Citation
Trends and Predictors of Cigarette Smoking and Its Association on Cognitive Performance among HIV Seropositive and Seronegative Men

Material Information

Title:
Trends and Predictors of Cigarette Smoking and Its Association on Cognitive Performance among HIV Seropositive and Seronegative Men
Creator:
Akhtar, Wajiha Z
Place of Publication:
[Gainesville, Fla.]
Florida
Publisher:
University of Florida
Publication Date:
Language:
english
Physical Description:
1 online resource (124 p.)

Thesis/Dissertation Information

Degree:
Doctorate ( Ph.D.)
Degree Grantor:
University of Florida
Degree Disciplines:
Epidemiology
Committee Chair:
COOK,ROBERT L
Committee Co-Chair:
BEYTH,REBECCA J
Committee Members:
MANINI,TODD M
PRICE,CATHERINE ELIZABETH
PLANKEY,MICHAEL
Graduation Date:
12/18/2015

Subjects

Subjects / Keywords:
AIDS ( jstor )
Cigarette smoking ( jstor )
Cohort studies ( jstor )
Disease risks ( jstor )
Diseases ( jstor )
Hispanics ( jstor )
HIV ( jstor )
Self reports ( jstor )
Tobacco smoking ( jstor )
Trajectories ( jstor )
Epidemiology -- Dissertations, Academic -- UF
epidemiology -- hiv -- msm -- smoking
Genre:
bibliography ( marcgt )
theses ( marcgt )
government publication (state, provincial, terriorial, dependent) ( marcgt )
born-digital ( sobekcm )
Electronic Thesis or Dissertation
Epidemiology thesis, Ph.D.

Notes

Abstract:
Men who have sex (MSM) with men account for the largest proportion of people living with HIV (PLWH). In this current era of antiretroviral therapy use, lifestyle factors may be a greater threat than the virus to long-term survival and quality of life in MSM and others living with HIV. The prevalence of current cigarette smoking among PLWH has been estimated to be approximately 35% to 70%, more than twice the rate of the general US population. It is unclear how smoking behavior in MSM has changed in the last 25 years. Studies have shown that smoking can increase the risk of HIV-associated infections, cardiovascular disease, and immunologic dysfunction. Emerging evidence suggests that smoking may be associated with cognitive decline in older adults in the general population. Little is known on the effects of smoking on cognitive decline in PLWH. Using data from the Multicenter AIDS Cohort Study, we characterized the annual prevalence of smoking in MSM and evaluate the trend by HIV serostatus. We described the demographic characteristics associated with long-term cigarette smoking behavior among PLWH using a trajectory approach. Last, we longitudinally assessed the association of smoking on cognitive performance among HIV-seropositive and HIV-seronegative MSM. This study addressed a critical gap in the literature by understanding long-term smoking behavior and its effects on cognitive decline. A comprehensive understanding of important modifiable factors that have a greater threat to long-term survival and quality of life in PLWH will enable us to have better screening and treatment guidelines for this subpopulation. Additionally, public health messages on smoking can be better targeted based on our results. ( en )
General Note:
In the series University of Florida Digital Collections.
General Note:
Includes vita.
Bibliography:
Includes bibliographical references.
Source of Description:
Description based on online resource; title from PDF title page.
Source of Description:
This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Thesis:
Thesis (Ph.D.)--University of Florida, 2015.
Local:
Adviser: COOK,ROBERT L.
Local:
Co-adviser: BEYTH,REBECCA J.
Electronic Access:
RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2016-12-31
Statement of Responsibility:
by Wajiha Z Akhtar.

Record Information

Source Institution:
UFRGP
Rights Management:
Applicable rights reserved.
Embargo Date:
12/31/2016
Classification:
LD1780 2015 ( lcc )

Downloads

This item has the following downloads:


Full Text

PAGE 1

TRENDS AND PREDICTORS OF CIGARETTE SMOKING AND ITS ASSOCIATION ON COGNITIVE PERFORMANCE AMONG HIV SEROPOSITIVE AND SERONEGATIVE MEN By WAJIHA ZEENAT AKHTAR A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2015

PAGE 2

© 2015 Wajiha Zeenat Akhtar

PAGE 3

To Ami and Abu

PAGE 4

4 ACKNOWLEDGMENTS I thank Dr. Elena An dresen for bringing me to the University of Florida and mentoring me both personally and professionally. I thank Dr. Bob Cook, for blindly taking over when Dr. Andresen left, without knowing what he was going to get himself into. I thank him for his persis tence, dedication, and helping me get to where I am today. I thank Dr. Michael Plankey for his countless hours of help, guidance, and mentorship. He selflessly devoted his time and effort solely for the advancement of epidemiology and HIV/AIDS. I thank the rest of my dissertation committee, Drs. Rebecca Beyth, Todd Manini, and Catherine Price, who pushed me to become a better researcher and writer. I also thank the Department of Epidemiology for their continued support throughout the years. Data for this s tudy were collected by the Multicenter AIDS Cohort Study (MACS) with centers at Baltimore (U01 AI35042) ; Chicago (U01 AI35039) ; Los Angeles (U01 AI35040) ; and Pittsburgh (U01 AI35041) . The MACS is funded primarily by the National Institute of Allergy and I nfectious Diseases (NIAID), with additional co funding from the National Cancer Institute (NCI), the National Institute on Drug Abuse (NIDA), and the National Institute of Mental Health (NIMH). I thank Janet Schollenberger and the rest of the staff at the Data Coordinating Center for their endless help in obtaining and updating the dataset for this study. I would also like to thank the participants of the MACS who have devoted so much of their time for so many years. I thank my parents , Shameem and Zeenat Akhtar, who sacrificed everything to ensure that my brother and I can have everything . I thank my brother , Tariq Syed, who has always been there f or m e . would have been without him. I thank my in la ws , Ibrahim and Saleha Khaleel, for their

PAGE 5

5 longtime support, and my brother in laws , Safi, Omar, and Ali , for their encouragement. I thank Sajid Bhai, Faheema Baji, Adam, Aaliya, Aisha, and Ibraheem, who took me in and treated me like family while living in Gainesville, FL. I would also like to thank my husband, Awais. Because of you, I am here, writing this dissertation. Your passion for your work and family motivated me to understand what was important to me and drove me to continue my research. Your cont inued support and encouragement kept me going. Finally, I thank our adorable son, Idris. The excitement I see in your eyes and how eager you are to take on this world, drives me to learn and teach you. (no pressure) .

PAGE 6

6 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ .. 4 LIST OF TABLES ................................ ................................ ................................ ............ 8 LIST OF FIGURES ................................ ................................ ................................ .......... 9 LIST OF AB BREVIATIONS ................................ ................................ ........................... 10 ABSTRACT ................................ ................................ ................................ ................... 12 CHAPTER 1 BACKGROUND AND SIGNIFICANCE ................................ ................................ ... 14 Smoking and HIV ................................ ................................ ................................ .... 14 Cognitive Decline and HIV ................................ ................................ ...................... 19 Smoking, Cognitive Decline, and HIV ................................ ................................ ..... 23 Study Population ................................ ................................ ................................ ..... 26 2 TRENDS AND PREDICTORS OF CIGARETTE SMOKING AMONG HIV SEROPOSITIVE AND SERONEGATIVE MEN: THE MULTICENTER AIDS COHORT STUDY ................................ ................................ ................................ ... 40 Introduction ................................ ................................ ................................ ............. 40 Methods ................................ ................................ ................................ .................. 42 Study Design and Administration ................................ ................................ ..... 42 Main Outcome Measure ................................ ................................ ................... 42 Independent Variables ................................ ................................ ..................... 43 Data Analysis ................................ ................................ ................................ ... 45 Results ................................ ................................ ................................ .................... 46 Discussion ................................ ................................ ................................ .............. 50 3 LONG TERM CIGARETTE SMOKING TRAJECTORIES AMONG HIV SEROPOSITIVE AND SER ONEGATIVE MSM IN THE MULTICENTER AIDS COHORT STUDY ................................ ................................ ................................ ... 60 Introduction ................................ ................................ ................................ ............. 60 Methods ................................ ................................ ................................ .................. 62 Study Population ................................ ................................ .............................. 62 Main Outcome Measure ................................ ................................ ................... 63 Covariates of Interest ................................ ................................ ....................... 63 Data Analysis ................................ ................................ ................................ ... 64 Results ................................ ................................ ................................ .................... 66 Discussion ................................ ................................ ................................ .............. 68

PAGE 7

7 4 THE AS SOCIATION OF MID LIFE SMOKING STATUS ON PROCESSING SPEED AND MENTAL FLEXIBILITY AMONG HIV SEROPOSITIVE AND SERONEGATIVE OLDER MEN: THE MULTICENTER AIDS COHORT STUDY ... 79 Introduction ................................ ................................ ................................ ............. 79 Methods ................................ ................................ ................................ .................. 80 Study Population ................................ ................................ .............................. 80 Exposure of Interest ................................ ................................ ......................... 81 Outcome of Interest ................................ ................................ .......................... 82 Statistical Analysis ................................ ................................ ............................ 86 Results ................................ ................................ ................................ .................... 87 Reliable Change Index ................................ ................................ ..................... 87 Linear Mixed Models ................................ ................................ ........................ 88 Discussion ................................ ................................ ................................ .............. 89 Comparison with other studies ................................ ................................ ......... 89 Mechanisms ................................ ................................ ................................ ..... 90 Survivor Effects ................................ ................................ ................................ 91 Limitations ................................ ................................ ................................ ........ 92 5 CONCLUSIONS ................................ ................................ ................................ ... 106 Accomplishments of the Dissertation ................................ ................................ .... 106 Smoking Cessation in PLWH ................................ ................................ ................ 108 Resiliency ................................ ................................ ................................ ............. 108 Future Directions ................................ ................................ ................................ .. 108 Screening Tests for Cognitive Decline ................................ ........................... 108 The MACS Study ................................ ................................ ............................ 109 LITERATURE REVIEW ................................ ................................ ............................... 111 LIST OF REFERENCES ................................ ................................ ............................. 116 BIOGRAPHICAL SKETCH ................................ ................................ .......................... 124

PAGE 8

8 LIST OF TABLES Table page 1 1 Neuropsychological assessments in the MACS ................................ ................. 35 2 1 Characteristics of the MACS population at time of enrollment. ........................... 55 2 2 Prevalence ratios for cigarette smoking f rom univariate and multivariate analysis ................................ ................................ ................................ .............. 57 2 3 Prevalence ratios for smoking from univariate and multivariate an alysis among HIV seropositive participants ................................ ................................ .. 58 2 4 Prevalence ratios for smoking from univariate and multivariate analysis among HIV seropositive participants ................................ ................................ .. 59 3 1 Descriptive statistics of baseline covariates by trajectory group and HIV serostatus ................................ ................................ ................................ ........... 71 3 2 Estimate groups and group specific growth parameters for all participants in the MACS ................................ ................................ ................................ ........... 74 3 3 Factors associated with trajectory group membership, after adjusting for time varying covariates for all participants ................................ .......................... 75 3 4 Estimated trajectory groups and group specific growth parameters for HIV seropositive participants in the MACS ................................ ................................ 77 3 5 Factors associated with trajectory group m embership, before adjusting for time varying covariates, HIV seropositive participants ................................ ....... 78 4 1 Characteristics of the MACS Cohort Study as a function of smoking status at age 50 ................................ ................................ ................................ ................ 98 4 2 Neuropsychological raw test scores and standardized reliable change index scores by smoking status at age 50 and visit ................................ ................... 101 4 3 Associ ation of smoking history (age 50) and cognitive change over the subsequent 5 years among all participants (n=591) ................................ ......... 102 4 4 Association of smoking history (age 50) and cognitive change over the subsequent 5 years among HIV positive participants (n=220) ......................... 103 4 5 Association of cumulative pack years (at age 50) and cognitive change over the subsequent 5 years among all participants (n= 591) ................................ ... 104 4 6 Association of cumulative pack years (age 50) and cognitive change over the subsequent 5 years among HIV positive participants (n=220) ......................... 105

PAGE 9

9 LIST OF FIGURES Figure page 1 1 Factors associated with smoking on cognitive performance ............................... 36 1 2 Example of Trail Making Test A ................................ ................................ .......... 37 1 3 Example of Trail Making Test Part B. ................................ ................................ . 38 1 4 Example of Symbol Di git Modalities Test. ................................ .......................... 39 2 1 Change in smoking prevalence over time. ................................ .......................... 56 3 1 Trajectory groups of smoking consumption over time with adjustment for time constant and time varying variables in participants in the MACS ............... 73 3 2 Trajectory groups of smoking consumption over time with adjustment for time constant and time varying variables in HIV seropositive participants ......... 76 4 1 Example of Trail Making Test Part A ................................ ................................ .. 93 4 2 Example of Trail Making Test Part B. ................................ ................................ . 94 4 3 Example of Symbol Digit Moda lities Test. ................................ .......................... 95 4 4 Distribution of Trail Making A and B raw and log transformed scores at baseline (age=50) ................................ ................................ ............................... 96 4 5 Distribution of Symbol Digit Modalities raw and log transformed scores at baseline (age=50) ................................ ................................ ............................... 97 4 6 B Spline of cognitive performance test z scores over time ................................ . 99 4 7 Change in composite z scor e over time by smoking status at age 50 .............. 100

PAGE 10

10 LIST OF ABBREVIATIONS AD ADL Activities of Daily Living AIDS Acquired Immunodeficiency Syndrome ANI Asymptomatic neurocognitive impairment ART Antire troviral therapy BMI Body Mass Index CD4+ Cluster of differentiation 4 CESD Center for Epidemiological Studies Depression CHF Congestive heart failure CI Confidence interval CVD Cardiovascular disease CNS Central Nervous System HAD HIV as sociated dementia HAND HIV associated neurocognitive disorder HCV Hepatitis C Virus HIV Human immunodeficiency virus HR Hazards Ratio IQR Interquartile range IRB Institutional review board MACS Multicenter AIDS Cohort Study MI Myocardial Inf arction MND Mild neurocognitive disorder MSM Men who have sex with men OR Odds Ratio

PAGE 11

11 PLWH People living with HIV SD Standard Deviation VACS Veterans Aging Cohort WIHS

PAGE 12

12 Abstract of Dissertat ion Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy TRENDS AND PREDICTORS OF CIGARETTE SMOKING AND ITS ASSOCIATION ON COGNITIVE PERFORMANCE AMO NG HIV SEROPOSITIVE AND SERONEGATIVE MEN By Wajiha Zeenat Akhtar December 2015 Chair: Robert L Cook Major: Epidemiology Men who have sex with men (MSM) account for the largest proportion of people living with HIV (PLWH). In this current era of antire troviral therapy use, lifestyle factors may be a greater threat than the virus to long term survival and quality of life in MSM and others living with HIV. The prevalence of current cigarette smoking among PLWH has been estimated to be approximately 35% 70 %, more than twice the rate of the general US population. It is unclear how smoking behavior in MSM has changed in the last 25 years. Studies have shown that smoking can increase the risk of HIV associated infections, cardiovascular disease , and immun ologi c dys function. Emerging evidence suggests that smoking may be associated with cognitive decline in older adults in the general population. Little is known on the effects of smoking on cognitive decline in PLWH. Using data from the Multicenter AIDS Cohort Study, we characterized the annual prevalence of smoking in MSM and evaluate the trend by HIV serostatus. We describe d the demographic characteristics associated with long term cigarette smoking behavior among PLWH using a trajectory approach. Last, we lon gitudinally assess ed the

PAGE 13

13 association of smokin g on cognitive performance among HIV sero positive and HIV seronegative MSM. This study addressed a critical gap in the literature by understanding long term smoking behavior and its effects on cognitive decline . A comprehensive understanding of important modifiable factors that have a greater threat to long term survival and quality of life in PLWH will enable us to have better screening and treatment guidelines for this subpopulation. Additionally, public healt h messages on smoking can be better targeted based on our results.

PAGE 14

14 CHAPTER 1 BACKGROUND AND SIGNIFICANCE Since the introduction of antiretrovirals (ART) , HIV positive individuals are living longer than before (Justice, 2009). In 20 11 , people living with HIV (PLWH) over the age of 50 accounted for 2 6 % (313,200) of the population of PLWH (CDC, 20 15 ). Among newly diagnosed cases, older Americans are more likely than younger Americans to be diagnosed with HIV later in the course (CDC, 2015). As the epidemic r eaches its fourth decade, more information is needed on how the virus, therapy, behaviors associated with PLWH, and the natural aging process all interact with each other, in order to understand adverse health outcomes that PLWH may face. HIV and its treat ments affect the aging process or the development of morbidities that are associated with the aging process (High et al., 2012). Thus, more information is needed on how modifiable risk factors, such as smoking, can affect the aging process. Smoking and HIV The prevalence of current smokers among PLWH has been estimated to be over 40%, more than 20% greater than the general US population (Crothers et al., 2009; Lifson and Lando, 2012; Lifson et al., 2010; Tesoriero et al., 2010). PLWH who smoke, on average, smoke 16 23 cigarettes per day, an indicator for high level of nicotine HIV Study (WIHS) showed that 56% of the participants had reported that they were current smokers and h ad smoked more than a pack a day for 50% of their lives since adolescence (Feldman et al., 2006). Among MSM, estimates have ranged from 45% 49% (Royce and Winkelstein, 1990; Stall et al., 1999). Because of a history of exclusion and discrimination in other social settings, the social focus for many MSM has been

PAGE 15

15 gay identified bars and clubs, where the prevalence of smoking continues to be high (Reynolds, 2011). Past studies have established a few factors that are associated with smoking in PLWH. Factors inc lude race (African Americans), alcohol consumption, level of physical activity, Hepatitis C virus (HCV), substance use, obstructive lung disease, and cardiovascular disease (Helleberg et al., 2013; Crothers et al., 2009; Crothers et al., 2006; Crothers et al., 2005; Feldman et al., 2006). A known risk factor for cerebrovascular disease is smoking, which is more prevalent in older populations of both HIV seronegative and HIV seropositive individuals (Kilbourne et al., 2001; Valcour et al., 2004). Additionall y, lipid abnormalities in HIV infected individuals may add to atherosclerosis and an increased risk of cerebrovascular disease (Valcour et al., 2004). Therefore, studying the effects of smoking on cognitive decline is imperative in this population. Studies assessing smoking in PLWH support research from the general population showing that smoking is a risk factor for coronary artery disease, myocardial infarction, and stroke (Barbaro et al., 2003; Friis Møller et al., 2003; Lifson and Lando, 2012; Lifson et al., 2010; Petoumenos et al., 2011). Petoumenos et al. found that in those who stopped smoking during follow up, compared with those who had never smoked, the odds ratio for the risk of cardiovascular disease (CVD) decreased from 2.3 within the first year of stopping to 1.5 after more than three years. Grunfeld et al. (2009) used a multivariate analysis to demonstrate that both HIV and smoking were independent risk factors for greater carotid intima media thickness, which may contribute to preclinical athe rosclerosis and CVD events. Smoking can also

PAGE 16

16 increase the risk of altering cellular and humoral immune system function in PLWH. immune and inflammatory response to infect ious agents (Arcavi and Benowitz, 2004). Implementing tobacco cessation programs has been challenging in PLWH. Crothers et al. (2007) found that HIV care providers in the Veterans Aging Cohort 5 Site Study are less likely to recognize current smokers compa red with non HIV care providers. However, providers that recognize current smokers are more likely to make smoking treatment a low priority because of other competing priorities, economic barriers, or limited time for health promotion activities (Reynolds, 2009). Smoking can be measured by self report or biochemical tests that detect carbon monoxide, thiocyanate, or cotinine in saliva, blood, urine, or expired air (Jarvis et al., 1987). Appendix A shows that all studies, to date, assessing the association o f smoking on cognitive decline have used self report for their exposure variable, while cotinine measures were used to examine secondhand smoke exposure in self reported nonsmokers. Self report questionnaires consist of a set of questions designed to under stand t, former, and never smokers with an additional question that asks about current smoking

PAGE 17

17 reported duration, pack years can be calculated by multiplying packs smoked per day by years smoked. Major advantages of using a self report questionnaire are that it is the lea st expensive type of measurement and the responses can be immediately assessed and analyzed. Various exposure questions enable the researcher to evaluate the patterns of cumulative and passive smoking. Limitations in self reported smoking include participa unwillingness to disclose information about their tobacco use. But studies suggest that self 0.86) when compared to biochemical measures (So ulakova et al., 2012; Murray et al., 2002). The majority of studies in Appendix A dichotomized participants as either smokers or nonsmokers. Other studies categorized them as current, former, or never smokers. Measuring smoking this way may group people to gether that are dissimilar to each other. Categorization does not capture duration or intensity of smoking throughout years does quantify duration and intensity; however, it does not capture the fluctuations in lifetime smoking that can be observed in populations trying to quit or decrease cigarette smoking. For example, studies assessing smoking in PLWH have categorized smoking as current, former, and never smokers and have also analyzed it by using cumulativ e pack years separately (Levine A et al., 2010; Crothers et al., 2009; Crothers et al., 2006). Cumulative pack interquartile range (IQR) for pack years of current smokers ran ged from 3.8 to 16.5. In

