all assessments. No significant differences were detected between groups in symptom change over time between baseline and 18 months as measured by the PCL and PHQ-9.

of relapse, shorter than the median 1-year interval reported in a websurvey of current and former smokers (Yeomans et al., 2011). This is a positive observation given that future abstinence has been associated with multiple prior quit attempts (Kerr, Owen, & Capaldi, 2011). Likely this finding is due in part to the study's emphasis on re-engagement, which allowed veterans in both treatment conditions to re-engage in cessation services throughout the study. Veterans in the early reduction group were significantly more likely to be randomized to IC than SCC. This finding may be due in part to IC's reliance on proactive contacts to address factors associated with smoking and relapse in patients with PTSD, including exacerbation of smoking by psychiatric symptoms (Fu et al., 2007), severe withdrawal symptoms (Weinberger, Desai, & McKee, 2010), and possible deterioration in mood following quit attempts (Glassman, Covey, Stetner, & Rivelli, 2001; Killen, Fortmann, Schatzberg, Hayward, & Varady, 2003). Future study is needed to determine which aspects of IC are most helpful in promoting and maintaining abstinence. Notably, the temporary reduction cluster likewise included significantly more IC- than SCCrandomized veterans. Temporary cluster members made meaningful quit attempts, typically lasting over 90 days, with three-quarters making additional quit attempts following relapse, and accessed smoking cessation treatment at high rates, particularly early in treatment when relapse is common (Hughes, Keely, & Naud, 2004). Veterans in early and temporary reduction groups received approximately twice as many cessation visits and were twice as likely to receive recommended

pharmacotherapy (combination and long-term treatment) relative to those with no reduction between baseline and month six. Although tobacco use trajectories of the early and temporary reduction groups diverged after 6 months, the groups were similar in treatment received between 6 and 18 months. Thus lack of cessation aids, which has been associated with relapse (Zhou et al., 2009), did not appear to account for group differences in later abstinence. The temporary and early sustained reduction groups appeared similar at baseline with two important exceptions of higher depression and anxiety in the temporary reduction group. These findings are consistent with prior results showing that pre-quit negative affect and anhedonia are associated with relapse (Cook et al., 2010; Leventhal et al., 2008; McCarthy, Piasecki, Fiore, & Baker, 2006). Lower rates of current major depression (35% compared to approximately 50%) were seen in the early reduction group compared to all other groups. The early reduction group had consistently lower depressive symptoms over the study course relative to the no and temporary reductions groups, as well as lower PTSD symptoms relative to the no reduction group. However, the magnitude of the differences was small, with approximately 1 to 2 points separating the no and temporary reduction groups from the early reduction group on the PHQ-9 and 2 to 4 points separating the early and no reduction groups on the PCL after baseline. Further, we didn't detect differences between groups in changes in depression and PTSD symptom severity over time. This may be due in part to the measurement protocol; data collection occurred every three months and did not evaluate real time associations between negative affect, PTSD symptoms and lapse/relapse occasions, which may be important in detecting an influence of symptoms on relapse (Beckham et al., 2013) and may reveal larger, more clinically meaningful differences between groups. Future studies should evaluate symptom severity, disease burden and other psychological factors (e.g., anxiety sensitivity, distress tolerance, task persistence) associated with PTSD that may contribute in unique ways to sustained abstinence. 4.1. Clinical implications This study lends further support to the use of IC with smokers diagnosed with PTSD, with nearly two-thirds of veterans randomized to IC in the early reduction group. Early treatment engagement appears critical, as illustrated by the low engagement in the no reduction group, and reengagement efforts and ongoing access (Fiore et al., 2008) may encourage multiple quit attempts and result in extended abstinence among a subset of patients for whom early engagement efforts fail. However, for a significant percentage of veterans with PTSD, early and intensive treatment may prove insufficient to produce long-term abstinence. Even among a relatively homogenous group (i.e., combatrelated PTSD), those with more severe depression and anxiety may require additional and/or different intervention to sustain reductions in tobacco use. Assessment of major depression and anxiety levels at baseline and over the course of treatment may be helpful in determining which patients are likely to need more intensive intervention. 4.2. Limitations This study was limited primarily to older male veterans with PTSD. Further study is needed to replicate these smoking trajectories and evaluate findings in women, non-veterans, and smokers with other psychiatric disorders. Key factors associated to relapse, including anxiety sensitivity (Zvolensky, Bonn-Miller, Bernstein, & Marshall, 2006) and nicotine withdrawal, were not collected for this study. Such factors may be particularly important for individuals with PTSD, who endorse greater increases in anxiety in response to bodily arousal relative to those without PTSD (Feldner, Vujanovic, Gibson, & Zvolensky, 2008), and should be considered in future studies. Further, given that electronic diary studies have shown that mood and PTSD symptoms are related to smoking behavior, craving and lapse/relapse (Beckham et al., 2008,

