Receipt of Smoking Cessation Medications Among People With and Without Human Immunodeficiency Virus in the Veterans Aging Cohort Study (2003–2018)

Abstract

Background

Nicotine replacement therapy, bupropion, and varenicline are smoking cessation medications (SCMs) shown to be similarly effective in people with and without human immunodeficiency virus (PWH and PWoH, respectively), although rates of receipt of these medications are unknown.

Methods

We identified patients in the Veterans Aging Cohort Study with electronic health record-documented current smoking using clinical reminder data for tobacco use (2003–2018). We measured receipt of SCMs using Veterans Affairs pharmacy data for outpatient prescriptions filled 0–365 days after current smoking documentation. We used log-linear, Poisson-modified regression models to evaluate the relative risk (RR) for receiving SCM by human immunodeficiency virus (HIV) status, the annual rate of receipt, and rate difference among PWH relative to PWoH.

Results

The sample included 92 632 patients (29 086 PWH), reflecting 381 637 documentations of current smoking. From 2003 to 2018, the proportion receiving SCMs increased from 15% to 34% for PWH and from 17% to 32% among PWoH. There was no statistical difference in likelihood of receiving SCM by HIV status (RR, 1.010; 95% confidence interval [CI], .994–1.026). Annual rates of receiving SCM increased for PWH by 4.3% per year (RR, 1.043; 95% CI, 1.040–1.047) and for PWoH by 3.7% per year (RR, 1.037; 95% CI, 1.036–1.038; rate difference +0.6% [RR, 1.006; 95% CI, 1.004–1.009]).

Conclusions

In a national sample of current smokers, receipt of SCM doubled over the 16-year period, and differences by HIV status were modest. However, fewer than 35% of current smokers receive SCM annually. Efforts to improve SCM receipt should continue for both groups given the known dangers of smoking.

Receipt of smoking cessation medications doubled among patients with and without HIV in the VA between 2003 and 2018. Although annual rates of receiving medications were comparable for both populations, there is considerable room for improvement.

Cigarette smoking is the leading cause of morbidity and mortality worldwide and is highly prevalent among and particularly dangerous for people with human immunodeficiency virus (PWH) [1]. Greater than 40% of PWH currently smoke [2, 3] in contrast to 14% among the general US population [4]. Smoking cessation medications, such as nicotine replacement therapy (NRT), bupropion, and varenicline, are effective prescription pharmacotherapies approved by the US Food and Drug Administration (FDA) for treatment of tobacco use disorder [5]. They have been shown to promote smoking cessation among PWH comparably to studies conducted among those without human immunodeficiency virus (PWoH) [6–8]. However, recent evidence suggests that there may be human immunodeficiency virus (HIV)-related disparities in receiving evidence-based pharmacologic treatments [9]. This is concerning because cumulative health impacts of tobacco use and HIV leave PWH at greater susceptibility and severity of chronic illnesses such as cardiovascular disease [10, 11], chronic obstructive pulmonary disease [12, 13], and cancer [14, 15] relative to PWoH.

There have been multiple changes in recommendations for use of smoking cessation medications for tobacco cessation over the last several decades. The US Public Health Service introduced Clinical Practice Guidelines for providers in 2000 to prescribe at least 1 smoking cessation medication in the form of NRT or bupropion in the absence of contraindications [16]. These were updated in 2008 to (1) include varenicline, (2) recommend administering 2 smoking cessation medications in combination with counseling, and (3) represcribe medications after relapse [5]. Although pharmacotherapy has been the standard care in treating tobacco use disorder for the past 2 decades, it is unknown how rates of receiving NRT, bupropion, and varenicline have changed among PWH compared with PWoH.

The Veterans Health Administration (VHA) is the nation's largest provider of HIV care and home to the Veterans Aging Cohort Study (VACS). The VACS is a multisite, longitudinal, prospective cohort study of PWH and controls who are HIV-uninfected assembled from national VHA data. In the VHA—where veterans with HIV report higher rates of tobacco use than those without HIV and non-veterans [17]—tobacco use reduction is a priority. In 2003, the VHA issued guidance to remove restrictions on prescribing NRT and/or bupropion to increase access to these highly effective pharmacologic treatments for all patients who smoked [18].

Therefore, the VHA offers an important opportunity to study differences in receipt of smoking cessation medications between PWH and PWoH longitudinally over a period marked by policy and clinical changes. We used VACS data to assess likelihood of receiving these medications, annual rates of receipt, and rate differences among PWH relative to a PWoH over a 16-year period in a national sample of VHA patients.