PAGE 18

18 to 30 and 5 to 30, respectively. In another study by Crothers (2009), the IQR for current smokers and former smokers ranged from 7 to 30 and 5 to 28, respectively. Although these studies have categorized participants based on smoking status, their pack years do not differ from one another. To date, only one study has assessed smoking on cognitive decline using longitudinal data in the general popul ation and no studies have done so in PLWH (Sobia et al., 2012). In Sobia et al., participants were first categorized as current, recent ex , long term ex , or never smokers. At follow up, smokers were further categorized as persistent smokers (continued sm oking), intermittent smokers (quitters who started smoking again), and quitters (stopped smoking). Studies that have assessed cigarette smoking on other outcomes have also discussed the limitation of using pack years (Lubin and Caporaso, 2006; Pandeya et a l., 2008). Lubin and Caporaso compared smokers who smoked at low intensity for a long period of time with participants who smoked at high intensity for a short period of time. They found that smoking at lower intensity for a longer duration was associated with a higher risk for lung cancer compared with smokers at a higher intensity for a shorter duration. In order to group participants based on their smoking behavior longitudinally, we measured smoking using group based trajectories. Many studies have used group based trajectories to measure alcohol consumption in the general population and in PLWH (Platt et al., 2010; Cook et al., 2012; Jacob et al.; 2013). We utilized semi parametric, group based logistic models, which will be explained in detail later on , to identify smoking trajectories among participants within the dataset. A group based logistic model is a useful technique to assess change over time, and it has a number of

PAGE 19

19 advantages. First, the method can help one discover potentially meaningful traje ctories that could have been overlooked (Nagin, 2005). All available data can be utilized to estimate the trajectories, and the process can take into account missing values. There are a few limitations in using group based trajectories. Trajectory modeling does not account for past smoking information, but using established cumulative pack years would show that the relationship remains consistent. The procedure needs at least three time points for proper estimation and four to five for quadratic or cubic tr ends (Andruff and Louvet, 2009). Loss to follow up can also decrease the statistical precision and may introduce bias if the data are not missing at random or missing completely at random (Andruff and Louvet, 2009). Cognitive Decline and HIV Cognition can be thought of as a process to take in and understand a sensory input or knowledge. As time progresses, the human brain will decline in cognitive capacity and will be marked by a decrease in brain volume (Drag et al., 2009). Impairment in cognition in PLWH is known as HIV associated neurocognitive disorder (HAND). In PLWH, HIV first enters the central nervous system (CNS) during the acute infection, but the neurons remain uninfected (Valcour et al., 2011a). It is theorized that, once in the CNS, the virus es tablishes a reservoir that resists ART and causes extensive inflammation, which then leads to neuronal dysfunction and synaptodendritic injury (Valcour et al., 2011a). HIV staining has shown that the virus is concentrated in the subcortical deep gray matte r structures (Gabrieli et al., 1995; Gray et al., 2001; Woods et al., 2009). The effects of the virus are most prominent in the basal ganglia, frontal neocortex, and white matter tracts connecting these regions (Woods et al., 2009). Cognitive disorders ass ociated with HIV share features with diseases such as

PAGE 20

20 processes (bradyphrenia and bradykinesia) but can often recall information if given enough time (Kartikeyan et al., 2007). Specifically, patients with HIV who develop AIDS have reported experiencing memory deficits, slower information processing, motor deficits, and attention and concentration problems (Hardy et al., 2009). Patients with symptomatic stages of HIV have reported problems with attention, concentration, learning, memory, psychomotor speed, and speed of processing (Hardy et al., 2009). In everyday terms, PLWH who have the most severe HAND can experience increased confusion (e.g., being unaware of their illness, havi ng difficulty following directions); delayed or no verbal responses; and difficulty in generating ideas (e.g., not being able to communicate what they want). Other symptoms include difficulty in reading, losing track of conversations, handwriting deteriora tion, slowness in thinking, and stiffness in the legs (Janssen, 1991). As these symptoms become more severe, cognitive (ADL) , such as bathing, dressing, eating, and maintaining per sonal hygiene. It is important to note that even patients diagnosed with the mildest form of HAND are less likely to adhere to medication recommendations, struggle to perform complex daily tasks, and have a worse quality of life, difficulty in obtaining em ployment, and a shorter survival rate (Albert SM et al., 1999; Berger et al., 2005; Farinpour et al., 2003; Garvey et al., 2008; Garvey et al., 2009; Tozzi et al., 2007). As PLWH begin to age, there could be a rise in HAND because of the interactive effect s of immune function and aging on the CNS. As many as 40% of HIV seropositive individuals suffer from HAND (Sacktor et al., 2001;

PAGE 21

21 Lindl et al., 2010). Conservative estimates suggest that HAND diagnoses will increase five to tenfold by 2030 (Lindl et al., 2010). Several studies have assessed possible causal factors that are associated with cognitive decline in PLWH (Valcour et al., 2004; Heaton et al., 2011; Wright et al., 2010; Heaton et al., 2010; Becker et al., 2011; Becker et al., 2009; Fabbiani et al., 2012). One model for cognitive decline in HIV infected older adults posits five different causal factors (Valcour et al., 2004): (1) synergistic effects of HIV and neuropathology related to aging; (2) hypothyroidism/vitamin B12 deficiency and other metabo lic conditions; (3) changes in immune function; (4) neurological changes that are associated with aging; and (5) cerebrovascular disease. More recent studies have included factors such as age, current and nadir CD4 count, race/ethnicity, education, BMI, sm oking, alcohol abuse, medication adherence, glomerular filtration rate (GFR), recreational drug abuse, Hepatitis B, Hepatitis C, and CVD (Heaton et al., 2011; Wright et al., 2010; Heaton et al., 2010; Becker et al., 2011; Becker et al., 2009; Fabbiani et a l., 2012). Cerebrovascular and cardiovascular copathology is even more important because of the many factors that can increase it in PLWH. ART use, for example, that includes a protease inhibitor is known to increase the risk of cardiovascular and cerebrov ascular disease (Hardy and Vance, 2009). The current gold standard for diagnosing HAND was designed using neuropsychology testing batteries. These tests result in patients with HAND being assigned to one of three levels of impairment. HIV associated dement ia (HAD) is the most severe form of impairment, requiring test performance to be greater than two standard deviations below the mean in two cognitive domains with evidence of impaired

PAGE 22

22 daily function (Valcour et al., 2011b). Mild neurocognitive disorder (MN D) requires less severe impaired daily function and impairment in cognitive domains (Valcour et al., 2011b). Last, asymptomatic neurocognitive impairment (ANI) requires test performance to be greater than one standard deviation below the mean in two cognit ive domains with an absence of identifiable functional deficits (Valcour et al., 2011b). Past studies have used at least three different methods to measure cognitive decline longitudinally. First, participants often complete a comprehensive neurocognitive test battery encompassing different domains of cognitive function. Some studies use the best available normative standards to correct testing scores for age, education, sex, and ethnicity. The test scores are then converted to demographically corrected sta ndard scores. Classification of neurocognitive impairment can be calculated using a published objective algorithm (Woods et al., 2004). Second, other studies standardize each score into z scores by subtracting the mean test scores of matched HIV seronegati ve reference populations, or their own baseline groups, and then dividing that by the standard deviation. Some studies define impairment as z using z t et al., 2010; Heaton et al., 2011). Z scores allow for comparability across tests and can be used to build composite scores. But composite scores become problematic when they are compared to other studies because authors use different tests to make up th eir composite scores. Finally, studies have also used differences in raw scores to determine the aggregate rate of change for each participant (Sabia et al., 2012; Vo et al., 2013). We use d this method because it comes with the ability to use all available data, handle

PAGE 23

23 different lengths of follow up, and to take into account that repeated measures for each participant are correlated. The intercept and slope will be assessed as random effects. This will allow for individual differences in cognitive performan ce at baseline and the rate of decline. Although this method is less clinically meaningful compared with the other methods above, we can still c ompare change in decline and without a cutoff that has differed across studies. Smoking, Cognitive Decline, and HIV Considering the information provided in the first two sections, it is understood that smoking is a risk factor for many vascular diseases, such as atherosclerosis and thrombosis, which may increase the risk of cerebrovascular diseases and vascular deme ntia (Ott et al., 1998). Furthermore, recent studies have shown that cognitive changes in older PLWH are likely to include a cerebrovascular pathology (Sacktor et al., 2010; Valcour et al., 2004). Figure 1 1 displays factors that may affect smoking and cog nitive decline in PLWH. Based on the literature stated earlier, factors such as age, race, education, and time of enrollment are associated with both smoking and cognitive decline thus, they may act as confounding variables. Smoking is also associated with alcohol use, substance abuse, and HCV. The relationship of these factors on cognitive decline has not been established in HIV seropositive adults. Smoking also causes comorbidities and other factors that may increase the risk of cognitive decline, thus, t hese factors may act as intermediate variables. These include hypertension, cardiovascular disease, CD4 nadir, inflammation, and cerebrovascular disease. Last, studies have shown that ART use and depression are associated with cognitive decline, but their relationships with smoking have not yet been established.

PAGE 24

24 disease on the general population has yielded inconsistent results (Almeida et al., 2002). The pooled effects of the case control studies show a protective effect (OR=0.74, 95%CI: 0.66, 0.84) while cohort studies show an association (RR=1.99, 95%CI: 1.33, 2.98). A more recent review assesses 19 studies with at least 12 months of follow up (Anstey et al., 2007). Outcome measu disease, vascular dementia, any dementia, cognitive performance at follow up, cognitive performance change, cognitive decline, and cognitive impairment. The pooled effects of current smokers to never smokers on A a relative risk of 1.79 (95%CI: 1.43, 2.23) and 1.78 (95%CI: 1.28,2.47), respectively. The relationship remains consistent when comparing current smokers with former and never smokers on cognitive decline (RR=1 .41 95%CI: 1.16, 1.71). A review from Durazzo, Meyerhoff, and Nixon compiles peer reviewed studies assessing the consequences of chronic cigarette smoking on neurocognitive or neurobiological performance. Their review showed that smoking appears to be asso ciated with poor executive function, cognitive flexibility, general intellectual abilities, processing speed, and working memory (Durazzo et al., 2010). A more recent study utilizes six assessments of smoking status over 25 years to establish an associatio n with cognitive decline in the general population (Sabia et al., 2012). To our knowledge, this is the first study that has established longitudinal smoking patterns using follow up data. The study found that men who continued smoking over follow up were m ore likely to experience greater decline in cognitive tests for memory, vocabulary, and executive functioning (Sabia et al., 2012).

PAGE 25

25 The relationship between smoking and cognitive decline in PLWH remains unclear. To date, three studies have assessed the ass ociation of smoking on cognitive performance in PLWH. Durazzo et al. cross sectionally assessed 44 HIV seropositive alcohol drinkers and found that smokers were more likely to perform worse than non smokers on tests for auditory verbal learning, auditory v erbal memory, and cognitive efficiency (Durazzo et al., 2007). Wojna et al. performed a cross sectional study of 56 participants and found no statistically significant associations between cognitive performance and current smoking or past smoking history i n PLWH (Wojna et al., current smokers and learning, memory, and global cognitive functioning. After adjusting for education and HCV, the association was no longer statistically significant (Bryant et al., 2013). With the literature that is currently available for the assessment of smoking and cognitive decline, we aim to address a few scientific gaps. First, there is a need to better assess smoking exposure in participants longi tudinally in order to understand changes in smoking behavior. This will be the first study to investigate the relationship between smoking and long term cognitive decline, using a large sample of participants with repeated measures and a longitudinal resea rch study design. We will then use this measure to longitudinally assess cognitive function in PLWH again, the first study of its kind in this subpopulation. Using repeated measures, we can assess cumulative changes in both exposure and outcome and can det ermine causality. As PLWH begin to age, more information is needed on how behaviors such as smoking, which is more prevalent in PLWH will affect how they age. Studies assessing

PAGE 26

26 smoking and cognitive performance have been focused on the HIV seronegative po pulation. The MACS study described in detail later on is an ideal cohort to test this hypothesis. The study, which encompasses a diverse group of men who are assessed longitudinally, gives us the opportunity to analyze in more novel ways. The studies propo sed are expected to have a significant impact for PLWH. Understanding the concepts involved in HIV and aging can improve clinical management for PLWH. HIV increases the risk and progression of many common infectious and noninfectious non AIDS conditions. A full understanding of the etiology of these conditions can enable better screening and treatment guidelines of non AIDS conditions for PLWH. More information can justify earlier or more aggressive antiretroviral treatment. Mateen and Mills have urged rese archers to seek strategies in addressing issues related to HIV and aging related cognitive decline. This includes testing reversible causes of dementia (Mateen and Mills, 2012). If an association does exist, more aggressive smoking cessation strategies fo r PLWH will be needed to improve future cognitive performance. While filling these gaps, the studies proposed will also provide a major public health impact on this subpopulation. First, along with other emerging studies on the detrimental effects of smoki ng on morbidity and mortality in PLWH, guidelines for treatment and cessation strategies can be recommended to clinicians that offer care to PLWH. Additionally, this research can build on the growing research on the mechanism for aging in PLWH. Study Popul ation The Multicenter AIDS Cohort Study (MACS) has been described in detail elsewhere (Kaslow et al., 1987). It is an ongoing longitudinal study of the natural and

PAGE 27

27 treated history of HIV infection and AIDS among MSM. The study identified risk factors for i ncidence and clinical expression of the infection (Kaslow et al., 1987). Since 1984, a total of 6,973 men have been recruited at four centers: the Baltimore and Washington, DC, area, Chicago, Los Angeles, and Pittsburgh. Recruitment strategies differ among sites. Initially, in Baltimore and Washington, DC, participants were recruited through media publicity and personal communications between the investigators and gay activists. In Los Angeles, participants were recruited from pre existing AIDS study cohort s and organizations. In Pittsburgh, investigators recruited from gay bars and bathhouses for screening studies, which led them to volunteer for the MACS. Demographic characteristics at each site were almost identical (Kaslow et al., 1987). Small difference s were observed in number of partners and infections among the sites. From April 1987 to December 1990, the MACS recruited 231 additional men to increase the ethnic and racial minorities. The cohort is ideal for the proposed study because of its repeated m easures and large sample size. MSM still remains the largest HIV exposure group in the United States, encompassing 53% of new HIV cases (Hall et al., 2008). Participants return ed biannually for a detailed interview, physical exam, and blood draw for labora longitudinal neuropsychological test performance to assess the effects of HIV on the brain and nervous system (Miller et al., 1990). The full battery is performed every two years, while the Trai l Making and Symbol Digit Modalities tests are administered every six months. The subgroup undergoing the neuropsychological assessments ha d been

PAGE 28

28 selected without systematic bias in an effort to represent the entire MACS cohort (McArthur et al., 1993). Cig arette Smoking . The MACS has collected data on cigarette consumption. Based on answers to a detailed interview, participants are classified as never, former, and current Amount of packs smoked is categorized by the MACS as follows: less than ½ pack per day; at least ½ but less than 1 pack per day; at least 1 but less than 2 packs per day; and 2 or more packs per day. We defined pack years based on the amount of cigarettes smoked per day. We calculated it by determining the average pack (based on the choices from above) and multiplying by 0.5. If a participant smoked 1 to 2 packs a day, then his current smoking exposure will be calculated as 1.5 x 0.5 years = 0.75 pack years. This measure d pack days for one year for that specific visit. The MACS also assessed the length of time participants had smoked before commencing the study. Participants were asked at what age they had begun smoking cigarette s and how many packs had they smoked during their heaviest smoking periods. If participants were former smokers, the men were asked for the number of years they had stopped smoking cigarettes. Using this information, the MACS calculated baseline cumulative smoking for each participant, and continued to add onto it while the participants remained in the study. Cognitive Performance . Measures for cognitive performance were assessed for the last specific aim. The assessments used in the MACS and the cognitive functions sensitive to impairment found in PLWH are shown in Table 1 1 .

PAGE 29

29 For the third study we used the Trail Making A and B and the Symbol Digit Modalities tests. These t ests were picked because of the large amount of repeated measures in the dataset and because past studies suggest that smoking is associated with poor mental flexibility and inhibitory abilities (Trail Making B) and processing speed (Trail Making A, Symbol Digit Modalities Test) (Durazzo et al., 2010). Below is more information on each tes t. Trail Making A (Figure 1 2). P articipants were given a sheet of paper with numbers (1 25) randomly shown on an 8 by 11 inch portrait oriented paper. Participants were instructed to, as rapidly as possible, connect the numbers in ascending order. If the y made an error, the error was quickly pointed out and the participant corrected and number, the tester waited exactly ten seconds before correcting the participant. The time and the total number of prompts given to the participant were recorded during each visit. The dependent variable used for the current study was time to completion. Trail Making B (Figure 1 3). Participants were given an 8x11 inch portrait oriented paper with numbers (1 13) and letters (A L). Similar to Part A, participants were required to connect a series of numbers and letters such that they alternated between the number s and letters (1 A 2 B 3 C, etc.). The rules for timing, scoring, and correcting errors are the same as Part A. The dependent variable used for the current study was time to completion. Symbol Digit Modalities (Figure 1 4). Participants were given a sheet of paper with a box on the top of the page. The box has a symbol with a corresponding number.

PAGE 30

30 Below the key are boxes with symbols -and the participant must fill in each box with the corresponding number. Participants are told to fill the boxes as quickly as they can and that when they get to the end of the first line to go on to the second line. If a mistake is made, they are asked to write over it. They are not to skip any boxes. Each participant was given 90 seconds to complete as many as they can. One p oint was given for each box with a correct number. The final dependent variable used was total correct in 90 seconds. Long term cognitive performance was measured two ways. First, the three tests were assessed separately as raw scores. Second, a composite score was created using all three tests described above by first standardizing the raw scores on each test to z scores (mean=0; SD=1) using the mean and SD at baseline in the entire cohort for each test. The z scores was then averaged to yield a composite score, seen to minimize problems due to measurement error. Covariates of Interest (Figure 1 1) . Participants recorded date of birth was used to assess age for each visit. The MACS measures race as follows: White, non Hispanic; White, Hispanic; Black, non Hispanic; Black, Hispanic; American Indian or Alaskan Native, Asian or Pacific Islander, Other, or Other Hispanic. Because of the small number of Hispanics, American Indian or Alaskan Native, Asian or Pacific Islander, Other, or Other Hispanic in the first wave, we grouped them together. Our study categorizes race as White, non Hispanic; Black, non Hispanic; and other. Education was assessed at the baseline of the study and during follow up. Education was categorized as 8th grade or less, 9th,10th, and 11th grades, 12th grade, at least one year of college but no degree, four years of college/received degree, some graduate

PAGE 31

31 work, and post graduate degree. Information from the baseline visit will be used and categorized as follows: less than high school; high s chool diploma; some college or more. Presence of ever having depression was defined as Center for Epidemiological Studies Depression (CES D) Scale score of 16 or more. The CES D scale is an instrument that records the frequency of psychological symptoms du ring the past week with a value of 0 amount of the time (3 the time (5 7 days per D scale and a score of 16 or more was used to determine probable cases of clinical depression. Alcohol drinks w as measured at each visit using self report. The question asked for the number of d rinks per day since last visit and the MACS categorized it as 1 2 drinks; 3 4 drinks; 5 6 drinks; 7 9 drinks; 10 or more drinks; None. For our analysis, participants were categorize d by, no drinks; 1 2 drinks; 3 or more drinks. Diabetes was assessed in tw o different ways. The first was using fasting glucose Respondents who self reported that they were diagnosed with diabetes were also marked as having diabetes. For each vis it, data from past visits were added in order to establish the number of times each participant had diabetes. Participants who reported marked as having diabetes.

PAGE 32

32 Hyperte nsion was measured using blood pressure measurements from each Like diabetes, data from past visits were added in order to establish the number of time each participan t had hypertension throughout the study. Duration of hypertension before the age of 50 was used as a continuous variable. Self reported angina, heart attack, and congestive heart failure was measured in the questionnaire section for each visit. Participant Comorbidities listed included : 1. Angina or chest pain caused by your heart . 2. Heart attack or myocardial infarction (MI) . 3. Congestive Heart Failure ( CHF ). Participants were noted as having an angina, heart attack, or congestive heart if they reported having any of the above at any time after before age of 50. Because many participants began the study without initiating ART, and many participants changed their ART regimen during th e course of the MACS duration of ART was used in the analysis of the study. Self reported stroke was measured in the questionnaire section for each visit. idities listed included the following: had stroke , had mini stroke or transient ischemic attack . If a participant answered yes to one of these questions before the age of 50, the participant will then be classified as having a stroke.

PAGE 33

33 Self reported HCV was measured in the questionnaire section for each visit. participants were then class ified as having HCV. Self reported cancer was measured in the questionnaire section for each visit. d yes, the participants where then classifies as having cancer. Cumulative pack years. For every visit, the MACS categorized amount of packs smoked as follows: less than ½ pack a day ; at least ½ pack, but less than one pack per day; a t least 1 but less th an 2 packs; 2 or more packs per day. We summarized cumulative pack years based on the amount of cigarettes smoked per day. This can be calculated by using the average pack based on the above question and multiplying by 0.5. If a participant smoked 1 to 2 p acks a day, then their current smoking exposure would be calculated as 1.5 x 0.5 years = 0.75 pack years. This measures pack days for one year. A count variable was used to add each pack year for each visit. Cumulative pack years before baseline was assess ed by asking each participant their age when they began smoking cigarettes and the number of packs they smoked most during their

PAGE 34

34 baseline visit. Multiplying the number of years by the amount they smoked most was added to the baseline cumulative pack year t otal. IRB Review . Written informed consent was obtained from all subjects participating in the study. The MACS protocols were approved by institutional review boards (IRB) for e MACS behavioral working group has also reviewed and approved this proposed study.

