Smoking is the leading preventable cause of death and disease in the United States. HIV-infected individuals smoke tobacco at nearly 3 times the rate of the general population.1–7 These extraordinary smoking rates are associated with greater AIDS-related morbidity,8,9 greater non-AIDS–related morbidity (including cardiovascular disease, pulmonary disease, and non-AIDS cancers),7,10–15 and greater mortality.16,17 Smoking significantly impacts the progression and outcome of HIV disease, and has been identified as the leading contributor to premature mortality among people with HIV.18 In fact, a study from Denmark estimated that HIV-positive individuals lose more years from smoking than from HIV infection in and of itself.19 In our view, shared by others in the field,4,18,20–24 the single greatest health behavior change that could reduce mortality in this group is smoking cessation. This is well aligned with the personal goals of the target group as surveys have shown that most HIV-infected smokers want to quit.13,25–27 Thus, establishing the efficacy of interventions for HIV-infected smokers is of critical importance.28 Although many interventions are effective in helping general population smokers quit,21 research examining the outcomes of smoking cessation treatments for individuals with HIV is limited.29,30 Based on the alarming prevalence of smoking and the health-related consequences of smoking among HIV-infected smokers, the 2008 PHS Guideline for Treating Tobacco Use and Dependence specifically called for research that evaluates the efficacy of counseling, and targeted interventions among HIV-infected smokers.21 To our knowledge, there have been 8 randomized controlled trials evaluating the efficacy of behavioral interventions plus nicotine replacement targeting smoking cessation among people living with HIV/AIDS who smoke.1,31–37 Some trials demonstrated efficacy33,38 and several showed no effect.31,32,34,36,39 The purpose of this meta-analysis was to address the question whether behavioral interventions aimed at smoking cessation led to increased smoking abstinence as measured by expired carbon monoxide (ECO) 7-day point prevalence abstinence (PPA) rates compared with control interventions among people living with HIV who smoke.
English language articles from January 1980 through February 2016 were searched. The criteria for inclusion were randomized controlled trials of HIV-infected smokers with or without nicotine replacement therapy with a primary endpoint of carbon monoxide–verified 7-day point prevalence. The studies were restricted to smokers of any age with HIV or AIDS with self-reported cigarette use. Studies were excluded if they allowed participants to be enrolled in multiple smoking cessation studies that co-occurred with the study of interest, were not randomized control trials, or used an endpoint of urine nicotine and cotinine. There were no sample size limitations.
PubMed (1980 to 12/2014), Cochrane (1980 to 12/2014), CINHAL (1980-02/2016), PsychINFO (1980-02/2016), Google Scholar (1980-02/2016), and a literature review through bibliographies of related articles were searched.
The electronic search strategy was conducted on PubMed, CINHAL, PsychINFO, and Cochrane, using the following search terms: HIV, AIDS, immune compromised, smokers, behavioral intervention, NRT, nicotine patch, patch treatment, individual therapy, group therapy, motivational interviewing, person counseling, telephone counseling, brief counseling, counseling, behavior. Search filters were used to restrict studies to randomized controlled trials published in English and with human subjects. Google Scholar search included the key phrases such as HIV, randomized controlled trials, and smoking cessation. The bibliographies of included articles were searched for any additional studies.
Search results were reviewed for any duplicate studies. Two authors (A.K. and Y.D.) independently reviewed all study titles and abstracts for eligibility criteria. If studies met the eligibility criteria through title and abstract, the full article was reviewed. Thirteen full articles were reviewed, and 8 were selected. One study was excluded because the primary endpoint was of urine nicotine and cotinine. Four studies were excluded for not being randomized controlled trials. All the included used ECO-verified 7-day PPA rates for verification of abstinence. Cohen Kappa statistics that assesses the chance-corrected agreement between reviewers was 0.94. The PRISMA flow diagram (Fig. 1) summarizes the selection process.
The primary endpoint was ECO level–verified 7-day PPA.
Data Collection Process
Two authors (A.K. and Y.D.) independently extracted data from each of the 8 trials. All the data were reviewed for discrepancies. Any disagreement between the 2 reviewers was resolved by consensus.
Risk of Bias in Individual Studies
All included studies were independently reviewed by 2 reviewers and the validity and reliability were determined according to the Cochrane approach, including adequate sequence generation, allocation concealment, blinding, completeness of outcome data reporting, selective outcome reporting, and presence of other sources of bias.