METHODS

Study Data Sources and Sample

In this study, we used VACS data, drawn from national VHA data beginning in 1997. The VACS includes all PWH receiving care in the VHA matched 1:2 to PWoH by 5-year blocks of age, race/ethnicity, sex, and site of care [19, 20]. The cohort is updated annually to include veterans with new diagnoses of HIV infection and matched controls receiving care in the same fiscal year.

For the present study, we identified VACS patients with documentation of current smoking in the electronic health record (EHR) resulting from an automated clinical reminder for tobacco use between January 1, 2003 and December 31, 2018. Clinical reminders are embedded in the EHR and serve to prompt providers to ask their patients about tobacco use. These data have been collected nationally since 1999 and have substantial agreement with smoking status based on survey response [21]. Current smoking status was assigned yearly based on the first EHR-documentation of current smoking in each calendar year. For patients contributing multiple observations, we limited the sample to 1 observation per person per year and included subsequent observations if they occurred at least 365 days after the one prior.

We linked these data to demographic and pharmacy information extracted from the VHA Corporate Data Warehouse (CDW). The CDW is a national repository that incorporates data from the clinical and administrative systems into 1 standard database structure for all users of VHA care [22].

Outcomes

The primary outcome of interest was a dichotomous measure identifying receipt of any smoking cessation medications (NRT, bupropion, and varenicline), based on outpatient prescriptions filled from VHA pharmacy data in the EHR 0–365 days after the date documenting current smoking. This outcome measure was generated for each observation in which there was a positive tobacco screen.

To distinguish bupropion used to treat tobacco use disorder rather than severe depression, we dropped prescribed dosages greater than or equal to 200 mg. However, it was not possible to identify whether bupropion was used to treat both conditions among our cohort.

In secondary outcomes, we measured receipt of each individual medication 0–365 days after the date of each positive tobacco screen.

Primary Independent Measure

HIV status was defined as having 2 outpatient or 1 inpatient diagnostic code from the International Classification of Disease, Ninth Revision, Clinical Modification for either acquired immune deficiency syndrome (042) or HIV (V08), consistent with prior research conducted in the VHA [19]. For patients contributing multiple observations to the data, we updated HIV status accordingly.

Moderator

To reflect changes over time in policies and clinical recommendations regarding smoking cessation medications, we defined time as a continuous measure represented by calendar year. We assigned each observation to the calendar year in which it was documented.

Covariates

Age at time of each observation (continuous), sex (male versus female), race/ethnicity (Black, Hispanic, other, and White), and site of care were collected via the EHR for each patient. “Other” race included American Indian or Alaska Native, Asian, and Native Hawaiian or Pacific Islander.

Analysis

We summarized characteristics of PWH and PWoH as individuals and at each observation of the study period. We then compared prevalence of primary and secondary outcomes among all observations by HIV status in bivariate analyses.

In regression analyses, generalized log-linear regression with a modified Poisson working model were used to estimate the relative risk (RR) of receiving smoking cessation medications and each individual medication with 95% confidence interval (CI) [23]. Because we allowed patients to contribute to multiple years (given tobacco screens at least 1 year apart), we clustered by patient and specified an exchangeable correlation structure to account for repeated measures. We also applied a robust sandwich estimator to calculate standard errors, obtain a valid inference for a large sample and common outcome, and account for misspecification of the correlation structure.

The inclusion of time in the model allowed us to estimate the relative risk of receiving smoking cessation medications in one calendar year compared with the previous calendar year (annual rate of receipt) among PWoH, holding all other variables constant. Finally, we included the HIV*year interaction term to estimate the annual rate difference among PWH relative to PWoH. A significant P value (<.05) for the interaction term would indicate that the difference in annual rate of receipt between the 2 populations is statistically different from zero.

Patient Consent Statement

This study was approved by the Institutional Review Boards of Yale University (0309025943) and VA Connecticut Healthcare System (AJ0001). It has been granted a waiver of informed consent and is compliant with the Health Insurance Portability and Accountability Act.

RESULTS

Between January 1, 2003 and December 31, 2018, 92 632 patients had documentation of current smoking, reflecting 381 637 observations (1–14 per patient) over the 16-year period. Among all patients, 29 086 had documented HIV (31%), contributing 109 473 observations to the analyses. The median number of observations contributed per individual was greater among PWoH than PWH (median number 4 versus 3, respectively; P < .001).

A summary of patient characteristics of individuals contributing to the data is presented in Table 1, both overall and by HIV status. The population was predominantly male (97.4%) with an average age of 49 years at the time of their first observation. Approximately 50% were Black, reflecting the composition of PWH who smoke in the VHA. People with HIV were more likely to be younger and female, and less likely to have received smoking cessation medications during the study period compared with PWoH, although differences were numerically small. Comparing individual medications, 41% and 43% of PWH and PWoH, respectively, had received NRT (P < .001); 21% and 24% had received bupropion (P < .001); and 5.3% and 5.2% had received varenicline (P = .43) (Table 1). For detailed demographic information of patients contributing observations for each year of the study period, see Appendix 1.