PAGE 35

35 Table 1 1 . Neuropsychological a ssessments in the MACS Cognitive Domain MACS Neuropsychological Assessment Psychomotor Speed Trail Making A* Mental Flexibility/Inhibi tory Trail Making B* Psychomotor Speed Symbol Digit Modalities Test* Fine Motor Coordination Grooved Pegboard Test, dominant and non dominant hand Verbal Memory, Learning, and Long Term Recall Rey Auditory Verbal Learning Test (RAVLT)* Visuoconstructi on, Visual Memory Rey Osterrieth complex figure, copy, delayed recall Attention and Working Memory CalCap Reaction Test *Tests that will be used in Specific Aims 3

PAGE 36

36 ART Use Depressed Mood BMI Smoking Cognitive Decline Race Edu cation Age Time of Enrollment Hypertension Cardiovascular Disease Inflammation CD4 Nadir Cerebrovascular Disease Persistent Inflammation Alcohol Use Substance Abuse HCV Figure 1 1 . Factors associa ted with smoking on cognitive performance

PAGE 37

37 Figure 1 2 . Example of Trail Making Test A

PAGE 38

38 Figure 1 3 . Example of Trail Making Test Part B.

PAGE 39

39 Figure 1 4 . Example of Symbol Digit Modalities Test.

PAGE 40

40 CHAPTER 2 TRENDS AND PREDICTORS OF CIGA RETTE SMOKING AMONG HIV SEROPOSITIVE AND SERONEGATIVE MEN: THE MULTICENTER AIDS COHORT STUDY Introduction Research suggests that the prevalence of smoking in men who have sex with men (MSM) is higher than men in the general population ( Skinner W, 1994; Sk inner W and Drug O, 1996; Royce R and Winkelstein W, 1990; Lifson A et al., 2010; Crothers K et al., 2009 ). This group has the largest proportion of people living with HIV (PLWH). Because of a history of exclusion and discrimination in other social settin gs, the social focus for MSM has been gay identified bars and clubs, where the prevalence of smoking continues to be high ( Reynolds N, 2009 ). Recently, two studies focusing on HIV seropositive women and indigent adults reported the prevalence of current sm oking to be 39% and 67.3%, respectively ( Hessol NA et al., 2014; Vijayaraghavan M et al., 2014 ). Further, recent studies have assessed smoking in MSM but have not assessed the overall changing trend ( Robinson WT et al., 2011; Ompand DC et al., 2014 ). Studi es assessing smoking in PLWH support research from the general population showing that smoking is a risk factor for coronary artery disease, myocardial infarction, and stroke ( Lifson A et al., 2010; Crothers K et al., 2009 , Lifson AR and Lando AH, 2012; Te soriero JM et al., 2010 ). Smoking is also the number one reason for non AIDS defining cancers in PLWH ( Worm SW et al., 2014 ). Implementing tobacco cessation programs has been challenging in PLWH. Crothers et al. (2007) found that HIV care providers in the Veterans Aging Cohort 5 Site Study are less likely to recognize current smoking as a problem compared with non HIV care providers ( Crothers K et al., 2007 ). Furthermore, providers who identify current smokers are less

PAGE 41

41 likely to make smoking treatment a pri ority because of other competing health concerns, economic barriers, or limited time for health promotion activities ( Reynolds N, 2009; Crothers K et al., 2007 ). Although the overall prevalence of smoking has decreased in the general population ( Centers f or Disease Control and Prevention, 2014 ), it is unclear whether this trend also holds among HIV seropositive and negative MSM. Compared to the general population, HIV seropositive and seronegative MSM may have higher rates of other addictive behaviors such as alcohol use and drug use that are likely to increase the risk of cigarette smoke ( Mckirnan DJ et al., 2006 ). More importantly, if different trends exist in this subpopulation, then culturally tailored public health messages may be needed to promote hea lth behavior change interventions among HIV seropositive and negative MSM. The Multicenter AIDS Cohort Study (MACS) is an ongoing longitudinal study of men who have had sex with men. With nearly 30 years of longitudinal data on almost 7,000 HIV seropositi ve and HIV seronegative MSM, the MACS is an ideal cohort to study trends in smoking in MSM over time because of its large sample size, continued enrollment and repeated measures of smoking status. The aims of this study were (1) to evaluate differences in trends of cigarette smoking and change in daily consumption among HIV seropositive and seronegative men over time by birth cohort from 1984 2012, (2) examine predictors of smoking prevalence and smoking consumption among total MSM and HIV seropositive men.

PAGE 42

42 Methods Study Design and Administration The Multicenter AIDS Cohort Study (MACS) is an ongoing prospective cohort study of the natural and treated histories of HIV infection among MSM in the United States ( Kaslow R et al., 1987 ). The study has been descri bed in detail previously ( Kaslow R et al., 1987 ) . A total of 6,972 men were recruited at four centers: Baltimore/Washington DC, Chicago, Los Angeles, and Pittsburgh. Men were recruited in three waves, 4,954 in 1984 1985, 668 in 1987 1991, and 1,350 in 2001 2003. Informed consent was obtained from all participants, and the MACS study protocol was approved by the institutional review boards of each of the participating centers. Participants of the MACS return biannually for detailed interviews, physical exam inations, and collection of blood and laboratory testing. At each study visit, the men are asked detailed information about their smoking history since their previous visit. The questionnaires are available online at http://www.statepi.jhsph.edu/macs/forms.html . This present study utilizes a prospective cohort design to examine the association between demographic characteristics with self reported smoking. We utilized all data from the three wave s since smoking behavior was captured since their initial visit. The study sample included 6,577 men who reported their smoking behavior during their initial visit and at least one more visit. The median person years in the study was 9.6 years (interquarti le range: 5.4 18.5 years). Main Outcome Measure Current smoking was collected based on answers to a detailed interview. Participants were classified as never, former, and current smokers at each visit. These questions

PAGE 43

43 Participants who answered yes to both questions were categorized as current smokers. Participants were categorized as former smokers if they answered yes to the first question and no to the second. Never smokers were participants who answered no to both questions. Quantity of cigarette packs smoked were categorized by the MACS as follows: less than ½ pack per day; at least ½ but less than 1 pack per day; at least 1 but less than 2 packs per day; and 2 or more packs p er day. For the current study, the quantity consumed was dichotomized as less than one pack per day, and 1 pack or more a day. The MACS also assessed the length of time participants had smoked prior to joining the study. Participants were asked at what a ge they had begun smoking cigarettes and how many packs they had smoked during their heaviest smoking periods. If participants were former smokers, the men were asked for the number of years they had stopped smoking cigarettes. Using this information, we c alculated baseline cumulative smoking for each participant, and continued to add onto it while the participants remained in the study. Independent Variables Age at the each visit was calculated using self reported recorded date of birth and was treated a s a continuous variable. Self reported race at enrollment was categorized as follows: White non Hispanic; White Hispanic; Black non Hispanic; Black Hispanic; American Indian or Alaskan Native, Asian or Pacific Islander, other, or other Hispanic. Because of the small number of Hispanics (n=631), American Indian or Alaskan Native (n=23), Asian or Pacific Islander (n=32), and other (n=39) were grouped them together with other. Self reported educational attainment was collected from the most recent

PAGE 44

44 semi annual visit and was categorized as high school diploma or less, some college or college degree, and graduate work or more (reference group). We dichotomized participants into two groups based on their time of enrollment as either before or after 2001; before 20 01 was the reference group. Baseline characteristics differed by time of enrollment. Participants enrolled in 2001 or after were more likely to be younger, HIV seronegative, black non Hispanic, Hispanic, express depressive symptoms, have a high school dipl oma or less, unemployed, and were smokers at baseline. HIV serostatus was assessed using enzyme linked immunosorbent assay with and at every semiannual visit for p articipants who were initially HIV seronegative. Standardized flow cytometry was used to quantify CD4+ T lymphocyte subset levels by each MACS site ( Giorgi JV et al . , 1990; Schenker E et al., 1993 ). Through the course of the longitudinal study, 17.3% of HI V seronegative MSM were diagnosed with HIV. Self reported employment status was dichotomized as employed or unemployed. Self reported alcohol use was measured using questions about frequency of drinking and average number of drinks the participant consum ed since his last visit. Participants were categorized as no drinks since last visit, low moderate (1 2 drinks per day or 3 4 drinks per day no more than once a month), moderate heavy (3 4 drinks per day for more than once a month or 5 or more drinks per d ay for less than once a month), and binge (5 or more drinks for at least once a month) ( SAMHSA, 2014 ). Marijuana use, hospitalization in the last 6 months, and frequency of depressive symptoms (Occasionally or Most/All Days versus Rarely or Some days) were dichotomized.

PAGE 45

45 Data Analysis We first calculated the prevalence of smoking as the number of smokers over the total number of participants for each year. Smoking prevalence was stratified by birth cohort, and HIV serostatus. We used the Cochran Mantel Hae nszel test to measure the difference between birth cohort and HIV serostatus. We also used the Cochran Armitage trend test to assess changes by calendar year. We repeated this calculation to assess the prevalence of smoking one pack or more per day. We als o stratified by birth cohort and HIV serostatus. Univariate analyses were used to describe characteristics of the population as a function of HIV serostatus. Poisson regression with robust error variance was used to estimate prevalence ratios for smoking ( Zou G, 2004 ). Univariate and multivariate analysis were first done for all MSM and then HIV seropositive MSM. We used SAS 9.2 GENMOD procedure (SAS Institute, Cary, North Carolina, USA). We included age, race, education, employment, HIV serostatus, time o f enrollment, depressive symptoms, alcohol use, hospitalization, and marijuana use in the univariate models. Covariates with statistical significance at p<0.05 were entered into an exploratory multivariate model. Missing values for smoking status (n=113), age (n=1), race (n=2), education (n=50), and alcohol use (n=147) were imputed with values from the subsequent visit. We tested an interaction term for HIV serostatus and time of enrollment, and then stratified the analysis by HIV+ and HIV men. Two sensit ivity analyses were conducted to understand the changes in association based on seroconverters and lost to follow up. First, seroconverters were analyzed as HIV seropositives and were then removed to assess changes in baseline characteristics and longitudi nal associations (n=5,865). Additionally, we removed all participants that

PAGE 46

46 were lost to follow up to assess the same baseline characteristics and longitudinal associations (n=1,966). Results Baseline demographic data for HIV seropositive and HIV seroneg ative men are shown in Table 2 1. The prevalence of smoking was slightly higher among HIV seropositive men (44.1%) compared with HIV seronegative men (37.9%). HIV seropositive men were more likely to be black non Hispanic (19% compared with 13.2%) or other (11.5% compared with 7.2%), enrolled in 2001 or after (21.0% compared with 15.8%), have a high school diploma or less (19.0% compared with 14.0%), have depressive symptoms (25.6% compared with 21.6%), and were hospitalized in the past 6 months (6.4% compa red with 4.1%). The annual smoking prevalence has declined over time in the MACS. Shown in Figure 2 1a, 38.9% of participants enrolled in the MACS in 1984 smoked. In 2012, the prevalence sharply declined to 11.8% (test for trend p<0.0001). Among participa nts who were enrolled in 2001 or after, the prevalence of smoking in 2002 was 53.9% and 36.9% in 2012 (test for trend p<0.0001). Shown in Figure 2 1b, differences in prevalence were also observed by birth cohort. Participants in the oldest birth cohort (19 14 1934) had the lowest prevalence of smoking while participants in the youngest birth cohort (1960 1969) had the highest (test for trend p<0.05). However, in participants recruited after 2001, the oldest (1940 1949) and youngest (1970 1992) birth cohorts had a lower prevalence of smoking compared with participants born between 1950 1959 (test for trend p<0.05) (Figure 2 1b). Among participants who were enrolled before 2001, HIV seropositive men had a higher prevalence of smoking compared with HIV seronega tive men (Figure 2 1c). The

PAGE 47

47 rate of decline of smoking was the same among both groups. There were no differences observed among HIV seropositive and HIV seronegative men who were enrolled after 2001 (Figure 2 1 ). Similarly, there were no differences observ ed in the quantity of daily cigarettes smoked by birth cohort or HIV serostatus (not shown). In the multivariate analysis using the total sample, black non Hispanic, lower education, time of enrollment, living in Chicago and Pittsburgh, marijuana use and alcohol use were positively associated with smoking prevalence in MSM (Table 2 2). To understand the analysis more clearly, among all MSM participants, having less than a high school diploma was associated with a 20% (95% CI, 1.15 1.25) higher prevalence o f smoking compared with MSM who had attended graduate school or more; a 10% (95% CI, 1.07 1.15) higher prevalence for black, non Hispanic men compared with white, non Hispanic men; a 12% (95% CI, 1.06 1.14) higher prevalence of smoking if enrolled in 2001 or after; a 5% (95% CI, 1.01 1.08) higher prevalence of smoking if participants lived in Chicago compared with participants living in Baltimore/Washington DC; a 5% (95% CI, 1.01 1.08) higher prevalence of smoking if participants living in Pittsburgh compar ed with participants living in Baltimore/Washington DC; an 11% (95% CI, 1.09 1.14) higher prevalence rate if they were marijuana users; and a 11% (95% CI, 1.09 1.14) higher prevalence rate if they were binge drinkers. Being HIV seropositive was positively associated with smoking prevalence in the univariate analysis, but the association was no longer significant after adjusting for covariates. Although the association with HIV serostatus was no longer significant after adjusting for covariates, we further e xamined potential predictors of smoking using HIV related variables among HIV seropositive men (Table 2 3). Having less than a high

PAGE 48

48 school diploma was associated with a 18% (95% CI, 1.10 1.27) higher prevalence of smoking compared with MSM who had attended graduate school or more; a 13% (95% CI, 1.06 1.21) higher prevalence of smoking if enrolled in 2001 or after; a 9% (95% CI, 1.05 1.13) higher prevalence of smoking if participants used marijuana; a 2% (95% CI, 1.01 1.04) higher prevalence of smoking if p articipants exhibited depressive symptoms; and a 14% (95% CI, 1.08 1.16) higher prevalence of smoking if participants were binge drinkers compared to those who did not drink. CD4+ count and HAART use were not statistically significant associated with smok ing prevalence in the multivariate analysis. However, HIV seropositive men had a 4% (95% CI, 1.02 1.06) higher prevalence of smoking if they had a detectable viral load. The same analysis was performed to observe prevalence ratios for smoking one pack or more by among participants who were smokers. Shown in Table 2 4, among all MSM participants, black, non Hispanic, and other were less likely to smoke more than one pack a day compared to White non Hispanics. Having less than a high school diploma was assoc iated with a 10% (95% CI, 1.07 1.13) higher prevalence of smoking one pack or more per day compared with MSM who attended graduate school or more; a 6% (95% CI, 1.03 1.09) higher prevalence if they lived in Pittsburgh compared with those who lived in Balti more/Washington, DC; a 3% (95% CI, 1.00 1.05) higher prevalence for unemployed MSM compared with those that were employed; a 6% (95% CI, 1.04 1.09) higher prevalence rate if they were binge drinkers compared with MSM who were non drinkers; an 11% (95% CI, 1.08 1.09) higher prevalence in marijuana users compared with non users; and 3% (95% CI, 1.00 1.05) higher prevalence in MSM who were HIV seropositive compared with HIV seronegative MSM. After further

PAGE 49

49 stratifying by HIV serostatus, CD4+ count was still not statistically significant associated with a higher prevalence of amount smoked among HIV seropositive men (Table 2 5). HIV seropositive men had a 1% (95% CI, 1.00 1.02) higher prevalence of smoking if they had a detectable viral load. An interaction ter m was introduced to further assess the association of cigarette smoking on HIV serostatus by time of enrollment. There were no statistically significant interactions for both outcomes. After removing seroconverters from the dataset, we assessed changes in baseline characteristics and longitudinal associations. There were no major differences in baseline characteristics (data not shown). There were no major differences in prevalence ratios for smoking in multivariate analysis except for a statistically sign ificant positive association in participants enrolled after 2001 (data not shown). Additionally, we removed all participants that were lost to follow up in order to assess the same baseline characteristics and longitudinal associations. Because participant s who were enrolled before 2001 were more likely to have died or have been lost to follow up, baseline characteristics were similar to the characteristics of participants enrolled after 2001. Compared to the participants analyzed in the study, participants who were not lost to follow up, were more likely to be Black non Hispanic, enrolled after 2001, be unemployed, and be current smokers (data not shown). After conducting the same multivariate analysis, the same predictors were positively associated with pr evalence of smoking in total MSM, with the exception of hospitalization in the last 6 months. Additionally, being HIV seropositive was associated with a 4% (95% CI, 1.01 1.06) higher prevalence of cigarette smoking compared with HIV seronegative participan ts (data not shown).

PAGE 50

50 Discussion In this study, although prevalence of smoking in MSM remain high, there were no differences by HIV serostatus. We found that among all men in the MACS, the prevalence of current smoking has been declining significantly, wit h a greater likelihood of current smoking among men in certain subgroups. These include black, non Hispanic men, participants with lower education, those who were enrolled in 2001 and after, binge drinkers, and marijuana users. Multivariate analysis on HIV seropositve participants showed that, CD4+ cell count and HAART use were not statistically significantly associated with prevalence of smoking, but detectable viral load was. Additionally, we found that the prevalence of smoking one pack or more per day w as positively associated among HIV seropositive men. Lower education, unemployment, alcohol use and detectable viral load were positively associated with prevalence of smoking one pack or more per day among HIV seropositive men. Race, alcohol consumption, level of physical activity, depression, and substance abuse have been shown to be associated with smoking among PLWH ( Crothers K et al., 2009; Crothers K et al., 2007; Mckirnan DJ et al., 2006 ). The low prevalence of smoking among HIV seropositive and HIV seronegative MACS participants enrolled before 2001 may be lower than expected because of loss to follow up or death. Being in a cohort for nearly 30 years may have also modified their smoking behaviors because they are aware of being observed. The 2001 e nrollment in the MACS increased the number of minority participants in order to better represent the current HIV positive MSM in the US. Though there is a decline in prevalence of smoking among this group, they were still more likely to smoke compared with the earlier enrolled group. This can reflect the overall historical shift of smoking prevalence in the population as whole. We

PAGE 51

51 found the prevalence of smoking among those enrolled in 2001 and later to be similar to other HIV subpopulations ( Crother K et a l., 2009; Hessol NA et al. , 2014; Agaku IT, King BA, Dube SR., 2014 Interagency HIV Study that showed that 39% of women living with HIV were current smokers in 2011 ( Hessol NA et al., 2014 ). The Ve terans Aging Cohort found that 45.9% of HIV seropositive patients were current smokers ( Crothers K et al., 2009 ). Tesoriero et al . , reported the prevalence of smoking among PLWH in New York State to be 59% ( Tesoriero JM et al ., 2010 ). Lifson et al 2010, a ssessed smoking prevalence for 5,472 HIV seropositive men enrolled in 33 countries, and found that 40.5% were current smokers ( Lifson AR et al., 2010 ). Prevalence of smoking in MSM enrolled in 2001 and later was similar to older published studies ( Skinner W, 1994; Skinner W and Drug OM, 1996; Stall RD et al., 2005 ). For the first time in 2013, the National Health Interview Survey (NHIS) established a sexual orientation question for their annual health survey. The landmark addition will enable surveillance and long term monitoring of the Healthy People 2020 goals to improve the health, safety, and well being of lesbian, gay, and bisexual populations ( Ward B et al., 2013 ). In 2013, 27.2% of gay men between the ages of 16 64 were current smokers compared with 22.3% of straight men of the same age group. Although the prevalence of smoking differed in our current study, the NHIS did not include a question on HIV serostatus. Additionally, a recent study utilizing data from the National Survey on Drug Use and Healt h examined the association between sociodemographic characteristics and smoking status among HIV positive individuals ( Pacek LR, Harrel PT, Martins SS, 2014 ). Among their participants, 40% of PLWH were

PAGE 52

52 smokers, and were more likely to smoke if they were pr eviously married, binge drinkers, and were in lifetime drug and alcohol treatment. For both outcomes in this study, site was associated with higher prevalence of in 19 94 but allowed for exemptions for smoking in ventilated employee smoking rooms. The exemption still remains in effect ( Centers for Disease Control and Prevention, 2011 ). Washington, DC/Baltimore, MD and Chicago, IL have 100% smoke free laws in all non hosp itality workplaces, restaurants and bars ( Centers for Disease Control and Prevention, 2011 ). Pennsylvania has enacted smoke free worksites but allow exemptions for smoking in ventilated restaurants. Compared to participants in Washington, DC/ Baltimore, MD , MSM living in Chicago, IL and Pittsburgh, PA were more likely to smoke. More importantly, site was not statistically significant associated with smoking prevalence or quantity of cigarettes consumed among HIV seropositive participants. Among PLWH, dete ctable viral load was associated with prevalence of smoking while CD4+ cell count and HAART use were not statistically significant in this study. Previous studies assessing the association of smoking with the progression of HIV disease have yielded inconsi stent results. Royce and Winkelstein found that smoking increased CD4+ cell count but it was less pronounced among PLWH while Kabali et al did not find a statistically significant association between smoking and CD4+ cell count and viral load ( Royce R and Winkelstein W, 1990; Kabali C et al., 2011 ). It has been suggested that nicotine may alter the metabolism of HAART by increasing clearance and decreasing its efficacy, thus increasing viral load among smokers ( Wojna V, et al.,

PAGE 53

53 2007 ). This may have been wh y detectable viral load was associated with smoking prevalence in our study while HAART use remained statistically insignificant. It has been suggested that PLWH maintain the belief that they will not live long enough to suffer the adverse effects of tobac co use and therefore are not concerned about smoking cessation ( UHHS, 2014 , Burkhalter JE et al., 2005 ). However, as the life expectancy of PLWH continues to improve, there is a need to focus on modifiable risk factors such as smoking that will further red uce morbidity and mortality from non AIDS conditions ( Robinson WT et al., 2011 ). Additionally, in one study, 63% of current smokers reported they were interested in quitting smoking ( Reynolds NR, Neidig JL, Wewers ME , 2004) but low self efficacy (believin g that they will not be able to quit) is a strong predictor of non enrollment in smoking cessation programs ( Lloyd Richardson EE et al., 2008 ). As mentioned before, Crothers et al. (2007) showed that HIV providers were more likely to not identify patien ts who were current smokers when compared with non HIV providers. Providers who do identify patients that are current smokers often feel that smoking cessation is low priority because of competing priorities, economic barriers, and that it may impose an ad ditional burden on someone who is living with HIV ( Lifson AR and Lando AH, 2012 , Mamary EM, Bahrs D, Martinez S , 2002 ). A strength of our study is the large, diverse sample of HIV seropositive and HIV seronegative MSM representing four different cities wit We were able to assess the prevalence of smoking over time and determine factors associated with prevalence of smoking. However, our study also has limitations. Like most studies assessing cigarette smoking, we relied on s elf reported data and were

PAGE 54

54 unable to confirm smoking status with biomarkers such as salivary or blood cotinine. Additionally, as mentioned, the very low reported prevalence rate for men who were enrolled before 2001 could have been because of loss to follo w up, death, or participation bias. Our study shows that prevalence of smoking remains high among certain subpopulations of MSM including HIV ser o positive MSM. There is need for a continued effort to target MSM and PLWH with evidence based tobacco cessati on treatments. We also documented a strong tendency among men in the MACS to decrease smoking consumption. Other studies can build on this research and identify the predictors of successful smoking cessation. Understanding these predictors of smoking pre valence over time can inform targeted intervention for HIV seropositive and seronegative MSM to mitigate smoking associated comorbidities (i.e., heart disease, cancer) among HIV seropositive men and smoking related health disparities among MSM in general.