Synthesis of Results
We calculated the relative risk (RR) ratios and the weighted pooled RR ratios across studies (Stata 12.0: metan command). We used the DerSimonian and Laird (random effects) model to provide weight estimates for each study (Fig. 2). We chose the random-effects model as it provides a more conservative estimate of weighting than the fixed effect (Mantel–Haenszel method) when one is concerned that the fixed-effects assumption, namely that the true effect is the same in each study, may not be met. The Q statistic and I2 statistic were used to evaluate heterogeneity. The Q statistic quantifies the magnitude of heterogeneity, whereas the I2 statistic quantifies the total variation due to between-study variance. Publication bias was not formally tested, as tests for publication bias (eg, funnel plot) may be too low to distinguish chance from real asymmetry when using meta-analytic techniques with 10 studies or less.22,40
The search of PubMed, Cochrane, CINHAL, PsychINFO, Google Scholar, and a bibliography literature review yielded a total of 17,384 citations after duplicate removal. A total of 17,371 were excluded after review of abstracts indicated eligibility criteria were not met. The remaining 13 full text articles were reviewed for eligibility. Four articles were excluded because they were not randomized controlled trials and 1 for a primary endpoint of urine nicotine and cotinine (Fig. 1).
A total of 8 studies with 1822 subjects were identified for inclusion in the review. All 8 studies were randomized controlled trials in the English language. All studies included nicotine replacement therapy either by prescription, distribution, or referral as a part of the intervention. The study interventions ranged from 4 to 11 sessions, and 5 of the 8 studies had at least 8 sessions.33,34,36–38 The studies delivered the behavioral intervention in a range of formats. Three studies used telephone counseling.36,38,41 One used group therapy.33 Two used computer-based interventions,31,34 and 3 used individual therapy.31,32,35 The primary endpoints for all 8 studies were ECO-verified 7-day PPA. All studies compared the intervention group to brief counseling or self-help control conditions (Table 1).
Risk of Bias Within Studies
All 8 trials were at low risk for bias because of inadequate sequence generation or incomplete data reporting of the primary outcome. There is, however, risk for publication bias. A funnel plot was constructed and by inspection there was no suggestion of publication bias; however, in a meta-analysis with less than 10 studies, funnel plots may not be reliable.40 Since these were behavioral studies, the nature of the interventions precluded blinding and allocation concealment.
Results of Individual Studies
Three of the 8 clinical trials reviewed reported treatment efficacy. Wewers et al37 found statistically significant increase in abstinence at 8 weeks comparing those randomized to the nurse managed, peer led, smoking cessation intervention (n = 8) with the control condition (n = 7) (67.5% vs. 0%). Vidrine et al38 found a statistically significant increase in abstinence at 3 months among those randomized to the telephone intervention group (n = 48) compared with usual care control (n = 47) (36.8% vs. 10.3%). A larger study conducted by Vidrine et al36 found a statistically significant increase in abstinence at 3 months for those randomized to the telephone group (n = 236) compared with the control group (n = 238) (11.9% vs. 3.4%).
Five of the 8 clinical trials reviewed did not find treatment efficacy.31–35 Lloyd-Richardson et al32 found no difference in abstinence at 6 months among those randomized to the motivational enhancement arm (n = 232) compared with those in standard of care (n = 212) (9% vs. 10%).
Humfleet et al31 found no difference in abstinence after 52 weeks among those randomized to individual counseling (n = 69) or computer-based intervention (n = 58) compared with brief counseling control (n = 82) (20.4% vs. 25.6% vs. 19.8%). Shuter et al34 found no difference in abstinence at 3 months among those randomized to web-based intervention (n = 69) compared with standard-of-care control (n = 69) (10.3% vs. 4.3%). Moadel et al33 found no difference in abstinence at 3 months among those randomized to group therapy (n = 72) compared with the control condition (n = 73) (19.2 vs. 9.7%). Stanton et al35 found no evidence that 4 in-person sessions (n = 131) improved cessation rates compared with the standard of care (n = 131) at 3 months (8.5% vs. 9.1%) (Table 1).