C.A. Malte et al. / Addictive Behaviors 41 (2015) 238–246

2013), including these measures in future studies evaluating quit trajectories among smokers with PTSD would be useful. Tobacco use data are based on self-report and thus may overestimate days of abstinence. In addition, abstinence rates cannot be inferred from mean tobacco use days, so these results cannot be compared to abstinence rates reported in other studies. 4.3. Conclusions Despite these limitations, to our knowledge this is the first study to evaluate demographic, psychiatric and intervention components across different smoking trajectories in a large sample of carefully diagnosed smokers with PTSD over a long time period. Results suggested that cessation and relapse trajectories varied distinctly, with a sizable proportion of veterans relapsing following a meaningful quit and others achieving abstinence well into treatment after multiple quit attempts. Additional research is needed to determine how and when to intervene with individuals experiencing significant baseline depression and anxiety, who appeared to be vulnerable to relapse despite high levels of cessation treatment utilization. Going forward, further examination of these trajectories and their applicability to individuals with other mental health conditions, as well as to women and younger veterans, may lead to improved cessation treatments and outcomes. Role of funding sources This material is based upon work supported by the U.S. Department of Veterans Affairs, Office of Research and Development Cooperative Studies Program (CSP #519). The Veterans Affairs Puget Sound Center of Excellence in Substance Abuse Treatment and Education, the U.S. Department of Veterans Affairs, Office of Clinical Research and Development, and the Veterans Affairs Mid-Atlantic Mental Illness Research, Education, and Clinical Center provided additional support. The views expressed in this article are those of the authors and do not necessarily reflect the position or policy of the U.S. Department of Veterans Affairs. Contributors Ms. Malte contributed to the design of the study, coordinated study interventions at participating sites and conducted analyses. Drs. McFall and Saxon designed the study, wrote the protocol, and created the integrated invention that is the focus of this study. Drs. Beckham and Carmody contributed to the study design and protocol development and provided ongoing oversight of the study. Dr. Unger was the principal investigator at a study site, directing the day to day study operations. Dr. Dennis designed and provided ongoing consultation and review of all statistical analyses. All authors contributed to and have approved the final manuscript. Conflict of interest The authors of this work have no conflicts of interest to disclose. Acknowledgments The authors wish to thank Ying Lu, Ph.D., Bruce Chow, M.S., and Lisa Zehm, M.S. at the VA Palo Alto Cooperative Studies Program Coordinating Center for their ongoing support and input.

References Beckham, J.C., Calhoun, P.S., Dennis, M.F., Wilson, S.M., & Dedert, E.A. (2013). Predictors of lapse in first week of smoking abstinence in PTSD and non-PTSD smokers. Nicotine & Tobacco Research, 15(6), 1122–1129. http://dx.doi.org/10.1093/ntr/nts252. Beckham, J.C., Dennis, M.F., McClernon, F.J., Mozley, S.L., Collie, C.F., & Vrana, S.R. (2007). The effects of cigarette smoking on script-driven imagery in smokers with and without posttraumatic stress disorder. Addictive Behaviors, 32(12), 2900–2915. http://dx. doi.org/10.1016/j.addbeh.2007.04.026. Beckham, J.C., Feldman, M.E., Vrana, S.R., Mozley, S.L., Erkanli, A., Clancy, C.P., et al. (2005). Immediate antecedents of cigarette smoking in smokers with and without posttraumatic stress disorder: a preliminary study. Experimental and Clinical Psychopharmacology, 13(3), 219–228. http://dx.doi.org/10.1037/1064-1297.13. 3.219. Beckham, J.C., Kirby, A.C., Feldman, M.E., Hertzberg, M.A., Moore, S.D., Crawford, A.L., et al. (1997). Prevalence and correlates of heavy smoking in Vietnam veterans with chronic posttraumatic stress disorder. Addictive Behaviors, 22(5), 637–647. Beckham, J.C., Wiley, M.T., Miller, S.C., Dennis, M.F., Wilson, S.M., McClernon, F.J., et al. (2008). Ad lib smoking in post-traumatic stress disorder: an electronic diary study. Nicotine & Tobacco Research, 10(7), 1149–1157. http://dx.doi.org/10. 1080/14622200802123302. Carmody, T.P., Vieten, C., & Astin, J.A. (2007). Negative affect, emotional acceptance, and smoking cessation. Journal of Psychoactive Drugs, 39(4), 499–508.