PAGE 55

55 Table 2 1 . Characteristics of the MACS population at time of e nrollment . 1 CESD Score of 16 or greater 2 The median was presented because of the unsymmetric al distribution of pack years. 3 Baseline was available for second new recruits. 4 Detectable viral load >40 copies/mL HIV HIV+ Total Men Age (mean, SD) 34.9 (8.6) 34.0 (7.7) 34.4 (8.2) Race (%) White, non Hispanic 79.6 (2535) 69.5 (2358) 74.4 (4893) Black, non Hispanic 13.2 (421) 19. 0 (644) 16.2 (1065) Other 7.2 (228) 11.5 (389) 9.4 (617) Enrolled after 2001 15.8 (503) 21.0 (712) 18.5 (1215) Site of Enrollment Baltimore/Washington DC 27.7 (882) 22.7 (771) 25.1 (1663) Chicago 21.7 (692) 24.7 (837) 23.3 (1529) Los Angele s 23.6 (752) 33.3 (1129) 28.6 (1881) Pittsburgh 27.0 (860) 19.3 (654) 23.0 (1514) Education (%) High School Diploma or Less 14.0 (443) 19.0 (639) 16.6 (1082) Some College or College Degree 46.5 (1471) 53.5 (1800) 50.1 (3271) Graduate work or More 39.5 (1247) 27.5 (927) 33.3 (2174) Unemployed 8.9 (282) 9.4 (313) 9.1 (595) Smoking Status Non smoker 41.4 (1310) 36.8 (1216) 39.1 (2526) Former Smoker 20.7 (653) 19.7 (630) 19.9 (1283) Current Smoker 37.9 (1198) 44.1 (1457) 41.1 (2655) Depressed Symptoms 1 (%) 21.6 (454) 25.6 (560) 23.6 (1014) Cumulative Pack years (median) 2 2.1 3.8 3.0 Marijuana Use 3 74.3 (199) 75.5 (247) 75.0 (446) Hospitalization in the past 6 months 4.1 (131) 6.4 (215) 5.3 (346) HAART No therapy 64.3 (7 01) Monotherapy 1.0 (11) Combined Therapy 2.4 (26) Potent Therapy 32.3 (352) Detectable Viral Load 4 70.2 (697)

PAGE 56

56 0 10 20 30 40 50 60 70 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 1914-1939 1940-1949 1950-1959 1960-1969 0 10 20 30 40 50 60 70 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 1940-1949 1950-1959 1960-1969 1970-1992 0 10 20 30 40 50 60 70 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 Seronegative Seropositive 0 10 20 30 40 50 60 70 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 Seronegative Seropositive d) c ) b ) a ) F igure 2 1 . Change in smoking prevalence over time. A) Annual prevalence of current smoking in the MACS by birth cohort, first wave (1984 2012). B) Annual prevalence of current smokin g in the MACS by birth cohort, second wave (2002 2012). C) Annual prevalence of current smoking in the MACS by serostatus, first wave (1984 2012). D) Annual prevalence of current smoking in the MACS by serostatus, second wave (2002 2012)

PAGE 57

57 Table 2 2 . Prevalence r atios for c igarette s moking from u nivariate and m ultivariate a nalysis Univariate Analysis Mu ltivariate Analysis Total Men p value Total Men p value Age .996 (.995 .996) <0.0001 0.999 (.998 1.001) 0.393 Race White, non Hispanic Reference Reference Black, non Hispanic 1.13 (1.10 1.15) <0.0001 1.11 (1.07 1.15) <0.0001 Other 1. 02 (.99 1.05) 0.177 0.95 (0.92 0.95) 0.028 Education High School Diploma or Less 1.25 (1.21 1.28) <0.0001 1.20 (1.15 1.25) <0.0001 Some College or College Degree 1.10 (1.08 1.12) <0.0001 1.09 (1.06 1.25) <0.0001 Graduate Work or More Referen ce Reference Enrolled after 2001 1.14 (1.11 1.17) <0.0001 1.12 (1.06 1.12) <0.0001 Unemployed 1.12 (1.10 1.14) <0.0001 1.06 (1.03 1.10) 0.0001 Site Baltimore/Washington DC Reference Reference Chicago 1.04 (1.02 1.07) .0015 1.05 (1.01 1.0 8) 0.009 Los Angeles 1.01 (0.98 1.03) 0.439 1.01 (0.98 1.04) 0.534 Pittsburgh 1.07 (1.04 1.09) <0.0001 1.05 (1.01 1.08) 0.009 Alcohol use None Reference Reference Low Moderate 1.21 (1.19 1.23) <0.0001 1.06 (1.03 1.08) <0.0001 Moderat e Heavy 1.34 (1.31 1.37) <0.0001 1.06 (1.02 1.11) 0.007 Binge 1.48 (1.44 1.53) <0.0001 1.12 (1.09 1.16) <0.0001 Marijuana use 1.14 (1.13 1.16) <0.0001 1.11 (1.09 1.14) <0.0001 Hospitalization in the last 6 months 1.03 (1.02 1.04) <0.0001 1.02 (0.9 9 1.03) 0.051 Depressed Symptoms 1.05 (1.04 1.07) <0.0001 1.00 (0.99 1.05) 0.02 HIV seropositive 1.05 (1.03 1.07) <0.0001 1.01 (1.00 1.06) 0.01 Enrolled after 2001*HIV seropositive 0.97 (0.94 1.00) 0.065

PAGE 58

58 Table 2 3 . Prevalence r atios for s m o king from u nivariate and m ultivariate a nalysis among HIV seropositive p articipants Univariate Analysis Multivariate Analysis HIV seropositive p value HIV seropositive p value Age 0.996 (.995 .997) <0.0001 0.999 (.0996 1.00) 0.35 Race White, non Hispanic Reference Reference Black, non Hispanic 1.10 (1.06 1.13) <0.0001 1.05 (0.99 1.12) 0.094 Other 1.01 (0.97 1.04) 0.6986 0.94 (0.87 1.02) 0.133 Education High School Diploma or Less 1.23 (1.19 1.27) <0.0001 1.18 (1.10 1.27) <0. 0001 Some College or College Degree 1.10 (1.07 1.13) <0.0001 1.07 (1.02 1.12) 0.007 Graduate Work or More Reference Reference Enrolled after 2001 1.13 (1.08 1.18) <0.0001 1.13 (1.06 1.21) 0.0004 Unemployed 1.11 (1.08 1.13) <0.0001 1.05 (0.99 1.10) 0.109 Site Baltimore/Washington DC Reference Chicago 1.03 (0.99 1.07) 0.134 Los Angeles 1.00 (0.97 1.03) 0.962 Pittsburgh 1.04 (0.99 1.08) 0.081 Alcohol use None Reference Reference Low Moderate 1.11 (1.07 1 .15) <0.0001 1.07 (1.03 1.12) 0.0003 Moderate Heavy 1.13 (1.06 1.20) <0.0001 1.08 (0.99 1.16) 0.069 Binge 1.22 (1.16 1.28) 0.0001 1.14 (1.08 1.21) <0.0001 Marijuana use 1.11 (1.09 1.14) <0.0001 1.09 (1.05 1.13) <0.0001 Hospitalization in the last 6 months 1.02 (1.01 1.04) 0.0013 1.02 (0.99 1.04) 0.234 Depressed Symptoms 1.07 (1.05 1.09) <0.0001 1.02 (1.01 1.04) 0.001 CD4 + T >500 Reference 201 500 1.00 (.99 1.02) 0.643 1.02 (0.99 1.04) 0.101 HAART Use No Therapy 1.04 (1.02 1.06) <0.0001 1.00 (0.96 1.03) 0.885 Monotherapy 1.01 (0.99 1.03) 0.421 1.01 (.93 1.10) 0.8 06 Combined Therapy 0.99 (0.96 1.01) 0.309 1.03 (0.96 1.10) 0.425 Potent ART Reference Reference Detectable Viral Load 1.06 (1.04 1.08) <0.0001 1.04 (1.02 1.06) <0.0001

PAGE 59

59 Table 2 4 . Prevalence ratios for s moking from univariate and m ultivar iate a nalysis among HIV seropositive p articipants Univariate Analysis Multivariate Analysis Total Men p value Total Men p value Age .997 (.996 .998) <0.0001 0.999 (.999 1.001) 0.445 Race White, non Hispanic Reference Reference Black, non Hispanic 0.95 (0.94 0.97) <0.0001 0.96 (0.94 0.98) 0.0004 Other 0.93 (0.92 0.95) 0.177 0.93 (0.91 0.95) <0.0001 Education High School Diploma or Less 1.13 (1.11 1.16) <0.0001 1.10 (1.07 1.13) <0.0001 Some College or College Degree 1.06 (1.05 1.08) <0.0001 1.04 (1.02 1.05) <0.0001 Graduate Work or More Reference Reference Enrolled after 2001 0.95 (0.94 0.96) <0.0001 1.02 (0.99 1.04) 0.145 Unemployed 1.03 (1.02 1.05) <0.0001 1.03 (1.00 1.05) 0.024 Site Baltimore/Washington DC Refe rence Reference Chicago 1.01 (0.99 1.03) 0.323 1.03 (0.96 1.10) 0.544 Los Angeles 0.99 (0.98 1.01) 0.646 1.00 (0.98 1.02) 0.653 Pittsburgh 1.07 (1.04 1.09) <0.0001 1.06 (1.03 1.09) 0.012 Alcohol use None Reference Reference Low Moder ate 1.07 (1.06 1.09) <0.0001 1.02 (1.00 1.04) 0.026 Moderate Heavy 1.12 (1.07 1.16) <0.0001 1.05 (1.01 1.03) 0.017 Binge 1.13 (1.44 1.53) <0.0001 1.06 (1.04 1.09) <0.0001 Marijuana use 1.07 (1.06 1.09) <0.0001 1.11 (1.08 1.09) 0.0016 Hospitalizat ion in the last 6 months 1.00 (0.99 1.01) 0.390 Depressed Symptoms 1.04 (1.02 1.05) <0.0001 1.02 (1.01 1.04) 0.0095 HIV seropositive 1.02 (1.01 1.42) 0.002 1.03 (1.00 1.05) 0.008 Enrolled after 2001*HIV seropositive 0.99 (0.96 1.03) 0.848

PAGE 60

60 CHAPTER 3 LONG TERM CIGARETTE SMOKING TRAJECTORIES AMONG HIV SEROPOSITIVE AND SERONEGATIVE MSM IN THE MULTICENTER AIDS COHORT STUDY Introduction The prevalence of current smoking among persons living with HIV (PLWH) has been estimated to be over 40%, mo re than 20% greater than the general US population (Mdodo R et al., 2015; Crothers K et al., 2009; Lifson AR and Lando AH, 2012; Lifson AR et al., 2010; Tesoriero JM et al., 2010; Pacek LR et al., 2014) . PLWH who smoke, smoke an average of 16 23 cigarettes per day, an indicat or of high nicotine dependence (Bernard A et al., 2007) . Studies from the 1990s suggest that smoking rates in men who have sex with men (MSM) were high ranging from 45% 49% (Royce RA and Winkelstein W, 1990; Stall RD et al., 1999) . Beca use of a history of exclusion and discrimination in other social settings, the social focus for many MSM has been gay identified bars and clubs, where the prevalence of smoking is thought to be high (Reynolds NR, 2009) . Recent data from the National Health Interview Survey, suggest that the current prevalence of smoking has dramatically shifted, 27.2% of gay men between the ages of 16 64 were current smokers compared with 22.3% of straight men of the same age group (Agaku I, 2014). However, it is not clear whether the lower rate of smoking today reflects that MSM are less likely to smoke, more likely to quit smoking, or whether they failed to initiate smoking. This rate likely reflects responses of MSM as a group to progressive public health policies regulat ing tobacco sales and use. Although the use of highly active antiretroviral therapy (HAART) has dramatically reduced HIV related morbidity and mortality, HIV seropositive individuals are now reaching ages at which smoking related disease rapidly increase s (Justice AC, 2010) . Studies assessing smoking in PLWH are similar to research from the general

PAGE 61

61 population showing that smoking is a risk factor for coronary artery disease, myocardial infarction, lung cancer, and stroke (Lifson AR and Lando AH, 2012 , Lifson AR et al., 2010 , Barbaro G et al., 2003; Friis Moller N et al., 2003) . Petoumenos et al. found that compared with those who had never smoked, those who stopped smoking during follow up, the odds ratio for the risk of cardiovascular disease (CVD) decreased from 2.3 within the first year of stopping to 1.5 after more than three years. Lit tle is known about long term smoking patterns among PLWH. A majority of studies have categorized smoking as current, former, or never smokers. Categorization Measures suc h as pack years does quantify duration and intensity; however, it does not capture the fluctuations in lifetime smoking that can be observed among those who quit or decrease cigarette smoking. The use of long term patterns of smoking provides a longitudina l measure that can be compared across different groups of interest. Given the study gaps described above, we constructed and characterized multiple long term trajectories of cigarette smoking among HIV positive and HIV negative MSM. We assessed whether HI V serostatus was associated with trajectory group membership. Finally, we analyzed multiple trajectories among HIV positive MSM, and examined how these trajectories varied by HIV specific time variant covariates. We used data derived from an ongoing longit udinal study with repeated measures over a period of 28 years. Dynamic models with both time constant and time varying covariates were used to evaluate differences in trajectories of cigarette smoking.

PAGE 62

62 Methods Study Population The Multicenter AIDS Cohor t Study (MACS) is an ongoing prospective cohort study of the natural and treated histories of HIV infection among MSM in the United States (Kaslow RA et al., 1987; Dudley J et al., 1995). A total of 6,972 men were recruited at four centers: Baltimore/Washi ngton DC, Chicago, Los Angeles, and Pittsburgh. Men were recruited in three waves, 4,954 in 1984 1985, 668 in 1987 1991, and 1,350 in 2001 2003. MACS participants complete study visits every 6 months during which they are tested for HIV (if HIV seronegativ e), provide a blood sample for storage in a repository for future research, undergo a physical examination, and complete questionnaires, which collect demographic, psychosocial, behavioral, medical history and health services data. The questionnaires are a vailable online at http://www.statepi.jhsph.edu/macs/forms.html . Informed consent was obtained from all participants, and the MACS study protocol was approved by the institutional review boards o f each of the participating centers. Participants of the MACS return biannually for detailed interviews, physical examinations, and collection of blood and laboratory testing. At each study visit, the men are asked detailed information about their smoking history since their previous visit. This present study utilizes a prospective cohort design to examine the association between demographic characteristics with self reported smoking trajectories. We utilized all data from three waves because smoking behav ior was captured since their initial visit. This analysis follows cigarette consumption from semiannual visits 1 through 57 of the MACS. The study sample included 6,535 men who reported their smoking

PAGE 63

63 behavior during their initial visit and at least one fol low up visit. The median person years in the study was 9.6 years (interquartile range: 5.4 18.5 years). Main Outcome Measure At each visit, current smoking status was assessed as part of the study questionnaires. Participants were classified as never, fo rmer, and current smokers at Quantity of cigarette packs smoked were categorized by the MACS as follows: less than ½ pack per day; at least ½ but less than 1 pack per day; at least 1 but less than 2 packs per day; and 2 or more packs per day. Using these measures, we constructed long term trajectories. Covariates of Interest We examined the following characteristics measured at the index visit in our analysis: age, race (indicated by two dummy variables, race included three categories: non Hispanic black, and other with non Hispanic white as the reference group), and education (indexed by high school diploma or less and having some college or more as the reference group) (Akhtar WZ et al., 2015) . Because participants enrolled after 2001 were more likely to be younger, HIV seronegative, black non Hispanic, and have a high school diploma or less, we also assessed time of enrollment in our analysis. When studying health trajectories over an extended period of observations, many of the covariates can change over time. Therefore, we conducted additional analyses by incorporating time varying covariates. In o ur analysis, a number of covariates (alcohol use, marijuana use, CD4 cell count, viral load, and HAART use) could vary with time

PAGE 64

64 (Cook RL et al., 2013; Lim SH et al., 2012) . Finally, we constructed a binary variable to identify those who had died (n=2124) or dropped out or censored (n=2527) of the study during the period of observation and were included in our model. These measures were treated as confounding variables instead of predictors in our model to assess the degree to which the findings depend on a (Hedeker D and Gibbons R, 2006) . Data Analysis We used group based, semi parametric mixture models to identify distinct trajectories of smoking among participants in the MACS, using the SAS PROC TRAJ program developed by Nagin and colleagues (Nagin D, 2005) . PROC TRAJ uses maximum likelihood estimation and yields parameter estimates that define a) trajectory shape and b) trajectory group membership probabilities. The two stage model selection process described by Nagin was used to define the optimal number of trajectory groups and the order of the polynomial needed to model the shape of each trajectory. A pre set rule that all trajectories are linear was used to structure the first stage search, and a zero inflated Poiss on model was specified for the amount smoked per visit. The optimal number of latent trajectory classes was determined by: 1) using the Bayes factor approximation to compare the difference in the Bayesian information criterion (BIC) scores between competin g models (Shwarz G, 1978) ; 2) calculating the average posterior probabilities for each class; and 3) assessing the utility of the latent classes in practice, including the similarity of trajectories between classes and the number of cases within each traje ctory class.

PAGE 65

65 To identify the distinct trajectories of cigarette smoking, we explored various models including intercept only and linear, quadratic, or cubic functions of time. Although a linear function may be sufficient in describing some trajectories, we chose to apply a quadratic function for some of the groups. Given the number of repeated observations available, the quadratic functions are more flexible in identifying the distinct trajectories. On the basis of changes in the BIC score as the number of trajectory groups increased from 2 to 10 and whether there was any overlap between the confidence intervals of adjacent trajectories, we chose 4 groups because the improvement in BIC began to level off after 4 groups. To first understand smoking behavior a mong our participants, we derived the basic trajectory groups by estimating a model in which smoking is a function of time only by wave, without any other covariates. Second, we evaluated baseline characteristics and time varying covariates in the probabil ities of belonging to the identified trajectories among all participants of the study. We repeated this same process among HIV positive MSM to adjust for HIV specific time varying covariates. This allowed for joint estimation of the parameters that describ e the shape of trajectory group curves and adjusted odds ratios (AORs) for the relationship between the covariates of interest and trajectory group membership. We used the joint estimation process because it yields standard errors that account for the unce rtainty of group assignments. To further confirm our results we used generalized linear mixed models (PROC GLIMMIX). Each trajectory was a separate outcome and the trajectory group that smoked the least was treated as the reference group. Only the intercep t was allowed to vary between subjects, and the regression slopes were assumed to be fixed effects.