Synthesis of Results
The meta-analysis of the 8 studies evaluating the efficacy of behavioral interventions to increase smoking abstinence among HIV-infected smokers yielded a moderate statistically significant effect size for abstinence (RR: 1.51; 95% CI: 1.17 to 1.95). Heterogeneity of the studies was assessed and found to be moderate (I2 = 43.3%; χ2 = 14.1, d.f. = 8; P = 0.08). When stratified by the number of sessions, those studies with ≥8 sessions had a large statistically significant effect on abstinence rates (RR: 2.88; 95% CI: 1.89 to 4.61) and no heterogeneity (I2 = 0.0%; χ2 = 1.93; d.f. = 4; P = 0.749). The studies with fewer than 8 sessions or less had nonsignificant results (RR: 1.05; 95% CI: 0.77 to 1.44) and no heterogeneity (I2 = 0.0%; χ2 = 1.27, d.f. = 3; P = 0.737) (Fig. 2).
A sensitivity analysis was conducted to better quantify the effect of the study by Wewer et al on the overall weighted effect of the stratified sample. In the stratified analysis (ie, studies with 8 sessions or more vs. studies with less than 8 sessions), the overall effect including the Wewer's study was RR 2.88 (1.80–4.61), whereas the overall effect excluding the Wewer's study was an RR 2.70 (1.67–4.37).
Our systematic review and meta-analysis of the 8 published randomized controlled trials of behavioral tobacco-treatment interventions targeting HIV-infected smokers demonstrated a significant effect in terms of increased abstinence. When stratified by the total number of sessions, studies with 8 or more sessions had a large statistically significant effect size for abstinence, whereas those studies with fewer than 8 sessions yielded nonsignificant results. There was no heterogeneity found with stratification by the total number of sessions. These results suggest that providing a greater number of sessions may be an important determinant of programmatic efficacy.
With respect to risk of bias, it is important to note that all the studies in the meta-analysis were deemed to be of high quality. They were all randomized controlled trials with adequate sequence generation and free of selective outcome reporting or incomplete outcome reporting in the primary measure.31–38 The mode of delivery of the behavioral interventions used in these studies differed between trials. Three studies used telephone counseling.36,38,41 One used group therapy.33 Two used computer-based interventions,31,34 and 3 used individual therapy.31,32,35 Even so, the heterogeneity of the studies was assessed and found to be moderate to low.
Although web-based programs have the advantage of flexibility in scheduling, low cost, and the ability to review material as needed, the 2 web-based studies in this meta-analysis had nonsignificant results. Both web-based programs used software that was at the sixth grade reading level. An important difference between the studies was that Shuter et al34 excluded participants with low literacy scores, whereas Humfleet et al31 did not. The former study found that a higher educational level was associated with more web pages visited and higher quit rates. This suggests that literacy or perhaps experience using web-based interventions (ie, as a result of the digital divide), which are both associated with higher education, may be important predictors of success with this type of intervention.
The advantage of peer-based interventions is that they may bridge potential barriers related to language, culture, and class, and thereby enhance treatment delivery, and they have proved promising in other areas of research.37,42,43 One of 2 studies that used peer support specialists yielded statistically significant large effect sizes, suggesting that this may be a particularly promising way of delivering targeted behavioral interventions in the future.33,37
Of note is the finding that the 2 studies that did not assess readiness to quit as a part of their inclusion criteria were also studies that found no treatment effect.31,32 This may be because of a greater number of participants who were either in precontemplation and/or contemplation and may have negatively contributed to abstinence rates in these trials. The implication of this finding may be that future trials will need to evaluate level of treatment readiness as an eligibility criterion for smoking cessation intervention.
Although this meta-analysis was limited to English language articles (ie, Tower of Babel Bias),44 it still appears that based on these data behavioral interventions targeted at the HIV-infected smokers had a significant effect on increasing smoking abstinence. When stratified by the length of intervention, those studies with interventions of 8 sessions or more had a statistically significant large effect size for abstinence, whereas those studies with interventions consisting of fewer than 8 sessions had nonsignificant results. Future studies may need to include eligibility criteria that emphasize readiness to quit as well and provide ≥8 intervention sessions. For those studies that are web-based assessment of literacy may be needed. Further study on the utility of peers is an emerging area of research that shows considerable promise.
1. Burkhalter JE, Springer CM, Chhabra R, et al. Tobacco use and readiness to quit smoking in low-income HIV-infected persons. Nicotine Tob Res. 2005;7:511–522.
