We measured the trend of cigarette smoking among HIV-seropositive and seronegative men over time from 1984 to 2012. Additionally, we examined the demographic correlates of smoking and smoking consumption. Six thousand and five hundred and seventy seven men who have sex with men (MSM) from the Multicenter AIDS Cohort Study (MACS) were asked detailed information about their smoking history since their visit. Prevalence of smoking and quantity smoked was calculated yearly from 1984 to 2012. Poisson regression with robust error variance was used to estimate prevalence ratios of smoking in univariate and multivariate models. In 2012, 11.8 and 36.9 % of men who were enrolled in the MACS before 2001 or during or after 2001 smoked cigarettes, respectively. In the multivariate analysis, black, non-Hispanic, lower education, enrollment wave, alcohol use, and marijuana use were positively associated with current smoking in MSM. HIV serostatus was not significant in the multivariate analysis. However, HIV variables, such as detectable viral load, were positively associated. Though cigarette smoking has declined over time, the prevalence still remains high among subgroups. There is still a need for tailored smoking cessation programs to decrease the risk of smoking in HIV-seropositive MSM.

To examine the association between demographic characteristics and long-term smoking trajectory group membership among HIV-seropositive and HIV-seronegative men who have sex with men (MSM). A cohort of 6552 MSM from the Multicenter AIDS Cohort Study were asked detailed information about their smoking history since their last follow-up. Group-based trajectory modeling was used to examine smoking behavior and identify trajectory group membership. Because participants enrolled after 2001 were more likely to be younger, HIV-seronegative, non-Hispanic black, and have a high school diploma or less, we also assessed time of enrollment in our analysis. Participants were grouped into 4 distinct smoking trajectory groups: persistent nonsmoker (n = 3737 [55.9 %]), persistent light smoker (n = 663 [11.0 %]), heavy smoker to nonsmoker (n = 531 [10.0 %]), and persistent heavy smoker (n = 1604 [23.1 %]). Compared with persistent nonsmokers, persistent heavy smokers were associated with being enrolled in 2001 and later (adjusted odds ratio [aOR] 2.35; 95 % CI 2.12-2.58), having a high school diploma or less (aOR 3.22; 95 % CI 3.05-3.39), and being HIV-seropositive (aOR 1.17; 95 % CI 1.01-1.34). These associations were statistically significant across all trajectory groups for time of enrollment and education but not for HIV serostatus. The overall decrease of smoking as shown by our trajectory groups is consistent with the national trend. Characteristics associated with smoking group trajectory membership should be considered in the development of targeted smoking cessation interventions among MSM and people living with HIV.

Smoking is a potential risk factor for age-related cognitive decline. To date, no study has examined the association between smoking and cognitive decline in men living with human immunodeficiency virus (HIV). The aim of this present study is to examine whether smoking status and severity in midlife is associated with a rate of decline in cognitive processing speed among older HIV-seropositive and HIV-seronegative men who have sex with men. Data from 591 older HIV-seropositive and HIV-seronegative men who have sex with men from the Multicenter AIDS Cohort Study were examined. All participants had information on smoking history collected before age 50 years and at least 5 years of follow-up after age 50. Smoking history was categorized as never smoker, former smoker, and current smoker and cumulative pack years was calculated. The raw scores of three neuropsychological tests (Trail Making A, Trail Making B, and Symbol Digit Modalities tests) were log transformed (Trail Making A and B) and used in linear mixed models to determine associations between smoking history and at least subsequent 5-year decline in cognitive processing speed. There were no significant differences in the rates of neurological decline among never smokers, former smokers, and current smokers. Findings were similar among HIV-seropositive participants. However, an increase of 5 pack-years was statistically significantly associated with a greater rate of decline in the Trail Making Test B score and Composite Score (β -0.0250 [95% CI, -0.0095 to -0.0006] and -0.0077 [95% CI, -0.0153 to -0.0002], respectively). We found no significant association between smoking treated as a categorical variable (never smoked, former smoker, or current smoker) and a small change in every increase of 5 pack-years on measures of psychomotor speed and cognitive flexibility. To optimize healthy aging, interventions for smoking cessation should be tailored to men who have sex with men.