PAGE 66

66 Results Men in the study were grouped into one of four distinct trajectories. Based on the trajectory models, we classified light smoking as less than ½ p ack per day (on average), and heavy smoking as greater than ½ pack per day (on average). The four patterns were: P ERSISTENT NON SMOKER : Accounting for 55.9% of all participants, persistent non smokers were characterized by nearly zero packs smoked along with a small negative linear slope. Participants in this group had little to no cigarettes smoked throughout the period of observation. P ERSISTENT LIGHT SMOK ER : Representing 11.0% of all participants, the course was distinguished by smoking approximately at least half pack a day and remained constant over time. H EAVY S MOKER TO N ON S MOKER : Participants in this group began with nearly smoking at least half pack but less than one pack per day but experienced substantial reduction over time, ending with smok ing almost no packs a day. They accounted for 10.0% of all participants in the study. P ERSISTENT H EAVY S MOKER : Characterizing 23.1% of the sample, the trajectory group exhibited a very high level of cigarette smoking that persisted over the observation pe riod. The trajectory did decline, but participants continued smoking more than half pack a day. At the baseline visit, black, non Hispanic men were more likely to be persistent heavy smokers among both HIV seronegative and seropositive men compared with w hite, non Hispanic men (Table 3 1). HIV seronegative participants in the second wave were more likely to be persistent heavy smokers compared with participants in the first

PAGE 67

67 wave (33.9% vs 19.8%). There were no differences in proportions among CD4 cell coun t, HAART use, and viral load by type of smoking trajectory. After adjusting for time constant and time varying covariates, we modeled long term smoking trajectories for all participants (Figure 3 1 and Table 3 2). As shown in Table 2, for all four trajec tory groups, cigarette smoking at a given time varied with not only time but also marijuana use and binge drinking. Several covariates of interest were associated with cigarette smoking group membership. Shown in Table 3, compared to the persistent non smo ker group, persistent heavy smoker group membership were associated with being enrolled in 2001 and after (AOR = 2.35 95% CI: 2.12 2.58), having high school diploma or less (AOR = 3.22 95% CI: 3.05 3.39). Time of enrollment and having a high school diploma or less was not only a significant predictor of cigarette smoking but also across all trajectory groups (p<0.0001). Additionally, participants that had died or dropped out were associated with persistent heavy smoker compared with the persistent non smoke r group. The variable was also significant predictor of across all trajectory groups (p<0.0001 and p=0.0002, respectively). To confirm our results, we used generalized linear mixed models. Each trajectory was a separate outcome and the persistent non smok er group was treated as the reference group. Black, non Hispanic MSMs, being enrolled 2001 and after, HIV serostatus, marijuana use, and binge drinking were all associated with persistent heavy smoker, heavy smoker to non smoker, and persistent light smoke r when compared with persistent non smoker groups (data not shown). Because we were also interested in HIV specific variables, we ran the same analysis using HIV seropositive participants and additionally adjusting for HIV specific

PAGE 68

68 time varying covariates. The patterns of smoking behavior remained the same (Table 5). Compared to the persistent non smoker group, persistent heavy smoker group membership was associated with being enrolled in 2001 and after (AOR = 2.13 95% CI: 1.67 2.59) and having a high schoo l diploma or less (AOR = 2.53 95% CI: 2.07 3.01). Discussion Our analysis of longitudinal data from the MACS identified four distinct, long term smoking trajectories in HIV seropositive and seronegative MSM. Reflecting the stability of smoking behavior ov er the lifetime, 21% of these men changed their smoking behavior during the study period. A tenth of our participants began with nearly smoking at least half pack but less than one pack per day but experienced substantial reduction over time, ending with s moking almost no packs a day. The remainder continued to either not smoke or smoke heavily. We demonstrated that among all participants, education, and time of enrollment were significant predictors across all trajectory groups. Among HIV seropositive part icipants, race was the only statistically significant predictor across all trajectory groups. Previous studies that examined smoking trajectories have presented results for trends among adolescents as they emerge into adulthood (Audrain McGovern J et al., 2004; Bernat DH et al., 2008; Hampson SE et al., 2013; Lessov Schlaggar CJ et al; 2008) . To our knowledge this is the first time smoking trajectories were used to establish distinct smoking patterns among MSM and PLWH using a large cohort study. Because t here is a difference among age groups, type of population, trajectory type, and gender, it is difficult to compare trajectory patterns across studies. However, the overall decrease of smoking as shown by our trajectory groups is consistent with the nationa l trend (Lifson AR and Lando HA, 2012; Agaku I et al., 2014) .

PAGE 69

69 Greenwood et al., found that a greater proportion of MSM reported cessation of tobacco use than reported current daily tobacco use. This indicates a vol untary inclination towards health promotion and recovery (Greenwood GL et al . , 2005) . This was also marked in our study, as there was evidence of reductions and quitting of smoking among MSM with demonstrated syndemic healthy conditions (Herrick AL, 2013) . Many studies have theorized that MSM have learned to overcome the negative effects of a certain exposure, how to cope with a traumatic experience, and how to avoid negative trajectories associated with risks (Kurtz SP et al., 2012; Stall R et al., 2008; F ergus S and Zimmerman MA, 2005) . Resilience among MSM has been defined as a process of adaptation and readjustment despite facing adversity (Rabkin JG et al., 1993) . The process of adaptation may involve psychological, social, and/or behavioral characteris tics (Fergus S and Zimmerman MA, 2006; Rabkin JG et al., 1993) . Over time, MSM have decreased recreational drug use, and increasingly participated in the gay rights movement, indicating health promotion, altruism and social justice (Mills TC et al., 2004; Stall R et al., 2001; Herrick AL et al., 2011; Kent M and Davis M, 2010) . Among MACS participants, a recent study showed that there were patterns of resilience against frequent stimulant drug use (Lim SH et al., 2012) . These patterns should aid in the design of intervention programs to continue to reduce cigarette smoking among MSM. Our study has several limitations. Although we used a large sample of HIV seropositive and seronegative MSM, those included in our samp le are still older and may be less diverse than those at highest risk of HIV in the United States. It has also been shown, that the MACS participants are a highly motivated group of MSM who

PAGE 70

70 have stayed in this study for a number of years and may differ fro m other MSM in the general population. To account for this, we added variables to assess how death and drop out were associated with trajectory group membership. Among all participants in the study and HIV seropositive participants, the variable was not on ly a predictor of cigarette smoking but also across the trajectory group. Specifically, the variable was statistically significant among the persistent light smoker group and persistent heavy smoker group among all MSM. Despite these limitations, the use of a large sample of HIV seropositive and seronegative MSM from different sites in the United States, the long term follow up, and the use of trajectory modeling are strengths of our study. Our findings expand current understandings of cigarette smoke patt erns among HIV seropositive and HIV seronegative MSM and should be considered in the development of targeted smoking cessation interventions among this population. Additionally, future studies should examine the factors that underlie resilience among MSM w ho quit smoking to find ways to incorporate them in interventions for those that continue to smoke. The preserved life expectancy of PLWH coupled with the increased rate of chronic illnesses associated with controlling viral loads emphasizes the need to id entify and test effective cessation treatments for smoking cessation in a vulnerable population.

PAGE 71

71 Table 3 1 . Descriptive s tatistics of baseline covariates by trajectory g roup and HIV serostatus HIV seronegative HIV seropositive Persistent non smo ker (n=1,925) Light smoker to non smoker (n=300) Heavy Smoker to non smoker (n=254) Persistent Heavy Smoker (n=699) Persistent non smoker (n=1,812) Light smoker to non smoker (n=363) Heavy Smoker to non smoker (n=277) Persistent Heavy Smoker (n=905) Age ( mean, SD) 33.6 (8.3) 35.0 (8.8) 34.0 (7.5) 35.6 (8.7) 33.3 (7.5) 34.2 (7.8) 33.1 (7.7) 34.0 (7.5) Race White, non Hispanic 63.2 (1600) 8.4 (213) 8.2 (208) 20.1 (509) 56.5 (1318) 9.5 (222) 7.5 (176) 26.5 (619) Black, non Hispanic 46.0 (193) 11 .9 (50) 6.0 (25) 36.2(152) 45.5 (289) 11.7 (74) 8.4 (53) 34.5 (219) Other 58.2 (132) 16.3 (37) 8.8 (20) 16.7 (38) 43.0 (205) 17.3 (67) 12.4 (48) 17.3 (67) Wave Enrolled before 2001 63.0 (1688) 9.0 (242) 8.2 (220) 19.8 (531) 56.8 (1520) 10.0 ( 268) 8.3 (222) 24.9 (666) 2001 and after 47.6 (237) 11.9 (59) 6.6 (33) 33.9 (169) 42.9 (292) 14.0 (95) 8.1 (55) 35.1 (239) Site of Enrollment Baltimore/Washingto n DC 65.3 (575) 9.9 (87) 8.5 (75) 16.4 (144) 55.3 (422) 10.9 (83) 7.5 (57) 26.3 ( 201) Chicago 61.5 (425) 9.6 (66) 5.9 (41) 23.0 (159) 52.2 (437) 9.0 (75) 9.0 (75) 29.9 (250) Los Angeles 64.2 (482) 9.5 (71) 10.3 (77) 16.1 (121) 55.8(630) 13.2 (149) 9.3 (105) 21.7 (245) Pittsburgh 51.8 (443) 9.0 (77) 7.0 (60) 32.2 (276) 51.4 (323) 8.9 (56) 6.4 (40) 33.3 (209) Education High School Diploma or Less 40.7 (180) 10.2 (45) 5.7 (25) 43.4 (192) 33.9 (214) 13.5 (85) 9.5 (60) 43.2 (273) Some College or College Degree 57.1 (838) 9.2 (135) 9.5 (140) 24.2 (355) 54.4 (967) 10.5 (187 ) 8.8 (157) 26.3 (467) Graduate Work or More 72.0 (895) 9.4 (117) 6.8 (85) 11.8 (147) 67.0 (618) 9.3 (86) 6.3 (58) 17.4 (160) Unemployed Depressive Symptoms CESD < 16 62.9 (1540) 9.3 (228) 7.7 (188) 20.1 (493) 57.5 (1401) 10.3 (251) 8.3 (202) 23.9 (583) 53.3 (360) 9.9 (67) 8.6 (58) 28.3 (191) 44.3 (369) 11.8 (98) 8.4 (70) 35.5 (296) Cumulative Pack Years 5.8 (13.8) 12.3 (15.0) 19.2 (17.2) 26.0 (18.7) 5.4 (13.4) 10.9 (12.9) 16.7 (15.9) 23.2 (15.5) Marijuana Use No 42.7 (29) 13.2 (9) 4.4 (3) 39.7 (27) 35.9 (28) 18.0 (14) 11.5 (9) 34.6 (27) Yes 37.8 (74) 13.8 (27) 6.6 (13) 41.8 (82) 26.1 (62) 17.7 (42) 8.0 (19) 48.3 (115) Hospitalization in the past 6 months No 61.1 (1859) 9.4 (285) 8.0 (242) 21.6 (12 3) 54.7 (1709) 11.0 (342) 8.3 (260) 26.0 (811) Yes 48.1 (63) 12.2 (16) 7.6 (10) 32.1 (42) 44.1 (94) 6.6 (14) 8.0 (17) 41.3 (88) CD4 + T cell count >500 52.3 (1072) 10.8 (221) 9.2 (188) 27.7 (568) 201 500 58.2 (603) 10.2 (106) 6.5 (67) 25.2 (261)

PAGE 72

72 Table 3.1 Continued SD = standard deviation; CESD = Center for Epidemiologic Studies Depression; HAART = highly active antiretroviral therapy HIV seronegative HIV seropositive Persistent non smoker (n=1,925) Light smoker to non smoker (n=300) Heavy Smoker to non smoker (n=254) Persistent Heavy Smoker (n=699) Persistent non smoker (n=1,812) Light smoker to non smoker (n=363) Heavy Smoker to non smoker (n=277) Persistent Heavy Smoker (n=905) CD4 + T cell count 50.6 (137) 13.3 (36) 8.1 (22) 28.0 (76) HAART No therapy 49.5 (335) 9.8 (66) 8.6 (58) 32.2 (218) Monotherapy 63.6 (7) 0.0 (0) 9.1 (1) 27.3 (3) Combined Therapy 45.0 (9) 15.4 (4) 3.9 (1) 26.9 (7) Potent Therapy 48.9 (114) 14.8 (51) 7.3 (25) 33.3 (115) Viral Load Not Detectable 46.9 (135) 12.2 (35) 6.6 (19) 34.4 (99) Detectable 44.5 (300) 11.9 (80) 9.6 (65) 34.0 (229) Died 4.0 (12) 5.4 (103) 9.1 (23) 11.6 (81) 52.1 (189) 57.8 (1047) 52.7 (146) 57.6 (521) Dropped Out / Censored 67.8 (204) 61.5 (1184) 52.2 (132) 65.9 (461) 16.5 (60) 15.1 (273) 13.4 (37) 18.0 (163 )

PAGE 73

73 Figure 3 1 . Trajectory groups of smoking consumption over time with adjustmen t for time constant and time varying variables in participants in the MACS

PAGE 74

74 Table 3 2 . E stimate groups and group specific growth parameters for all participants in the MACS Group 1 Group 2 Group 3 Group 4 Description Persistent non smoker Persistent light Smoker Heavy Smoker to non smoker Persistent Heavy Smoker Intercept 6.3264** 1.1471** 0.8572** 1.0924** Linear 0.0047** <0.0001** 0.0005 0.0001 Quadratic 0.00001** < 0.0001** Marijuana Use 1.1794** 0.2862** 0.0043 0.0081 Binge Drinker 0.1348* 0.0333* 0.0230 0.0091 Group Membership 55.9% 11.0% 10.0% 23.1% BIC • 71359.39 (N=123550) / 71309.39 (N=6525) * p < 0.05 ** p < 0.001 • Bayesian Information Criteria

PAGE 75

75 Table 3 3 . Factors asso ciated with t rajectory group m embership, a fter a djusting for time v arying covariates for all p articipants Persistent non smoker Persistent light Smoker AOR (95% CI) Heavy Smoker to non smoker AOR (95% CI) Persistent Heavy Smoker AOR (95% CI) Wald based chi squared statistic Age (mean, SD) Reference 1.02 (1.01 1.03) 1.01 (1.00 1.03) 1.01 (1.00 1.02) 0.23 Race White, non Hispanic Reference Reference Reference Reference Black, non Hispanic Reference 1.44 (1.16 1.73) 1.12 (0.76 1.48) 1.22 (1.03 1.42) 0.47 Other Reference 1.73 (1.44 2.02) 1.53 (1.19 1.88) 0.40 (0.09 0.70) <0.0001 Wave Enrolled before 2001 Reference Reference Reference Reference 2001 and after Reference 1.28 (0.96 1.60) 1.02 (0.61 1.44) 2.35 (2.12 2.58) <0.0001 Educ ation High School Diploma or Less Reference 1.76 (1.51 2.01) 1.30 (0.97 1.65) 3.22 (3.05 3.39) <0.0001 Some College or More Reference Reference Reference Reference HIV seropositive Reference 1.15 (0.93 1.37) 0.97 (0.72 1.23) 1.17 (1.01 1.34) 0. 36 Death Reference 1.32 (1.05 1.60) 1.22 (0.92 1.51) 1.95 (1.75 2.15) <0.0001 Drop out / Censored Reference 1.48 (1.24 1.72) 0.90 (0.64 1.17) 1.71 (1.53 1.89) 0.0002

PAGE 76

76 Figure 3 2 . Trajectory gro ups of smoking consumption over time with adjustment for time constant and time varying variables in HIV seropositive participants in the MACS.

PAGE 77

77 Table 3 4 . Estimated t rajectory groups and group s pecific growth parameters for HIV seropositive p articipants in the MACS Group 1 Group 2 Group 3 Group 4 Description Persistent non smoker Persistent light Smoker Heavy Smoker to non smoker Persistent Heavy Smoker Intercept 5.8115** 0.4994** 0.1214 0.9241** Linear 0.0063** 0.000** 0.0131** 0 .0003 Quadratic 0.0004** < 0.0001** Marijuana Use 1.1006** 0.1623* 0.0912 0.0295 Binge Drinker 0.2440** 0.0065* 0.0560 0.0179 CD4 Count 0.0016** 0.0002 <0.0001 0.0001** HAART Use 0.0888 0.1635** 0.0145 0.0059 Viral Load < 0.0001 < 0.0001 <0 .0001 < 0.0001 Group Membership 53.4% 11.1% 10.1% 25.4% BIC • 22339.38 (N=58083) / 22283.72 (N=3349) * p < 0.05 ** p < 0.001 • Bayesian Information Criteria

PAGE 78

78 Table 3 5 . Factors associated with trajectory group m embership, b efore a djusting for time varying c ovariates, HIV seropositive participants Persistent non smoker Light smoker to non smoker AOR (95% CI) Heavy Smoker to non smoker AOR (95% CI) Persistent Heavy Smoker AOR (95% CI) Wald based chi squared statistic Age (mean, SD) Reference 1.02 ( 1.00 1.04) 1.02 (0.99 1.05) 1.02 (1.00 1.04) 0.64 Race White, non Hispanic Reference Reference Reference Reference Black, non Hispanic Reference 2.23 (1.83 2.64) 0.88 (0.19 1.58) 1.30 (0.85 1.75) 0.001 Other Reference 1.65 (1.16 2.14) 2.42 ( 1.82 3.02) 1.42 (1.00 1.83) <0.0001 Wave Enrolled before 2001 Reference Reference Reference Reference 2001 and after Reference 1.75 (1.27 2.24) 1.39 (0.71 2.07) 2.13 (1.67 2.59) 0.447 Education High School Diploma or Less Reference 1.8 6 (1.47 2.25) 1.76 (1.17 2.35) 2.53 (2.07 3.01) 0.068 Some College or More Reference Reference Reference Reference Death Reference 1.49 (1.05 1.82) 1.13 (0.88 1.47) 2.31 (1.82 2.65) 0.002 Dropout / Censored Reference 1.41 (1.06 1.77) 0.48 (0.12 1.08 ) 1.61 (1.37 1.83) 0.0014

PAGE 79

79 CHAPTER 4 THE ASSOCIATION OF MID LIFE SMOKING STATUS ON PROCESSING SPEED AND MENTAL FLEXIBILITY AMONG HIV SEROPOSITIVE AND SERONEGATIVE OLDER MEN: THE MULTICENTER AIDS COHORT STUDY Introduction Since the introduction of ART, HIV positive individuals are living longer than before (Justice, 2009). In 2013, PLWH over the age of 50 accounted for 26% of the population of PLWH (CDC, 2015). It is predicted that, in 2015, more than half of PLWH in the United States will be 50 ye ars or older (Effros et al., 2008). As the epidemic reaches its fourth decade, more information is needed on how the virus, therapy, behaviors associated with PLWH, and the natural aging process all interact with each other, in order to understand health p roblems that PLWH may face. HIV and its treatments affect the aging process or the development of morbidities that are associated with the aging process (High et al., 2012). Impairment in cognition in PLWH is known as HIV associated neurocognitive disorder (HAND). Even in its mildest form, people living with HAND are less likely to adhere to medication recommendations, struggle to perform complex daily tasks, and have a worse quality of life, difficulty in obtaining employment, and a shorter survival rate ( Albert SM et al., 1999; Berger et al., 2005; Farinpour et al., 2003; Garvey et al., 2008; Garvey et al., 2009; Tozzi et al., 2007). As PLWH begin to age, there could be a rise in HAND because of the interactive effects of immune function and aging on the C NS. As many as 40% of HIV seropositive individuals suffer from HAND (Sacktor et al., 2001; Lindl et al., 2010). Conservative estimates suggest that HAND diagnoses will increase five to tenfold by 2030 (Lindl et al., 2010).