2. Kwong J, Bouchard-Miller K. Smoking cessation for persons living with HIV: a review of currently available interventions. J Assoc Nurses AIDS Care. 2010;21:3–10.
3. Lifson AR, Lando HA. Smoking and HIV: prevalence, health risks, and cessation strategies. Curr HIV/aids Rep. 2012;9:223–230.
4. Mamary EM, Bahrs D, Martinez S. Cigarette smoking and the desire to quit among individuals living with HIV. AIDS Patient Care STDS. 2002;16:39–42.
5. Moscou-Jackson G, Commodore-Mensah Y, Farley J, et al. Smoking-cessation interventions in people living with HIV infection: a systematic review. J Assoc Nurses AIDS Care. 2014;25:32–45.
6. Niaura R, Shadel WG, Morrow K, et al. Human immunodeficiency virus infection, AIDS, and smoking cessation: the time is now. Clin Infect Dis. 2000;31:808–812.
7. Palella FJ Jr, Phair JP. Cardiovascular disease in HIV infection. Curr Opin HIV AIDS. 2011;6:266–271.
8. Feldman JG, Minkoff H, Schneider MF, et al. Association of cigarette smoking with HIV prognosis among women in the HAART era: a report from the women's interagency HIV study. Am J Public Health. 2006;96:1060–1065.
9. Nieman RB, Fleming J, Coker RJ, et al. The effect of cigarette smoking on the development of AIDS in HIV-1-seropositive individuals. AIDS. 1993;7:705–710.
10. Akgun KM, Gordon K, Pisani M, et al. Risk factors for hospitalization and medical intensive care unit (MICU) admission among HIV-infected Veterans. J Acquir Immune Defic Syndr. 2013;62:52–59.
11. Calvo-Sanchez M, Perello R, Perez I, et al. Differences between HIV-infected and uninfected adults in the contributions of smoking, diabetes and hypertension to acute coronary syndrome: two parallel case-control studies. HIV Med. 2013;14:40–48.
12. De P, Farley A, Lindson N, et al. Systematic review and meta-analysis: influence of smoking cessation on incidence of pneumonia in HIV. BMC Med. 2013;11:15.
13. Metersky ML, Colt HG, Olson LK, et al. AIDS-related spontaneous pneumothorax. Risk factors and treatment. Chest. 1995;108:946–951.
14. Saves M, Chene G, Ducimetiere P, et al. Risk factors for coronary heart disease in patients treated for human immunodeficiency virus infection compared with the general population. Clin Infect Dis. 2003;37:292–298.
15. Winstone TA, Man SF, Hull M, et al. Epidemic of lung cancer in patients with HIV infection. Chest. 2013;143:305–314.
16. Lifson AR, Neuhaus J, Arribas JR, et al. Smoking-related health risks among persons with HIV in the strategies for management of antiretroviral therapy clinical trial. Am J Public Health. 2010;100:1896–1903.
17. Crothers K, Goulet JL, Rodriguez-Barradas MC, et al. Impact of cigarette smoking on mortality in HIV-positive and HIV-negative veterans. AIDS Educ Prev. 2009;21:40–53.
18. Davies TL, Gompels M, Johnston S, et al. Mind the gap: difference between Framingham heart age and real age increases with age in HIV-positive individuals-a clinical cohort study. BMJ Open. 2013;3:e003245.
19. Helleberg M, Afzal S, Kronborg G, et al. Mortality attributable to smoking among HIV-1-infected individuals: a nationwide, population-based cohort study. Clin Infect Dis. 2013;56:727–734.
20. Grinspoon SK, Grunfeld C, Kotler DP, et al. State of the science conference: initiative to decrease cardiovascular risk and increase quality of care for patients living with HIV/AIDS: executive summary. Circulation. 2008;118:198–210.
21. 2008 PHS Guideline Update Panel, Liaisons, and Staff. Treating tobacco use and dependence: 2008 update U.S. public health service clinical practice Guideline executive summary. Respir Care. 2008;53:1217–1222.
23. Lundgren JD, Battegay M, Behrens G, et al. European AIDS Clinical Society (EACS) guidelines on the prevention and management of metabolic diseases in HIV. HIV Med. 2008;9:72–81.