245

Collins, S.E., Eck, S., Torchalla, I., Schroter, M., & Batra, A. (2009). Validity of the timeline followback among treatment-seeking smokers in Germany. Drug and Alcohol Dependence, 105(1–2), 164–167. http://dx.doi.org/10.1016/j.drugalcdep.2009. 05.023. Cook, J.W., McFall, M., Calhoun, P.S., & Beckham, J.C. (2007). Posttraumatic stress disorder and smoking relapse: a theoretical model. Journal of Traumatic Stress, 20(6), 989–998. http://dx.doi.org/10.1002/jts.20275. Cook, J., Spring, B., McChargue, D., & Doran, N. (2010). Effects of anhedonia on days to relapse among smokers with a history of depression: a brief report. Nicotine & Tobacco Research, 12(9), 978–982. http://dx.doi.org/10.1093/ntr/ntq118. Feldner, M. T., Vujanovic, A. A., Gibson, L. E., & Zvolensky, M. J. (2008). Posttraumatic stress disorder and anxious and fearful reactivity to bodily arousal: a test of the mediating role of nicotine withdrawal severity among daily smokers in 12-hr nicotine deprivation. Experimental and Clinical Psychopharmacology, 16(2), 144–155. http://dx.doi.org/10.1037/1064-1297.16.2.144. Fiore, M.C., Jaén, C.R., Baker, T.B., Bailey, W.C., Benowitz, N.L., Curry, S.J., et al. (2008). Treating tobacco use and dependence: 2008 update. Clinical practice guideline. Rockville, MD: U.S. Department of Health and Human Services. Public Health Service. First, M.B., Spitzer, R.L., Gibbon, M., & Williams, J.B.W. (2001). Structured clinical interview for DSM-IV-TR Axis I disorders, Research Version, Patient Edition (SCID-I/P). New York: Biometrics Research Department, New York State Psychiatric Institute. Fu, S., McFall, M., Saxon, A., Beckham, J., Carmody, T., Baker, D., et al. (2007). Posttraumatic stress disorder and smoking: a systematic review. Nicotine & Tobacco Research, 9, 1071–1084. http://dx.doi.org/10.1080/14622200701488418. Glassman, A.H., Covey, L.S., Stetner, F., & Rivelli, S. (2001). Smoking cessation and the course of major depression: a follow-up study. Lancet, 357, 1929–1932. Hapke, U., Schumann, A., Rumpf, H.J., John, U., Konerding, U., & Meyer, C. (2005). Association of smoking and nicotine dependence with trauma and posttraumatic stress disorder in a general population sample. The Journal of Nervous and Mental Disease, 193(12), 843–846. Heatherton, T.F., Kozlowski, L.T., Frecker, R.C., & Fagerstrom, K.O. (1991). The Fagerstrom test for nicotine dependence: a revision of the Fagerstrom Tolerance Questionnaire. British Journal of Addiction, 86(9), 1119–1127. Hughes, J.R., Keely, J., & Naud, S. (2004). Shape of the relapse curve and long-term abstinence among untreated smokers. Addiction, 99(1), 29–38. Hyland, A., Borland, R., Li, Q., Yong, H.H., McNeill, A., Fong, G.T., et al. (2006). Individuallevel predictors of cessation behaviours among participants in the International Tobacco Control (ITC) Four Country Survey. Tobacco Control, 15(Suppl. 3), iii83–iii94. http://dx.doi.org/10.1136/tc.2005.013516. Kazis, L.E., Ren, X.S., Lee, A., Skinner, K., Rogers, W., Clark, J., et al. (1999). Health status in VA patients: results from the Veterans Health Study. American Journal of Medical Quality, 14(1), 28–38. Kerr, D.C., Owen, L.D., & Capaldi, D.M. (2011). The timing of smoking onset, prolonged abstinence and relapse in men: a prospective study from ages 18 to 32 years. Addiction, 106(11), 2031–2038. http://dx.doi.org/10.1111/j.13600443.2011.03500.x. Kessler, R.C., Sonnega, A., Bromet, E., Hughes, M., & Nelson, C.B. (1995). Posttraumatic stress disorder in the national comorbidity survey. Archives of General Psychiatry, 52, 1048–1060. Killen, J.D., Fortmann, S.P., Schatzberg, A., Hayward, C., & Varady, A. (2003). Onset of major depression during treatment for nicotine dependence. Addictive Behaviors, 28(3), 461–470. Kroenke, K., Spitzer, R.L., & Williams, J.B. (2001). The PHQ-9: validity of a brief depression severity measure. Journal of General Internal Medicine, 16(9), 606–613. Lasser, K., Boyd, J.W., Woolhander, S., Himmelstein, D.U., McCormick, D., & Bor, D.H. (2000). Smoking and mental illness: a population-based prevalence study. Journal of the American Medical Association, 284, 2606–2610. http://dx.doi.org/10.1001/ jama.284.20.2606. Leventhal, A.M., Ramsey, S.E., Brown, R.A., LaChance, H.R., & Kahler, C.W. (2008). Dimensions of depressive symptoms and smoking cessation. Nicotine & Tobacco Research, 10(3), 507–517. http://dx.doi.org/10.1080/14622200801901971. McCarthy, D.E., Piasecki, T.M., Fiore, M.C., & Baker, T.B. (2006). Life before and after quitting smoking: an electronic diary study. Journal of Abnormal Psychology, 115(3), 454–466. http://dx.doi.org/10.1037/0021-843X.115.3.454. McFall, M., Saxon, A.J., Malte, C.A., Chow, B., Bailey, S., Baker, D.G., et al. (2010). Integrating tobacco cessation into mental health care for posttraumatic stress disorder: a randomized controlled trial. The Journal of the American Medical Association, 304(22), 2485–2493. http://dx.doi.org/10.1001/jama.2010.1769. McFall, M., Saxon, A.J., Thaneemit-Chen, S., Smith, M.W., Joseph, A.M., Carmody, T.P., et al. (2007). Integrating smoking cessation into mental health care for posttraumatic stress disorder. Clinical Trials, 4(2), 178–189. http://dx.doi.org/10. 1177/1740774507076923. Olufade, A.O., Shaw, J.W., Foster, S.A., Leischow, S.J., Hays, R.D., & Coons, S.J. (1999). Development of the Smoking Cessation Quality of Life questionnaire. Clinical Therapeutics, 21(12), 2113–2130. Royston, P. (2004). Multiple imputation of missing values. Stata Journal, 4(3), 227–241. Weathers, F.W., Keane, T.M., & Davidson, J.R. (2001). Clinician-administered PTSD scale: a review of the first ten years of research. Depression and Anxiety, 13(3), 132–156. Weathers, F.W., Litz, B.T., Herman, D.S., Huska, J.A., & Keane, T.M. (1993). The PTSD Checklist (PCL): reliability, validity, and diagnostic utility. Paper presented at the Annual Meeting of the International Society for Traumatic Stress Studies. San Antonio, TX. Weinberger, A.H., Desai, R.A., & McKee, S.A. (2010). Nicotine withdrawal in U.S. smokers with current mood, anxiety, alcohol use, and substance use disorders. Drug and Alcohol Dependence, 108, 7–12.