PAGE 80

80 Smoking, a modifiable risk fact or, is highly prevalent among PLWH. Smoking is a risk factor for many vascular diseases, such as atherosclerosis and thrombosis, which may increase the risk of cerebrovascular diseases and vascular dementia (Ott et al., 1998). Furthermore, recent studies h ave shown that cognitive changes in older PLWH are likely to include a cerebrovascular pathology (Sacktor et al., 2010; Valcour et al., 2004). Additionally, public health messages have led many to give up smoking but the extent to which this change influen ces subsequent cognitive decline remains unclear. Within the general population, a growing body of research suggests that smoking increases the risk of cognitive decline (Sabia et al., 2012; Sabia et al., 2008; Collins et al., 2009; Knopman et al., 2009; Nooyens et al., 2008; Peters et al., 2009). The relationship between smoking and cognitive decline in PLWH remains unclear (Durazzo et al., 2007; Wojna et al., 2007; Bryant et al., 2013). Because of the increased prevalence of smoking in PLWH and inflammat ion exacerbated by the disease, t he aim of this present paper is to examine the association between midlife smoking s tatus and future rate of cognitive decline in HIV seropositive and HIV seronegative MSM. Additionally, we assessed whether a difference in cognitive decline exists by serostatus. Methods Study Population The Multicenter AIDS Cohort Study (MACS) has been described in detail elsewhere (Kaslow et al., 1987). It is an ongoing longitudinal study of the natural and treated history of HIV infectio n and AIDS among MSM. The study identified risk factors for incidence and clinical expression of the infection (Kaslow et al., 1987). Since 1984, a total of 6,973 men have been recruited at four centers: the Baltimore and Washington, DC, area, Chicago, Los Angeles, and Pittsburgh. Participants return biannually for a

PAGE 81

81 detailed interview, physical exam, and blood draw for laboratory testing. In 1986, the to assess the effects of HIV on the brain and nervous system (Miller et al., 1990). The full battery is performed every two years, while the Trail Making and Symbol Digit Modalities tests are administered every six months. The subgroup undergoing the neuropsychological assessme nts was selected without systematic bias in an effort to represent the entire MACS cohort (McArthur et al., 1993). In the beginning of 2005, neuropsychological assessments were administered to the entire cohort of active participants in the MACS. Among all participants, 5, 470 men c ompleted at least one test battery . Of the 1,502 men who did not complete any testing were on average, younger, less educated, more likely to use recreational drugs, white, non Hispanic (Becker et al., 2014). W e limited our study to MACS participants who had information on smoking behavior before the age of 50 and at least five years ( 1 0 visits) of follow up. Time was anchored at age 50. This yielded 591 men and 10,821 observations. Exposure of Interest The exposure of interest was smoking prior to age 50. Smoking behavior was collected based on answers to a detailed interview. Participants were classified as yes to both questions were categorized as current smokers. Participants were categorized as former smokers if they answered yes to the first question and no to the second. Never smokers were participants who answe red no to both questions. We measured smoking as never, former, and current smoker.

PAGE 82

82 Amount of packs smoked is categorized by the MACS as follows: less than ½ pack per day; at least ½ but less than 1 pack per day; at least 1 but less than 2 packs per day; and 2 or more packs per day. We defined pack years based on the amount of cigarettes smoked per day. We calculated it by determining the average pack (based on the choices from above) and multiplying by 0.5. If a participant smoked 1 to 2 packs a day, then his current smoking exposure was calculated as 1.5 x 0.5 years = 0.75 pack years. This measure d pack days for one year for that specific visit. To assess cumulative smoking using pack years, we added each pack year observed until the age of 50. Outcome o f Interest The outcome of interest was rate of change in cognitive performance based on three tests; Trail Making A, Trail Making B, and Symbol Digit Modalities. For Trail Making A (Figure 4 1) , participants were given a sheet of paper with numbers (1 25) scattered all over the page. Participants were asked to draw lines to connect the numbers in ascending order. If they made an error, the error was quickly pointed out and the participant corrected and continued the assessment. If the participant became con correcting the participant (SHARE Clinic Manual, Baltimore, MD) . The time the total numb er of prompts given to the participant were recorded during each visit (SHARE Clinic Manual, Baltimore, MD) . For this study, we used the total time the participants took to complete the task. Similar to Trail Making A, for Trail Making B (Figure 4 2) the p articipants were given a sheet of paper with numbers (1 13) and letters (A L). Participants must connect the circles in an ascending pattern. This time,

PAGE 83

83 participants must alternate between the numbers and letters (1 A 2 B 3 C, etc.). The rules for timing, scoring, and correcting errors are the same as Part A (SHARE Clinic Manual) . Time of completion was also used. For Symbol Digit Modalities (Figure 4 3) , participants were given a sheet of paper with a box on the top of the page. The box has a symbol with a corresponding number. Below the key are boxes with symbols -and the participant must fill in each box with the corresponding number. Participants are told to fill the boxes as quickly as they can and that when they get to the end of the first line to go o n to the second line (SHARE Clinic Manual, Baltimore, MD) . If a mistake is made, they are asked to write over it. They are not to skip any boxes. Each participant was given 90 seconds to complete as many as they can. One point was given for each box with a correct number. Each test was assessed separately, raw scores were log10 transformed in order to stabilize the variance and to have a better approximate of a normal distribution for Trails A and B . Shown in Figure 4 4, after log transforming Trails A, the kurtosis value changed from 9.76 to 0.44. For Trails B, the kurtosis value changed from 9.18 to 0.20. Data with high kurtosis tend to have a distinct peak, decline rapidly, and have heavy tails. Raw scores from Symbol Digit Modalities had a kurtosis value 0.93 and only increased after log transforming. Therefore, we assessed Symbol Digit Modalities as raw scores (Figure 4 5). Using the mean and standard deviation of baseline HIV negative participants at age 50 , we calculated z scores to standardize the raw scores and reliable change indices (RCI) to assess change over time . We averaged the z score for all three tests for a composite score.

PAGE 84

84 Covariates of Interest First, we included d emographic variables at age 50 . Because all participants were aged 50, we u sed year of birth recorded during baseline visit to assess cohort effects. Education was assessed at the baseline of the study and during follow up. Education was categorized as 8th grade or less, 9th,10th, and 11th grades, 12th grade, at least one year of college but no degree, four years of college/received degree, some graduate work, and post graduate degree. We categorized education as high school diploma or less, some college or college degree, and graduate work or more . The MACS measures race as follo ws: White, non Hispanic; White, Hispanic; Black, non Hispanic; Black, Hispanic; American Indian or Alaskan Native, Asian or Pacific Islander, Other, or Other Hispanic. Because of the small number of Hispanics, American Indian or Alaskan Native, Asian or Pa cific Islander, Other, or Other Hispanic we categorized race as black, non Hispanic; white, non Hispanic; and other. Health Behaviors at age 50 included alcohol consumption (yes/no) and marijuana use (yes/no). Self reported alcohol use was measured usin g questions about frequency of drinking and average number of drinks the participant consumed since his last visit. Participants were categorized as no drinks since last visit, low moderate (1 2 drinks per day or 3 4 drinks per day no more than once a mont h), moderate heavy (3 4 drinks per day for more than once a month or 5 or more drinks per day for less than once a month), and binge (5 or more drinks for at least once a month) (SAMHSA, 2014). Quantity of cigarette packs smoked were categorized by the MAC S as follows: less than ½ pack per day; at least ½ but less than 1 pack per day; at least 1 but less than 2 packs per day; and 2 or more packs per day. Prese nce of ever having depression was defined using the Center for Epidemiological Studies Depression ( CES -

PAGE 85

85 D) Scale score of 16 or more. The CES D scale is an instrument that records the frequency of psychological symptoms during the past week with a value of 0 3 for each e of the time (1 4 days CES D scale and a score of 16 or more is used to determine probable cases of clinical depression. Cumulative pack years was calculated based on the amount of cigarettes smoked per day. If a participant smoked 1 to 2 packs a day, then their current smoking exposure would be calculated as 1.5 x 0.5 years = 0.75 pack years. This meas ures pack days for one year. A count variable was used to add each pack year for each visit until the age of 50. Time dependent covariates included hypertension, incident self reported angina, incident self reported heart attack, and incident self reported stroke. Among HIV seropositive participants we also assessed CD4+ count, viral load, and HAART use as time dependent covariates. Hypertension was measured using blood pressure r Self reported angina, heart attack, and stroke was measured using the questionnaire section for each visit. a doctor or other medical provider ever told you that you had [comorbidity]. HIV ser ostatus was assessed using enzyme linked immunosorbent assay with confirmatory semiannual visit for participants who were initially HIV seronegative. Standardized flow

PAGE 86

86 cytometry was used to quantify CD4+ T lymphocyte subset levels by each MACS site (Giorgi et al . , 1990; Schenker et al., 1993). Statistical Analysis In order to understand the association between mid life smoking status and cognitive decline we used t wo different methods and compared our results: Reliable Change Index (RCI) were used to assess cognitive change from baseline until the last visit for each participant. RCI have been used in the assessment of cognitive change after neurological surgeries ( Price et al., 2014). The index was calculated using the following formula: RCI = [(Change Score) (Mean Change Score controlgroup ) / (SD controlgroup ). We assumed that the averaged control group change represents natural decline in cognition which was then su btracted from individual change. This value was determine the frequency of cognitive impairment at visits 1 and 10 . Linear Mixed Models (LMM) were used to estimate t he association between smoking status at age 50 and 5 year cognitive decline. We fitted the intercept as random effects to account for individual differences in baseline cognitive performance. The model included terms for year of birth, number of previous tests, education, and race (model 1) and the in teraction of each of the cova riates with time. We included the interaction of each covariate with time because we hypothesized that all covariates influence the rate of decline . We expanded the model to includ e time dependent variables and their interaction with time: drug use, depressive symptoms and health measures (model 2).We repeated the same analysis using HIV seropositive men, and also tested for an interaction. Finally, we repeated the same analysis us ing cumulative pack years as the exposure variable.

PAGE 87

87 Results Of the 5,470 men who completed one test battery, we limited our study to participants who had at least 10 visits after the age of 50 (n=591, a baseline z score of more than 2 across all tests (n =0), and free of HIV associated dementia (HAD) at baseline (n=0). This yielded a total of 531 men and 10,821 observations. The median number of visits was 17 (interquartile range: 13 22). Among the men in the MACS at age 50, current smokers were more like ly to be born between 1960 1969, black, non Hispanic, and have a college degree or less (Table 4 1). Current smokers at the age of 50 were also more likely to be depressed compared with never smokers (18. 9 % and 10.6 %, respectively). There were a total of 2 3 incidences of diabetes, heart attack, heart failure, stroke, or transient ischemic attack. Current smokers were more likely to be HIV seropositive compared with never smokers ( 42.4 % and 35. 1 %, respectively). Among HIV seropositive participants, there wer e no differences in CD4+ cell count and HAART use by smoking status at age 50. Figure 4 5 shows test performance z score over time for four random participants for each test and the composite score. We observed variance in performance for some participant s, while others declined without even taking account for practice effects. For all four participants, the composite score did indeed stabilize the variance from all three tests. Shown in Figure 4 6, there were no visual differences in unadjusted composite z scores over time by smoking status at age 50. Reliable Change Index Table 4 2 provides a summary of the raw neuropsychological variables and reliable change scores by smoking status. Among current smokers, at baseline visit, the mean time to completion for Trail Making A was 23.53 seconds (SD: 8.1) and 56.38 seconds

PAGE 88

88 (SD: 24.5) for Trail Making B. After 5 years, the mean time to completions for Trail Making A and B was 22.80 seconds (SD: 8.2) and 55.36 (SD:35.4). Using RCIs we used a cutoff point of 1.9 6 to determine the frequency of cognitive impairment at visits 1 and 10 . No participants met the criteria for Trail Making A, and for Trail Making B and Symbol Digit Modalities the pro portion of impairment was under 3% across smoking statuses (Table 4 2). Because of the small number of participants in our cognitive impairment group, we did not further analyze the association using this method. Linear Mixed Models Smoking Status . Table 4 3 shows the estimates of subsequent cognitive change over 10 visits (~5 years) among HIV seropositive and seronegative participants (n=591) . C ompared to never smokers, former smokers and current smokers had a greater rate of decline in Trail Making A ( 0.0023 (95% CI: 0.0063, 0.0016) and 0.0029 (95% CI: 0.0067, 0.0008), respectively); in Trail Making B ( 0.1876 ( 95% CI: 0.4422, 0.0670) and 0.0839 ( 95% CI 0.3271, 0.1593) , respectively); Symbol Digit Modalities ( 0.0035 ( 95% CI: 0.0076, 0.0006) and 0.0009 ( 95% CI: 0 . 0048, 0.0030) , respectively); and Composite Score ( 0.0549 ( 95% CI: 0.1410, 0.0311) and 0.0195 ( 95% CI: 0.0108, 0.0627) , respectively). However, these rate of declines were not statistically significant. We observed the same trend when we stratified our analysis to only HIV seropositive participants (T able 4 4). We tested a three way interaction of smoking status, HIV serostatus, and time, the association was not statistically significant (p=0.648) . We also ass ociated with cognitive performance across all tests.

PAGE 89

89 Cumulative Smoking . Table 4 5 shows the estimates of subsequent cognitive change over at least 10 visits using cumulative pack years among HIV seropositive and seronegative participants (n=591). An incr ease of one unit of five pack years was statistically significantly associated with a rate of decline in Trail Making B and Composite Score ( 0.025 (95% CI: 0.0095, 0.0006), and 0.0077 ( 0.0153, 0.0002). We repeated the same analysis for HIV seropositi ve men only, and the rates of decline for Trail Making A, B, Symbol Digit Modalities, and Composite Score were not statistically significant. Discussion Our analysis of data using smoking assessment until the age of 50 and three cognitive assessments for the subsequent years , found that compared to never smokers, current and f ormer smokers were not statistically significantly associated with an increased risk of cognitive decline in Trail Making A, Trail Making B, Symbol Digit Modalities and Composite Sco re . When we stratified participants by HIV serostatus, the associations remained consistent. However, when we assessed subsequent cognitive change using pack years, we found that an increase of a unit of five pack years was statistically significantly asso ciated with a rate of decline in Trail Makin B and our Composite Score. Comparison with other studies To date, three studies have assessed the association of smoking on cognitive performance in PLWH. Durazzo et al. cross sectionally assessed 44 HIV serop ositive alcohol drinkers and found that smokers were more likely to perform worse than non smokers on tests for auditory verbal learning, auditory verbal memory, and cognitive efficiency (Durazzo et al., 2007). Wojna et al. performed a cross sectional stud y of 56

PAGE 90

90 participants and found no statistically significant associations between cognitive performance and current smoking or past smoking history in PLWH (Wojna et al., curren t smokers and learning, memory, and global cognitive functioning. After adjusting for education and HCV, the association was no longer statistically significant (Bryant et al., 2013). Other studies that have used the MACS to study the effects of aging have also observed statistically insignificant findings. A recent study from Becker et al., examined the relationships between the apilopoprotein E (ApoE) gene, HIV disease, age, cognitive impairment and death (Becker et al., 2015). Becker et al., found that a lthough HIV infection was associated with incident cognitive impairment, ApoE4 status was not statistically significantly associated with incident cognitive impairment. Mechanisms Smoking is associated with deficiencies in executive functions, cognitive flexibility, general intellectual abilities, learning and memory processing speed, and working memory (Durazzo et al., 2010). Among older adults, smoking is associated with abnormal rates of brain volume loss, especially in anterior frontal regions, subco rtical nuclei, and commissural white matter. HIV staining has shown that the virus is concentrated in the subcortical deep gray matter structures (Gabrieli et al., 1995; Gray et al., 2001; Woods et al., 2009). The effects of the virus are most prominent in the basal ganglia, frontal neocortex, and white matter tracts connecting these regions (Woods et al., 2009). However, current smoking at age 50 was not statistically significantly different when compared to non smokers.

PAGE 91

91 It is understood that smoking is a risk factor for many vascular diseases, such as atherosclerosis and thrombosis, which may increase the risk of cerebrovascular diseases and could influence executive function via vascular pathways (Ott et al., 1998). Recent studies have shown that cognitiv e changes in older PLWH are likely to include a cerebrovascular pathology (Sacktor et al., 2010; Valcour et al., 2004). We attempted to include vascular covariates, but the incidence of such events were close to none. The Trail Making Test in one of the most widely used neuropsychological tests because of its sensitivity and general indicator of brain dysfunction (Stuss et al., 2001; Retzlaff, Butler, and Vanderploeg, 1992; Lezak, 1995). Response speed is one demand of the test. Studies have shown that p atients with frontal lobe damage tend to perform worse on Trail Making Test but may fail for many reasons. Survivor Effects Because we attempted to use a subpopulation of the MACS dataset, we focused on participants that were aging. In a recent study by B ecker et al., of the 5,470 men who ever had a cognitive performance test, 32% from these aging men were still active. They compared their first and last visits, and found that there was a significant improvement in performance on the Trails A and B tests. Therefore, selection bias from selective mortality or other forms of attrition may have occurred to participants after study enrollment (Weuve et al., 2012). When studying cognitive decline, impaired cognition strongly predicts morbidity, mortality, and at trition after study enrollment. With a risk factor like smoking, which is also associated with morbidity and mortality, we were vulnerable to bias due to selective attrition. Future studies should consider inverse probability weighting approach to c orrect analysis for attrition.

PAGE 92

92 Limitations Our study has limitations. Although we used a large sample of HIV seropositive and seronegative MSM, those included in our sample were from the first wave of enrollment and may be less diverse than those at greater ris k of HIV in the US. More importantly, MACS participants may not be generalizable to PLWH or at risk for HIV as they are a highly motivated group of MSM. Second, we used three tests in the MACS because they were assessed at every 6 months instead of 2 years . These three tests measured the same cognitive processes (executive functioning, speed of information processing). Future studies with tests that are meant for screening PLWH with cognitive decline due to aging may be warranted. Smoking was assessed by t he use of a self reported questionnaire. Other ways of assessing cigarette smoke are bio chemical tests that detect carbon monoxide, thiocyanate, or cotinine in saliva, blood, urine, or exp ired air (Jarvis et al., 1987). Though l imitations in self reported ion about their tobacco use, studies suggest that self reported smoking history is reliable 0.86) when compared to biochemical measures (Soulakova et al., 2012; Murray et al., 2002). However, our study is first of its kind to longitudinally assess smoking on cognitive declin e in this subpopulation. More research on understanding the risk factors for cognitive performance due to aging in this subpopulation is warranted. It is important to note, that although our results are not consistent by exposure type, smoking still remain s high in this subpopulation and its effects are well documented. There is a continued need for smoking cessation programs and programs for health promotion.

PAGE 93

93 Figure 4 1 . Example of Trail Making Test Part A

PAGE 94

94 Figure 4 2 . Example of Trail Making Test Part B.

PAGE 95

95 Figure 4 3 . Example of Symbol Digit Modalities Test.

PAGE 96

96 Seconds Seconds Log Transformed Seconds Log Transformed Seconds Figure 4 4 . Distribution of Trail Making A and B raw and log transformed scores at baseline (age=50)

PAGE 97

97 . Score Log Transformed Score Figure 4 5 . Distribution of Symbol Digit Modalities raw and log transformed scores at baseline (age=50)

PAGE 98

98 Table 4 1 . Characteristics of the MACS Cohort Study as a function of smoking status at age 50 Never Smoker N= 269 Former Smoker N= 140 Current Smoker N= 182 Birth Cohort 1914 1949 42 ( 22.7 ) 27 ( 9.9 ) 20 (16. 0 ) 1950 1959 79 ( 42.7 ) 116 ( 42.6 ) 49 ( 39.2 ) 1960 1969 64 ( 34. 6 ) 129 ( 47.4 ) 56 ( 44.8 ) Race White, non Hispanic 161 ( 87.0 ) 244 ( 89.7 ) 97 ( 77.6 ) Black, non Hispanic 19 ( 3.3 ) 18 ( 6.6 ) 26 ( 20.8 ) Site of Enrollment Baltimore/Washington DC 60 ( 32.4 ) 88 ( 32.4 ) 30 ( 24.0 ) Chicago 45 (2 4.3 ) 51 ( 18.8 ) 29 ( 23.2 ) Los Angeles 40 ( 21.6 ) 81 ( 29.8 ) 29 ( 19.3 ) Pittsburgh 40 ( 21.6 ) 52 ( 19.1 ) 37 ( 29.6 ) Education High School Diploma or Less 8 ( 4.3 ) 18 ( 6.6 ) 22 (1 7.6 ) Some College or College Degree 71 (3 8.9 ) 107 ( 39.3 ) 61 ( 48.8 ) Graduate Work or More 106 ( 57.3 ) 147 ( 54.0 ) 42 ( 33.6 ) Unemployed 3 (3.4) 1 (2.0) 1 (2.1) Depressive Symptoms 19 ( 10.6 ) 36 ( 13.5 ) 23 (18.9) Cumulative Pack Years 0 (0.0) 19.2 (20. 1 ) 35.5 ( 22.7 ) Marijuana Use 31 ( 20.1 ) 99 ( 38 .8) 51 ( 45.5 ) Hypertension 37 ( 44.7 ) 71 ( 47.3 ) 20 (2 9.4 ) Diabetes 3 (3.5) 4 (2.7) 4 (5.9) Had Angina 2 (2.7) 6 (4.6) 1 (1.6) Had Heart Failure 1 ( 1.4) 1 (0.8) Had Stroke 1 (1.4) HIV seropositive 65 (35. 1 ) 100 ( 36.8 ) 5 3 (4 2.4 ) CD4+ T cell Count (cells/uL) >500 62 (74.7) 36 (78.3) 37 (80.4) 200 500 18 (21.7) 10 (21.7) 8 (17.4) 3 (3.6) 1 (2.2) ART No Therapy 8 (29.6) 4 (23.5) 6 (37.5) Monotherapy 5 (18.5) 1 (5.9) 2 (12.5) Combined Therapy 6 (22.2) 8 (47.1) 2 (12.5) Potent Therapy 8 (29.6) 4 (23.5) 6 (37.5) CESD = Center for Epidemiologic Studies Depression; H AART = highly active antiretroviral therapy

PAGE 99

99 a) b ) c ) d ) Z Score Figure 4 6 . B Spline of cognitive performance test z scores over time by A) Trail Making A; B ) Trail Making B; C) Symbol Digit Modalities; D) Composite Score

PAGE 100

100 Figure 4 7 . Change in composite z score over time by smoking status at age 50 Never Smoker Former Smoker Current Smoker Visits Z Score