24. Niaura R, Chander G, Hutton H, et al. Interventions to address chronic disease and HIV: strategies to promote smoking cessation among HIV-infected individuals. Curr Hiv/aids Rep. 2012;9:375–384.
25. Drach L, Holbert T, Maher J, et al. Integrating smoking cessation into HIV care. AIDS Patient Care STDS. 2010;24:139–140.
26. Fuster M, Estrada V, Fernandez-Pinilla MC, et al. Smoking cessation in HIV patients: rate of success and associated factors. HIV Med. 2009;10:614–619.
27. Shuter J, Bernstein SL, Moadel AB. Cigarette smoking behaviors and beliefs in persons living with HIV/AIDS. Am J Health Behav. 2012;36:75–85.
28. Browning KK, Wewers ME, Ferketich AK, et al. Tobacco use and cessation in HIV-infected individuals. Clin Chest Med. 2013;34:181–190.
29. Hoffman AC, Starks VL, Gritz ER. The impact of cigarette smoking on HIV/AIDS: urgent need for research and cessation treatment. AIDS Educ Prev. 2009;21:1–2.
30. Nahvi S, Cooperman NA. Review: the need for smoking cessation among HIV-positive smokers. AIDS Educ Prev. 2009;21:14–27.
31. Humfleet GL, Hall SM, Delucchi KL, et al. A randomized clinical trial of smoking cessation treatments provided in HIV clinical care settings. Nicotine Tob Res. 2013;15:1436–1445.
32. Lloyd-Richardson EE, Stanton CA, Papandonatos GD, et al. Motivation and patch treatment for HIV+ smokers: a randomized controlled trial. Addiction. 2009;104:1891–1900.
33. Moadel AB, Bernstein SL, Mermelstein RJ, et al. A randomized controlled trial of a tailored group smoking cessation intervention for HIV-infected smokers. J Acquir Immune Defic Syndr. 2012;61:208–215.
34. Shuter J, Morales DA, Considine-Dunn SE, et al. Feasibility and preliminary efficacy of a web-based smoking cessation intervention for HIV-infected smokers: a randomized controlled trial. J Acquir Immune Defic Syndr. 2014;67:59–66.
35. Stanton CA, Papandonatos GD, Shuter J, et al. Outcomes of a tailored intervention for cigarette smoking cessation among Latinos living with HIV/AIDS. Nicotine Tob Res. 2015;17:975–982.
36. Vidrine DJ, Marks RM, Arduino RC, et al. Efficacy of cell phone-delivered smoking cessation counseling for persons living with HIV/AIDS: 3-month outcomes. Nicotine Tob Res. 2012;14:106–110.
37. Wewers ME, Neidig JL, Kihm KE. The feasibility of a nurse-managed, peer-led tobacco cessation intervention among HIV-positive smokers. J Assoc Nurses AIDS Care. 2000;11:37–44.
38. Vidrine DJ, Arduino RC, Lazev AB, et al. A randomized trial of a proactive cellular telephone intervention for smokers living with HIV/AIDS. AIDS. 2006;20:253–260.
39. Manuel JK, Lum PJ, Hengl NS, et al. Smoking cessation interventions with female smokers living with HIV/AIDS: a randomized pilot study of motivational interviewing. AIDS Care. 2013;25:820–827.
40. Ioannidis JP, Trikalinos TA. The appropriateness of asymmetry tests for publication bias in meta-analyses: a large survey. CMAJ. 2007;176:1091–1096.
41. Tesoriero JM, Gieryic SM, Carrascal A, et al. Smoking among HIV positive New Yorkers: prevalence, frequency, and opportunities for cessation. AIDS Behav. 2010;14:824–835.
42. Napoles AM, Ortiz C, Santoyo-Olsson J, et al. Nuevo Amanecer: results of a randomized controlled trial of a community-based, peer-delivered stress management intervention to improve quality of life in Latinas with breast cancer. Am J Public Health. 2015;105(suppl 3):e55–63.
43. Srinivas GL, Benson M, Worley S, et al. A clinic-based breastfeeding peer counselor intervention in an urban, low-income population: interaction with breastfeeding attitude. J Hum Lact. 2015;31:120–128.
44. Gregoire G, Derderian F, Le LJ. Selecting the language of the publications included in a meta-analysis: is there a tower of Babel bias? J Clin Epidemiol. 1995;48:159–163.