246

C.A. Malte et al. / Addictive Behaviors 41 (2015) 238–246

West, R., McEwen, A., Bolling, K., & Owen, L. (2001). Smoking cessation and smoking patterns in the general population: a 1-year follow-up. Addiction, 96(6), 891–902. http://dx.doi.org/10.1080/09652140020051013. Yeomans, K., Payne, K.A., Marton, J.P., Merikle, E.P., Proskorovsky, I., Zou, K.H., et al. (2011). Smoking, smoking cessation and smoking relapse patterns: a web-based survey of current and former smokers in the US. International Journal of Clinical Practice, 65(10), 1043–1054. http://dx.doi.org/10.1111/j.1742-1241.2011.02758.x. Zhou, X., Nonnemaker, J., Sherrill, B., Gilsenan, A.W., Coste, F., & West, R. (2009). Attempts to quit smoking and relapse: factors associated with success or failure from the ATTEMPT cohort study. [Multicenter study]. Addictive Behaviors, 34(4), 365–373. http://dx.doi.org/10.1016/j.addbeh.2008.11.013.

Zvolensky, M.J., Bonn-Miller, M., Bernstein, A., & Marshall, E. (2006). Anxiety sensitivity and abstinence duration to smoking. Journal of Mental Health, 15, 659–670. http:// dx.doi.org/10.1080/09638230600998888. Zvolensky, M.J., Gibson, L.E., Vujanovic, A.A., Gregor, K., Bernstein, A., Kahler, C., et al. (2008). Impact of posttraumatic stress disorder on early smoking lapse and relapse during a self-guided quit attempt among community-recruited daily smokers. Nicotine & Tobacco Research, 10(8), 1415–1427. http://dx.doi.org/10. 1080/14622200802238951.