PAGE 101

101 Table 4 2 . Neuropsychological raw t est s cores and s tan dardized reliable change index s cores by s moking s tatus at a ge 50 and v isit Visit 1 Visit 10 Visit 1 Reliable Change Score Visit 10 Reliable Change Score Frequency of cognitive impairment (RCI< 1.96) Trail Making A Never Smoker 23.78 (9.3) 22.13 (6.9) NA NA NA Former Smoker 21.15 (7.8) 20.63 (7.0) 0.08 (0.71) 0.15 (0.79) Current Smoker 23.53 (8.1) 22.8 (8.2) 0.33 (0.72) 0.12 (0.82) Trail Making B Never Smoker 49.69 (18.9) 44.48 (14.9) NA NA NA Former Smoker 48.74 (21.6) 44.79 (21.6) 0.15 (0.72) 0.08 (0.83) 2.6 (1) Current Smoker 56.38 (24.5) 55.36 (35.4) 0.03 (0.86) 0.09 (1.4) 2.5 (1) Symbol Digit Modalities Never Smoker 54.17 (8.5) 52.91 (10.1) NA NA NA Former Smoker 57.66 (10.6) 56.14 (10.9) 0.16 (0.83) 0.08 (0 .97) 2.9 (1) Current Smoker 53.17 (9.4) 51.9 (11.9) 0.24 (0.80) 0.30 (0.99) 3.1 (1)

PAGE 102

102 Table 4 3 . Association of smoking history (age 50) and cognitive change over the subsequent 5 years among all participants (n= 591 ) Cognitive change over 5 years Trail Making Test A Trail Making Test B Symbol Digit Modalities Composite Score Birth Cohort 1914 1949 0.0051 ( 0.0100, 0.0003) 0.8285 ( 1.1381, 0.5188) 0.0185 ( 0.0234, 0.0136) 0.3044 ( 0.1410, 0.0312) 1950 1959 0.0046 ( 0.0087, 0.000 4) 0.3285 ( 0.5941, 0.0629) 0.0077 ( 0.012, 0.0035) 0.1134 ( 0.2027, 0.0242) 1690 1969 Reference Reference Reference Reference Race White, non Hispanic Reference Reference Reference Reference Black, non Hispanic 0.0110 ( 0.0236, 0.0020) 0.3303 ( 1.1514, 0.4908) 0.0241 ( 0.0371, 0.0110) 0.1199 ( 0.3958, 0.1561) Other 0.0160 ( 0.0220, 0.0100) 0.3188 ( 0.0763, 0.7138) 0.0040 ( 0.0103, 0.0022) 0.0949 ( 0.0383, 0.2281) Education High School Diploma or Less 0.0020 ( 0.0085, 0.0 050) 0.4907 ( 0.9242, 0.0572) 0.0061 ( 0.0129, 0.0007) 0.1595 ( 0.0812, 0.1086) Some College or College Degree 0.0003 ( 0.0044, 0.0043) 0.0645 ( 0.2851, 0.3453) 0.0043 ( 0.0089, 0.0002) 0.0137 ( 0.1043, 0.0607) Graduate Work or More Reference Ref erence Reference Reference By Smoking Status Never Smoker Reference Reference Reference Reference Former Smoker 0.0023 ( 0.0063, 0.0016) 0.1876 ( 0.4422, 0.0670) 0.0035 ( 0.0076, 0.0006) 0.0549 ( 0.1410, 0.0311) Current Smoker 0.0029 ( 0. 0067, 0.0008) 0.0839 ( 0.3271, 0.1593) 0.0009 ( 00048, 0.0030) 0.0195 ( 0.0108, 0.0627) *p<0.05 **p<0.0001

PAGE 103

103 Table 4 4 . Association of smoking history (age 50) and cognitive change over the subsequent 5 years among HIV positive participants (n= 220 ) Cognitive change over 5 years Trail Making Test A Trail Making Test B Symbol Digit Modalities Composite Score Birth Cohort 1914 1949 0.0167 (0.0076, 0.0258) 0.5237 ( 0.0290, 1.0763) 0.0104 ( 0.0191, 0.0016) 0.1304 ( 0.0586, 0.3194) 1950 1959 0.0052 ( 0.0012, 0.0116) 0.0720 ( 0.3384, 0.4823) 0.0044 ( 0.0110, 0.0021) 0.0243 ( 0.1135, 0.1622) 1690 1969 Reference Reference Reference Reference Race White, non Hispanic Reference Reference Reference Reference Black, non Hispanic 0.0085 ( 0.0287, 0.0117) 0.7714 ( 0.5208, 2.0637) 0.0359 ( 0.0565, 0.0153) 0.2343 ( 0.1977, 0.6663) Other 0.0176 ( 0.0256, 0.0095) 0.2793 ( 0.2366, 0.7951) 0.0110 ( 0.0192, 0.0027) 0.0750 ( 0.0980, 0.2581) Education High School Diploma or Less 0.0 026 ( 0.0125, 0.0073) 0.1508 ( 0.0264, 0.9575) 0.0127 ( 0.0226, 0.0027) 0.0529 ( 0.2663, 0.1606) Some College or College Degree 0.0140 (0.0063, 0.0217) 0.4655 ( 0.1613, 0.6534) 0.0050 ( 0.0129, 0.0029) 0.0776 ( 0.0599, 0.2151) Graduate Work or Mor e Reference Reference Reference Reference By Smoking Status Never Smoker Reference Reference Reference Reference Former Smoker 0.1583 ( 0.0701, 0.3867) 0.4639 ( 0.8806, 0.0471) 0.1084 ( 0.0072, 0.0063) 0.1526 ( 0.2936, 0.0117) Current Smok er 0.0656 ( 0.2697, 0.1384) 0.1008 ( 0.4984, 0.2967) 0.1476 ( 0.0049, 0.0078) 0.0216 ( 0.1559, 0.1126) *p<0.05 **p<0.0001

PAGE 104

104 Table 4 5 . Association of cumulative pack years ( at age 50) and cognitive change over the subsequent 5 years among all parti cipants (n= 591 ) Cognitive change over 5 years Trail Making Test A Trail Making Test B Symbol Digit Modalities Composite Score Birth Cohort 1914 1949 0.0046 ( 0.0094, 0.0002) 0.8151 ( 1.1223, 0.5079)** 0.0179 ( 0.0228, 0.0130) 0.3008 ( 0.4 049, 0.1967)** 1950 1959 0.0039 ( 0.0080, 0.0003) 0.3062 ( 0.5729, 0.0394)* 0.0075 ( 0.0117, 0.0033) 0.1062 ( 0.1959, 0.0166)* 1690 1969 Reference Reference Reference Reference Race White, non Hispanic Reference Reference Reference Refere nce Black, non Hispanic 0.0110 ( 0.0240, 0.0020) 0.3859 ( 1.2231, 0.4514) 0.0257 ( 0.0390, 0.0125) * 0.1387 ( 0.4200, 0.1427) Other 0.0154 ( 0.0216, 0.0092) 0.3400 ( 0.0578, 0.7379) 0.0043 ( 0.0106, 0.0020) 0.1009 ( 0.0333. 0.2350) Education High School Diploma or Less 0.0018 ( 0.0084, 0.0049) 0.0862 ( 0.1953, 0.3677) 0.0087 ( 0.0281, 0.0106) 0.0211 ( 0.0740, 0.1162) Some College or College Degree 0.0007 ( 0.0037, 0.0050) 0.0028 ( 0.2474, 0.2418) 0.0022 ( 0.0061, 0.0018) 0.0098 ( 0. 0928, 0.0732) Graduate Work or More Reference Reference Reference Reference Cumulative Pack Years (per 5 pk years) 0.0006 ( 0.0010, 0.0003)* 0.0250 ( 0.0095, 0.0006)* 0.0003 ( 0.0006, 0.0001) 0.0077 ( 0.0153, 0.0002)* *p<0.05 **p<0.0001

PAGE 105

105 Table 4 6 . Association of cumulative pack years (age 50) and cognitive change over the subsequent 5 years among HIV positive participants (n= 220 ) Cognitive change over 5 years Trail Making Test A Trail Making Test B Symbol Digit Modalities Composite Score Birth Cohort 1914 1949 0.0160 (0.0073, 0.0246) 0.5237 ( 0.0290, 1.0763) 0.0104 ( 0.0191, 0.0016) 0.1304 ( 0.0586, 0.3194) 1950 1959 0.0047 ( 0.0016, 0.0111) 0.0720 ( 0.3384, 0.4823) 0.0044 ( 0.0110, 0.0021) 0.0243 ( 0.1135, 0.1622) 1690 1969 Reference Reference Reference Reference Race White, non Hispanic Reference Reference Reference Reference Black, non Hispanic 0.0107 ( 0.0306, 0.0093) 0.7714 ( 0.5208, 2.0637) 0.0359 ( 0.0565, 0.0153) 0.2343 ( 0.1977, 0.6663) Other 0.0173 ( 0.0253, 0.0092) 0.2793 ( 0.2366, 0.7951) 0.0110 ( 0.0192, 0.0027) 0.0750 ( 0.0980, 0.2581) Education High School Diploma or Less 0.0027 ( 0.0125, 0.0071) 0.1508 ( 0.0264, 0.9575) 0.0127 ( 0.0226, 0.0027) 0.0529 ( 0.2663, 0.1606) Some Colle ge or College Degree 0.0071 (0.0009, 0.0133) 0.4655 ( 0.1613, 0.6534) 0.0050 ( 0.0129, 0.0029) 0.0776 ( 0.0599, 0.2151) Graduate Work or More Reference Reference Reference Reference Cumulative Pack Years (per 5 pk years) 0.0003 ( 0.0313, 0.0091) 0.00 08 ( 0.4984, 0.2967) 0.0002 ( 0.0049, 0.0078) 0.0004 ( 0.1559, 0.1126) *p<0.05 **p<0.0001

PAGE 106

106 CHAPTER 5 CONCLUSIONS Accomplishments of the Dissertation Taken together, the three studies addressed several gaps in epidemiology of HIV. We sought to better understand how the prevalence of smoking changed in HIV seropositive and seronegative MSM. We also characterized smoking patterns over time among the same subpopulation. Finally, we assessed whether smoking was associated with cognitive performance in olde r HIV seropositive and seronegative MSM. We expected the first objective to provide new knowledge of the overall trend of smoking in PLWH and MSM. A lthough the prevalence of smoking in MSM remain high, there were no differences by HIV serostatus. We found that among all men in the MACS, the prevalence of current smoking has been declining significantly, with a greater likelihood of current smoking among men in certain subgroups. These include black, non Hispanic men, participants with lower education, thos e who were enrolled in 2001 and after, binge drinkers, and marijuana users. The results from this analysis h ave national implications since we can assess the rate of decline and compare it to the general population . We expected our second objective to map how smoking patterns change over time and whether or not we can predict these patterns using baseline characteristics. We were able to compare baseline characteristics of those that quit smoking and those that continued to smoke. We identified four distin ct, long term smoking trajectories in HIV seropositive and seronegative MSM. Reflecting the stability of smoking behavior over the lifetime, 21% of these men changed their smoking behavior during the study period. A tenth of our participants began with nea rly smoking at least half pack but less

PAGE 107

107 than one pack per day but experienced substantial reduction over time, ending with smoking almost no packs a day. The remainder continued to either not smoke or smoke heavily. We demonstrated that among all participa nts, education, and time of enrollment were significant predictors across all trajectory groups. Among HIV seropositive participants, race was the only statistically significant predictor across all trajectory groups. This laid groundwork for future target specific smoking cessation interventions . Results from the third objective brought light to an adverse health outcom e that can be caused by smoking, a modifiable risk factor. Our analysis of data using smoking assessment until the age of 50 and three cogn itive assessments for the subsequent years, found that compared to never smokers, current and former smokers were not statistically significantly associated with an increased risk of cognitive decline in Trail Making A, Trail Making B, Symbol Digit Modali ties and Composite Score. When we stratified participants by HIV serostatus, the associations remained consistent. However, when we assessed subsequent cognitive change using pack years, we found that an increase of a unit of five pack years was statistica lly significantly associated with a rate of decline in Trail Makin B and our Composite Score. Although our results are inconsistent, smoking is still highly prevalent and associated with known adverse health outcomes. Our analysis of data using smoking ass essment until the age of 50 and three cognitive assessments for the subsequent years , found that compared to never smokers, current and f ormer smokers were not statistically significantly associated with an increased risk of cognitive decline in Trail Mak ing A, Trail Making B, Symbol Digit Modalities and Composite Score . When we stratified participants by HIV serostatus, the

PAGE 108

108 associations remained consistent. However, when we assessed subsequent cognitive change using pack years, we found that an increase o f a unit of five pack years was statistically significantly associated with a rate of decline in Trail Makin B and our Composite Score. Smoking Cessation in PLWH Resiliency Future Directions Screening Tests for Cognitive Decline There is a need for cli nical measurement tools that will assess cognitive problems in aging PLWH. Subjective cognitive complaint tools such as the Cognitive Failures Questionnaire and the Prospective and Retrospective Memory Questionnaires have found to be helpful in assessing s everity of forgetfulness and cognitive problems. Objective cognitive tools such as the Mini Mental State Exam (MMSE) and the Montreal Cognitive Assessment (MOCA) are the recommended gold standards in PLWH. Studies that have aging PLWH should include tests such as the MOCA in their study for future study of cognitive decline. The MOCA, for example has been shown to be superior in sensitivity and specificity for classifying people with cognitive impairment (Sweet et al., 2011; Toglia et al., 2011). The MOCA takes about 10 minutes to administer and scores can range from 0 to 30. Scores that are 26 and higher are marked as normal for global cognition. The MOCA has been entrusted as a valid tool to detect global cognitive functioning among PLWH (Koski et al., 20 11). The short amount of time needed to administer the MOCA makes it feasible to be i ncluded in longitudinal studies.

PAGE 109

109 The MACS Study Mentioned briefly before, the MACS study was a convenient sample of MSM that were originally recruited through advocate networks and volunteers (Kaslow et al., 1987) . In some sites, media publicity and personal communications between the investigators and gay activists were used, while in others participants were recruited from pre existing AIDS study cohorts and organizati ons. As this cohort aged, we observed that these participants began to modify and improve their behaviors because of being part of a cohort study. The next recruitment focused on a minority MSM that were higher risk than the original cohort. Unfortunately, the cohort was younger, and we were not able to assess cognitive change among them. Cohorts addressing aging, such as the Veterans Aging Cohort Study (VACS) may be better suited to address risk factors associated with cognitive decline. Additionally, stud ies assessing how factors associated with resiliency (ie social support groups) may be associated with higher cognitive performance in late life.

PAGE 110

110

PAGE 111

111 APPENDIX LITERATURE REVIEW Studies Assessing Smoking on Cognitive Decline in Community Dwelling Older Adults and PLWH Study Name Length of Follow up time Smoking Exposure Variable Type Outcome Major Findings/Strengths and Limitations Active Smoking SOPS (Broe, 1998) 3 years Self report at baseline: asked whether they had ever smoked re gularly Categorical (current, former, or never smoker) DSM IV NINCDS ADRDA No consistent associations were observed for smoking and dementia or cognitive impairment. Study was assessing nutritional factors associated with just a factor that they included in their model. Cleveland Study of the Elderly (Ford, 1996) 4 years Self report at baseline Binary (ever or never smoker) SPMSQ No significant relationship between smoking and cognitive impairment was observed. Smoking was categorized into two groups. Possible misclassification of exposure since intensity and duration was The Kame Project (Graves, 1999) 2 years Self report at baseline Binary (curre nt or non smoker) CASI No significant associations were observed. The CASI test was used to establish presence of dementia. CASI is not sensitive to PLWH. Juan, 2004 2 years Self report Never Smokers: Never having a cigarette Past Smokers: Quit for at least 6 months Current Smokers: Reported smoking at least 1 cigarette a day Participants that intermittently smoked less than Categorical (current (based on pack years light, medium, heavy, and very heavy), former, or never smoker) DSM III R, MMSE Compared with light smokers, the risk for smokers with a medium level of exposure and even higher in the heavy smoking group. First study to categorize current smoker based on intensity. Ag ain, MMSE is not sensitive to PLWH.

PAGE 112

112 Studies Assessing Smoking on Cognitive Decline in Community Dwelling Older Adults and PLWH Study Name Length of Follow up time Smoking Exposure Variable Type Outcome Major Findings/Strengths and Limitations one a day were removed from the study Zutphen Elderly Study (Launer, 1996) 3 years Self Report Categorical (current, former (further divided by the time since they stopped smoking), or never smoker) MMSE Yearly cognitive performance change decreased in curr ent and former smokers compared to never smokers. First study to assess time since former smokers quit cigarette smoke. EURODEM (Launer, 1999) Varied Self report Categorical (current, former, or never smoker) DSM III R, NINCDS ADRDA Significant associat ions were observed in current smokers and not in former smokers. The authors conclude that it may be due to chance since many of the outcomes were related to aging. CSHA (Laurin, 2003) 5 years Self report Binary (ever or never smoker) CIND No significan t differences were observed. Smoking was a covariate that was being assessed. Little attention was paid in assessing it as an exposure. HAAS (Laurin, 2004) Varied Self report Binary (ever smoked nearly every day) DSM III, NINDS ADRDA; CADDTC No signific ant associations were observed. Smoking was a confounder that was being assessed for a larger study on dietary intake of antioxidants and risk of late life dementia. CSHA (Lindsay, 2002) 5 years Self report Binary (ever or never smoker) NINCDS ADRDA, Com pared to never smokers, ever smokers had a positive association with used a large sample and divided their group based on ever or never smoking. No dose response was able to be established. The study was also assessi ng any disease, so associations may be due to chance.

PAGE 113

113 Studies Assessing Smoking on Cognitive Decline in Community Dwelling Older Adults and PLWH Study Name Length of Follow up time Smoking Exposure Variable Type Outcome Major Findings/Strengths and Limitations WHICAP (Luschsinger, 2005) Varied Self report Binary (current or non smoker) NINCDS ADRDA disease in former smokers compa red to never smokers. The authors suspect that the association may be due to coincidence of common disorders in the elderly, an additive or synergistic pathogenesis of dementia, or misclassification of vascular dementia as AD. SOF (Lui, 2003) 4 years Self report of current smoking status at baseline Binary (current or non smoker) MMSE The relative risk for cognitive decline was greater in former smokers compared to never smokers. MMSE was used to assess cognitive impairment in the general population. MMSE is not sensitive to PLWH. WHICAP (Merchant, 1999) Varied Self report at baseline Categorical (current, former, or never smoker) DSM IV, NINCDS ADRDA, CDR was greater in current and former smokers compared to nev er smokers. BLSA (Moffat, 2004) Varied Self report Binary (current or non smoker) DSM III R, NINCDS ADRDA was greater in ever smokers compared to never smokers. EURODEM (Ott, 2004) Varied Self report Categorical (current, former, or never smoker) MMSE The relative risk for yearly cognitive performance change was 0.16 ( 0.22, 0.10) for current smokers versus never smokers. SITE (Paleologos, 1998) 4 years Self report Binary (ever or never smoker) MMSE Compared to never smokers, ever smokers had a positive association with incidence of cognitive impairment. Kungsholmen Project (Wang, 1999) Varied Self report Binary (ever or never smoker) DSM III R No significant associations were observed VA Normative Aging Study (Weisskopf, 2004) Varied Self report Categorical (current, former, or never smoker) MMSE No significant associations were observed

PAGE 114

114 Studies Assessing Smoking on Cognitive Decline in Community Dwelling Older Adults and PLWH Study Name Length of Follow up time Smoking Exposure Variable Type Outcome Major Findings/Strengths and Limitations Health and Lifestyle Survey (Whittington, 1997) 7 years Self report Binary (current or non smoker) Incidental memory, visuo spatial reasoning, reaction time tasks No significant associations were observed. The Hisayama Study (Yoshitake, 1995) Varied Self report Binary (current or non smoker) DSM III, NINCDS ADRDA, NINDS AIREN No significant associations were observed. White hall II Cohort Study (Sabia, 2012) 10 years Self report Categorical (current, recent ex, long term ex, or never smoker) At follow up, participants were categorized at persistent smokers (continued smoking), intermittent smokers (quitters who started smokin g again), and quitters (stopped smoking). Neuropsychological Batteries Men who continued smoking over follow up had greater decline in all cognitive tests. First study to use longitudinal data to assess smoking at follow up and assess cognition longitudina lly. H EPESE Study (Collins, 2009) 7 years Self report Binary (Smoker or non smoker) MMSE Smoking increased the risk of cognitive decline in older Mexican Americans. Passive Smoking English Longitudinal Study of Ageing (Llewellyn, 2009) Cross sectional Salivary Cotinine Categorical (Cotinine was categorized in quartiles) Neuropsychological Battery Higher cotinine concentration was associated with increased risk for cognitive impairment.

PAGE 115

115 Studies Assessing Smoking on Cognitive Decline in Community Dwelling Older Adults and PLWH Study Name Length of Follow up time Smoking Exposure Variable Type Outcome Major Findings/Strengths and Limitations Cardiovascular Health Cognition Study (Barnes, 2010) 6 years Self Report. Participants were asked if they ever lived with anyone that smoked regularly, the total number of years, and time period. Categorical, based on number of years: none/low, moderate, and high DSM IV Moderate and high SHS exposure levels were not in dependently associated with dementia risk. NHANES 99 01 (Akhtar, 2012) Cross Sectional Serum Cotinine Continuous, log transformed DSST Cognitive performance declined as serum cotinine concentration increased. Active Smoking in PLWH Durazzo, 2007 Cross sectional Self report Binary (Smoking and non smoking heavy drinkers) Neuropsychological Battery Smokers were more likely to perform worse than non smokers on tests for auditory verbal learning, auditory verbal memory, and cognitive efficiency. (Wojna et al., 2007) Cross sectional Self report Tobacco questionnaire: Fagerström Test for Nicotine Dependency Neuropsychological Battery No associations were observed. (Bryant et al., 2013) Cross sectional Self Report Categorized (Current smoker, never smoker, past smoker) Neuropsychological Battery Negative association existed in PLWH current smokers and learning, memory, and global cognitive functioning. After adjusting for education and Hep C, the association was no longer significant.

PAGE 116

116 LIST OF REFERENCES Acavi L and Benowitz NL (2004). Cigarette smoking and injection. Arch Intern Med. 164:2206 16. Agaku IT, King BA, Dube SR. Current tobacco use among adults in the United States, 2005 2012. MMWR 2014;63:29 34. Almeida OP, Hulse GK, Lawrence D, Flicker L ( 2002) Smoking as a risk factor for Alzheimer's disease: contrasting evidence from a systematic review of case control and cohort studies. Addiction. 97:15 28. Andruff H, Carraro N, Thompson A, et al. (2009) Latent class growth modelling: A tutorial. Tutori als in Quantitative Methods for Psychology. 5:11 24. Smoking as a risk factor for dementia and cognitive decline: a meta analysis of prospective studies. Am J Epidemiol. 166:367 78. Antinori A, Arendt G, Becker JT, et al. (2007). Updated research nosology for HIV associated neurocognitive disorders. Neurology. 69:1789 99. Barbaro G, Di Lorenzo G, Cirelli A, et al. (2003) An open label, prospective, observational study of the incidence of coronary artery d isease in patients with HIV infection receiving highly active antiretroviral therapy. Clin Ther 25:2405 18. Benard A, Bonnet F, Tessier JF, et al. (2007) Tobacco addiction and HIV infection: toward the implementation of cessation programs. ANRS CO3 Aquitai ne Cohort. AIDS Patient Care STDS. 21:458 68. Becket BW, Thames AD, Woo E, et al. (2011) Longitudinal change in cognitive function and medication adherence in HIV infected adults. AIDS Behav. 15:1888 1894. Becker JT, Kingsley L, Mullen J, et al. (2009) HIV serostatus, and cognitive dysfunction in gay and bisexual men. Neurology 73:1292 99. Bryant VE, Kahler CW, Devlin KN, et al. (2013) The effects of cigarette smoking on learning and memory performance among people living with HIV/AIDS. AIDS Care. 25:1308 16. Burkhalter JE, Springer CM, Chhabra R, et al. Tobacco use and readiness to quit smoking in low income HIV infected persons. Nicotine and Tobacco Research. 2005;7(4):511 522 Centers for Disease Control and Prevention. HIV/AIDS among persons aged 50 an d older. Available at: http://www.cdc.gov/hiv/pdf/library_factsheet_HIV_among_PersonsAged50andOld er.pdf . Centers for Disease Control and Prevention. State Smok e Free Laws for Worksites, Restaurants, and Bars United States, 2000 2010. MMWR 2011. 60(15);472 75

PAGE 117

117 Centers for Disease Control and Prevention. Current Cigarette Smoking Among Adults United States, 2005 2012.. Morbidity and Mortality Weekly Report 2014;63( 02):29 34 Crothers K, Griffith TA, McGinnis KA, et al. (2005). The impact of cigarette smoking on mortality, quality of life, and comorbid illness among HIV positive veterans. J Gen Intern Med. 20:1142 5. Crothers K, Butt AA, Giber CL, et al. (2006) Incre ased COPD among HIV positive compared to HIV negative veterans. Chest. 130:1326 33. Crothers K, Goulet JL, Rodriguez Barradas MC, et al. (2009) Impact of cigarette smoking on mortality in HIV positive and HIV negative veterans. AIDS Educ Prev . 21;40 53. Cy sique LA, Bain MP, Brew BJ, Murray JM (2011) The burden of HIV associated neurocognitive impairment in Australia and its estimates for the future. Sex Health. 8:541 50. Durazzo TC, Meyerhoff DJ, Nixon SJ (2010) Chronic Cigarette Smoking: Implications for N eurocognition and Brain Neurobiology. Int J Environ Res Public Health. 7:3760 3791. Durazzo, TC, Rothlind, JC, Cardenas V,et al (2007). Chronic cigarette smoking and heavy drinking in human immunodeficiency virus: consequences for neurocognition and brain morphology. Alcohol. 41 (7):489 501. Effros RB, Fletcher CV, Gebo K, et al. (2008) Aging and infectious diseases: workshop on HIV infection and aging: what is known and future research directions. Clin Infect Dis. 47:542 53. Fabbiani M, Ciccarelli N, Tana M, et al. (2013) Cardiovascular risk factors and carotid intima media thickness are associated with lower cognitive performance in HIV infected patients. HIV Medicine 14:136 144. Feldman JG, Minkoff H, Schneider MF, et al. (2006) Association of cigarette s moking with HIV prognosis among women in the HAART era: a report from the women's interagency HIV study. Am J Public Health. 96:1060 5. Fischer MJ, Wyatt CM, Gordon K, et al. (2010) Hepatitis C and the risk of kidney disease and mortality in veterans with HIV. J Acquir Immune Defic Syndr. 53:222 6. Frieberg MS, Chang CC, Skanderson M, et al. (2011) T he risk of incident coronary heart disease among veterans with and without HIV and hepatitis C . Circ Cardiovasc Qual Outcomes. 4:425 32. Friis Møller N, Sabin C A, Weber R, et al. (2003) Combination antiretroviral therapy and the risk of myocardial infarction. N Engl J Med. 349:1993 2003.

PAGE 118

118 Gabrieli, J. (1995). Contribution of the basal ganglia to skill learning and working memory in humans. In Houk JC, Davis JL, Be iser DG (Eds.), Models of Information Processing in the Basal Ganglia. Cambridge, MA: MIT Press. Giorgi JV, Cheng H, Margolick JB, et al. Quality controls in the flow cytometric measurement of T lymphocytes subsets: The Multicenter AIDS Cohort Study experi ence. Clin Immunol Immunopathol. 1990;55(2):173 186. Gray F, Adle Biassette H, Chretien F et al. (2001). Neuropathology and neurodegeneration in human immunodeficiency virus infection. Pathogenesis of HIV induced lesions of the brain, correlations with HI V associated disorders and modifications according to treatments. Neuropathol. 20:146 55. Greenwood GL, Paul JP, Pollack JP, et al. Tobacco use and cessation among a household based sample of US urban men who have sex with men. Am J Public Health. 2005;95: 145 151. Grunfeld C, Delaney JA, Wanke C, et al. (2009) Preclinical atherosclerosis due to HIV infection: carotid intima medial thickness measurements from the FRAM study. AIDS. 23:1841 9. Hardy D and Vance D (2009) The neuropsychology of HIV/AIDS in older adults. Neuropsychol Rev. 19:263 72. Heaton RK, Clifford DB, Franklin DR, et al. (2010) HIV associated neurocognitive disorders persist in the era of potent antiretroviral therapy: CHARTER study. Neurology 75:2087 96. Heaton RK, Franklin DR, Ellis RJ, et al. (2011) HIV associated neurocognitive disoders before and during the era of combination antiretroviral therapy: differences in rate, nature, and predictors. J Neurovirol. 17:3 16. Helleberg M, Afzal S, Kronborg G, et al. (2013) Mortality attributable to smoking among HIV 1 infected individuals: A nationwide, population based cohort study. Clinical Infect Dis. 56:727 734. Am J Prev Med. 2014;47(1):53 69 High KP, Brennan Ing M, Clifford DB, et al. (2012) HIV and aging: state of knowledge and areas of critical need for research. A report to the NIH Office of AIDS Research by the HIV and Aging Working Group. J Acquir Immune D efic Syndr. 60 Suppl 1:S1 18. Hogg R, Lima V, Sterne JA, et al. (2008) Life expectancy of individuals on combination antiretroviral therapy in high income countries: a collaborative analysis of 14 cohort studies. Lancet. 372:293 299. Jacob T, Blonigen DM, Upah R, et al. (2013) Lifetime drinking trajectories among veterans in treatment for HIV. Alcohol Clin Exp Res. 37;1179 87.

PAGE 119

119 Jarvis MJ, Pedoe H, Feyerabend C, et al. (1987) Comparison of tests used to distinguish smokers from non smokers. Am J of Pub Health . 77:1435 1438. Jones B, Nagin D, Roeder K. (2001) A SAS procedure based on mixture models for estimating developmental trajectories. Social Method Res . 29:374 93. Juan D, Zhou DH, Li J, et al. (2004) A 2 year follow up study of cigarette smoking and risk of dementia. Eur J Neurol. 4:277 82. Justice AC (2010) HIV and Aging: Time for a new paradigm. Curr HIV/AIDS Rep. 7:69 76. Kabali C, Cheng DM, Brooks D, et al. Recent cigarette smoking and HIV disease progression: No evidence of an association. AIDS Care. 2011;23(8):947 956. Kaslow RA, Ostrow DG, Detels R, et al. (1987) The Multicenter AIDS Cohort Study: rationale, organization, and selected characteristics of the participants. Am J Epidemiol. 126:310 8. Kilbourne AM, Justice AC, Rabeneck L, et al. (2001). General medical and psychiatric comorbidity among HIV infected veterans in the post HAART era. J Clin Epidemiol. 54 Suppl 1: S22 8. Koski L, Brouillette M J, Lalonde R, Hello B, Wong E, Tsuchida A, Fellows LK. Computerized testing augments pencil and paper tasks in measuring HIV associated mild cognitive impairment. HIV Medicine.2011;12:472 480 Launer LJ, Andersen K, Dewey ME, et al. (1999) Rates and risk factors for dementia and Alzheimer's disease: results from EURODEM pooled analyses. EURODEM Incidence R esearch Group and Work Groups. European Studies of Dementia. Neurology. 52:78 84 Launer LJ, Feskens EJ, Kalmijn S, Kromhout D (1996) Smoking, drinking, and thinking. The Zutphen Elderly Study. Am J Epidemiol. 143:219 27. Laurin D, Masaki KH, Foley DJ, et a l. (2004) Midlife dietary intake of antioxidants and risk of late life incident dementia: the Honolulu Asia Aging Study. Am J Epidemiol. 159:959 67 Levine AM, Seaberg EC, Hessol NA, et al. (2010) HIV as a risk factor for lung cancer in women: data from the J Clin Oncol. 28:1514 1519 Lifson AR and Lando HA (2012) Smoking and HIV: Prevalence, health risks, and cessation strategies. Curr HIV/AIDS Rep. 9;223 30. Lifson AR, Neuhaus J, Arribas JR, et al. (2010) Smoking related healt h risks among persons with HIV in the strategies for management of antiretroviral therapy council trial . Am J Public Health . 100;1896 906.

PAGE 120

120 Lloyd Richardson EE, Stanton CA, Papandonatos GD, et al. HIV positive smokers considering quitting: Differences by ra ce/ethnicity. Am Journal of Health Behavior. 2008;32(1)3 15. Lubin JH and Caporaso NE. (2006) Cigarette Smoking and lung cancer: modeling total exposure and intensity. Cancer Epidemiol Biomarkers Prev. 15; 517 523. Lucas GM, Mehta SH, Atta MG, et al. (200 7) End stage renal disease and chronic kidney disease in a cohort of African American HIV infected and at risk HIV seronegative participants followed between 1988 and 2004. AIDS. 21:2435 43. Mamary EM, Bahrs D, Martinez S. Cigaretts smoking and the desire to quit among individuals living with HIV. AIDS Patient Care and STDs. 2002;16(1):39 42. Mateen FJ and Mills EJ (2012) Aging and HIV related cognitive loss. JAMA . 308;349 50. Mckirnan DJ, Tolou Shams M, Turner L, et al. Elevated risk for tobacco use among men who have sex with men is mediated by demographic and psychosocial variables. Substance Use and Misuse. 2006;41:1197 1208 Miller EN, Selnes OA, McArthur JC, et al. (1990). Neuropsychological performance in HIV 1 infected homosexual men: The Multicenter AIDS Cohort Study (MACS). Neurology 40:197 203. Murray RP, Connett JE, Istvan JA, et al. (2002). Relations of cotinine and carbon monoxide to self report smoking in a cohort of smokers and ex smokers followed over 5 years. Nicotine Tob Res . 4:287 94. Ompan d DC, Kingdom M, Kupprat S, et al. Smoking and HIV related health issues among older HIV positive gay, bisexual, and other men who have sext with men. Behav Med. 2014;40:99 107 Ott A, Andersen K, Dewey ME, et al. (2004) Effect of smoking on global cognitiv e function in nondemented elderly. Neurology 62:920 4. Ott A, Slooter AJC, Hofman A, et al. (1998). Smoking and risk of dementia and Alzheimer's disease in a population based cohort study: the Rotterdam Study. Lancet. 351:1840 43. Pacek LR, Harrell PT, Mar tins SS. Cigarette smoking and drug use among a nationally representative sample of HIV positive individuals. Am J Addict. 2014;23:582 590. Pathan SS, Gottesman RF, Mosley TH, et al. (2011). Association of lung function with cognitive decline and dementia: the Atherosclerosis Risk in Communities (ARIC) Study. Eur J Neurol. 18:888 98. Petoumenos K, Worm S, Reiss P, et al. (2011). Rates of cardiovascular disease following smoking cessation in patients with HIV infection: results from the D:A:D study. HIV Med. 12:412 21.

PAGE 121

121 Platt A, Sloan FA, Costanzo P (2010). Alcohol consumption trajectories and associated characteristics among adults older than age 50. J Stud Alcohol Drugs. 71:169 179. Powell TM and Bagnell ME (2012). Dependen t Covariates. http://support.sas.com/resources/papers/proceedings12/168 2012.pdf . Reynolds NR, Neidig JL, Wewers ME. Illness representation and smoking behavior: A focus gro up study of HIV positive men. J Association of Nurses in AIDS Care. 2004;15(4)37 47. Reynolds N. Cigarette smoking and HIV: More evidence for action. AIDS educ Prev. 2009;21(3 suppl):106 121 Robinson WT, Brown MC, Moody Thomas S, et al. Smoking and experie nces with tobacco cessation among men who have sex with men: New Orleans, 2011. AIDS Behav. 2014;18 Suppl 3:325 32 Royce R, Winkelstein W. HIV infection, cigarette smoking and CD4+ T lymphocyte h Study. AIDS. 1990;4:327 333. Ryan H, Wortley PM, Easton A, et al. Smoking among lesbians, gays, and bisexuals: a review of the literature. Am J Prev Med. 2001;21:142 149 Sabia S, Elbaz A, Dugravot A, et al. (2012). Impact of smoking on cognitive decline in early old age: the Whitehall II cohort study. Arch Gen Psychiatry . 69:627 35. Sacktor N, Skolasky RL, Cox C, et al. (2010) Longitudinal psychomotor speed performance in human immunodeficiency virus seropositive individuals: impact of age and serostatus. J Neurovirol. 16:335 41. Schenker E, Hultin L, Bauer K, et al. Evaluation of a dual color flow cytometry immuphenotyping panel in a multicenter quality assurance program. Cytometry. 1993;14(3):307 317. Soulakova JN, Hartman AM, Liu B, et al. (2012) Relia bility of adult self reported smoking history: data from the tobacco use supplement to the current population survey 2002 2003 cohort. Nicotine Tob Res. 14:952 60. Shiels MS, Pfeiffer RM, Engels EA (2010). Age at cancer diagnosis among persons with AIDS in the United States. Ann Intern Med. 153:452 60. Silverberg MJ, Chao C, and Abrams DI (2009a). New insights into the role of HIV infection on cancer risk. Lancet Oncol. 10:1133 4. Silverberg MJ, Chao C, Layden WA (2009b). HIV infection and the risk of cance rs with and without a known infectious cause. AIDS. 23:2337 45.

PAGE 122

122 Singh Manoux A, Dugravot A, Kauffmann F, et al. (2011). Association of lung function with physical, mental and cognitive function in early old age. Age (Dordr) . 33:385 92. Skinner W. The preva lence and demographic predictors of illicit and licit drug use among lesbian and gay men. Am J Public Health. 1994;84:1307 1310. Skinner W and Otis M. Drug and alcohol use among lesbian and gay people in a southern US sample: epidemiological, comparative, and methodological findings from the Trilogy Project. J Homosex. 1996;30:59 91. Stall RD, Greenwood GL, Acree M, et al. Cigarette smoking among gays and bisexual men. Am J Public Health. 1999;89:1875 1878 Substance Abuse and Mental Health Services Administ ration (SAMHSA). Binge Drinking: Terminology and Patterns of Use. Available from: http://captus.samhsa.gov/access resources/binge drinking term inology and patterns use [Accessed February 15, 2014] Sulkowski MJ and Thomas DL (2003). Hepatitis C in the HIV infected person. Ann Intern Med. 138:197 207. Sweet L, Van Adel M, Metcalf V, Wright L, Harley A, Leiva R, Taler V. The Montreal Cognitive Asse ssment (MoCA) in geriatric rehabilitation: Psychometric properties and association with rehabilitation. International Psychogeriatrics . 2011 Tesoriero JM, Giervic SM, Carrascal A, Lavigne HE (2010) Smoking among HIV positive New Yorkers: Prevalence, frequen cy, and opportunities for cessation. AIDS Behav. 14;824 35. Mental State Examination and Montreal Cognitive Assessment in persons with mild subacute stroke: Relationship to functional outcome. Archives of Physical and Medical Rehabilitation. 2011; 92 :792 798. US Department of Health and Human Services, Office of Disease Prevention and Health Promotion. Healthy People 2020. Washington DC. Available from: http://www.healthypeople.gov/2020/default.aspx [Accessed September 1, 2014] Valcour VG, Shikuma CM, Watters MR, Sacktor NC (2004) Cognitive impairment in older HIV 1 seropositive individuals: prevalence and potential mechanisms. AIDS. 18 Suppl 1:S79 86. Valcour VG, Sithinamsuwan P, Letendre S, Ances B (2011a) Pathogenesis of HIV in the central nervous system. Curr HIV/AIDS Rep. 8:54 61. Valcour VG, Paul R, Chiao S, et al. (2011b) Screening for cognitive i mpairment in human immunodeficiency virus. Clin Infect Dis. 53:836 42. Vijayaraghavan M, Penko J, Vittinghoff E, et al. Smoking behaviors in a community based cohort of HIV infected indigent adults. AIDS Behav. 2014;18:535 543

PAGE 123

123 Vo QT, Cox C, Li X, et al. (20 13) Neuropsychological test performance before and after HIV 1 seroconversion: the Multicenter AIDS Cohort Study. J Neurovirol. 19:24 31. Ward B, Dahlhamer JM, Galinsky A, et al. Sexual orientation and health among US adults: National health interview surv ey, 2013. National Health Statistics Report. 2013;77 Available from: http://www.cdc.gov/nchs/data/nhsr/nhsr077.pdf [Accessed September 1, 2014] Weisskopf MG, Wright RO, Schwartz J, et al. (2004) Cumulative lead exposure and prospective change in cognition among elderly men: the VA Normative Aging Study. Am J Epidemiol 160:1184 93. Woods SP, Moore DJ, Weber E, et al. (2009) Cognitive neuropsychology of HIV associated neurocognitive disorders. Neuropsychol Rev 19:152 168. Wojna V, Robles L, Skolasky RL (2007) Associations of cigarette smoking with viral immune and cognitive function in human immunodeficiency virus seropositive women. J Neurovirol. 13:561 8. Womack JA, Goulet JL, Gibert C, et al. (2011) I ncreased risk of fragility fractures among HIV infected c ompared to uninfected male veterans. PLoS One. 6:e17217. Worm SW, Bower M, Reiss P, et al. Non AIDS defining cancers in the DAD study time trends and predictors of survival: A cohort study. BMC Infect Dis. 2014;13:471 Wright EJ, Grund B, Robertson K, et al . (2010) Cardiovascular risk factors associated with lower baseline cognitive performance in HIV positive persons. Neurology 75:864 73. Zou G. A modified poisson regression approach to prospective studies with binary data. Am J Epidemiol. 2004;159:702 706.

PAGE 124

124 BIOGRAPHICAL SKETCH Wajiha Akhtar received her Doctor of Philosophy from the Department of Epidemiology and G raduate C ertificate in G erontology in December 2015. She graduated from the University of Miami with a Bachelor of Science in b iomedical e ng ineering degree in 2008 and received her Master of Public Health degree from the University of Florida in 2010. During her time as a PhD student, she was a graduate research assistant for the Florida Office on Disability Health, teaching assistant, and an affiliated scholar with the Claude D. Pepper Older Americans Independence Center. Her research interests include HIV/AIDS, issues on aging, and quality of life.