Table 1.

Characteristics Among VACS Participants Who Currently Smoke by Year and HIV Status (N = 92 632), Between 2003 and 2018

Characteristic	Total Population (N = 92 632)	HIV+ (n = 29 086)	HIV− (n = 63 546)	P Value
HIV infected, %	31.4
Age, mean (SD)a	49.3 (10.4)	49.1 (10.3)	49.4 (10.5)	<.001
Male, %	97.4	97.1	97.6	<.001
Race, %Black	49.9	50.3	49.7	.08
Hispanic	7.4	7.4	7.4
Other	3.1	3.3	3.1
White	39.6	39.1	39.9
Number of observations during the study period, median (IQR)	4 (2–6)	3 (2–6)	4 (2–6)	<.001
Received smoking cessation medication during the study period, %	50.2	48.7	50.9	<.001
NRT	42.6	41.5	43.1	<.001
Bupropion	23.2	21.0	24.2	<.001
Varenicline	5.2	5.3	5.2	.43

Abbreviations: HIV, human immunodeficiency virus; IQR, interquartile range; NRT, nicotine replacement therapy; SD, standard deviation; VACS, Veterans Aging Cohort Study.

^aAge reported at first observation for participants contributing more than 1 observation.

Among all observations of current smoking between 2003 and 2018, the proportion receiving any smoking cessation medications within 1 year of reporting tobacco use increased from 15% to 34% for PWH and from 17% to 32% among those uninfected (Figure 1). In bivariate comparisons, the prevalence of receiving any medication was similar among PWH and PWoH (28% for both; P = .06). However, there were small, yet statistically significant, differences with regards to specific medications. The proportion that received NRT was greater among PWH than PWoH (21% vs 20%, respectively; P < .001). The proportion that received varenicline was also greater among PWH than PWoH (1.8 vs 1.6, respectively; P < .001), whereas the opposite was found in regards to bupropion (9% vs 10%, respectively; P < .001) (Table 2).

Figure 1.

Proportion of Veterans Aging Cohort Study participants who currently smoke and received smoking cessation medication by human immunodeficiency virus (HIV) status by year (2003–2018): any and individual medications. Abbreviation: NRT, nicotine replacement therapy.

Table 2.

Bivariate Association and Adjusted Relative Risk Between HIV Status and Receipt of Smoking Cessation Medication Among VACS Participants Who Currently Smoke (N = 92 632), Between 2003 and 2018

Bivariate Association and Adjusted Relative Risk of Receiving SCM Among All Observations of Current Smoking					Yearly Change in Rates of Receiving SCM Among All Observations of Current Smoking
Smoking Cessation Medication	HIV+ (n = 109 473)a	HIV− (n = 272 164)a	P Value	RR Comparing HIV+ and HIV− (95% CI)b	RR per Year for HIV− (95% CI)b	RR per Year for HIV+ (95% CI)c	RR Difference for HIV+ (95% CI)d
Any medication received	28.0%	27.7%	.06	1.010 (0.994–1.026)	1.037 (1.036–1.038)	1.043 (1.040–1.047)	1.0065 (1.0038–1.0092)
Nicotine replacement therapy	21.0%	20.3%	<.001	1.033 (1.015–1.052)	1.052 (1.050–1.053)	1.055 (1.050–1.059)	1.0028 (0.9997–1.0060)
Bupropion	9.4%	10.1%	<.001	0.921 (0.893 0.950)	0.993 (0.991–0.996)	1.001 (0.994–1.010)	1.0080 (1.0026–1.0135)
Varenicline	1.8%	1.6%	<.001	1.159 (1.086–1.237)	1.087 (1.081–1.093)	1.115 (1.097–1.133)	1.0281 (1.0158–1.0405)

Abbreviations: CI, confidence interval; HIV, human immunodeficiency virus; RR, relative risk; VACS, Veterans Aging Cohort Study.

^an = number of observations.

^bModel includes HIV status and year only.

^cAddition of RR for HIV⁻ and RR difference for HIV⁺.

^dModel includes HIV status, year, and interaction term (HIVxYear).

Similarly, regression models showed no statistical difference in likelihood of receiving medications between PWH and PWoH (RR, 1.010; 95% CI, .994–1.026). People with HIV were 3.3% more likely to receive NRT (RR, 1.033; 95% CI, 1.015–1.052) and 15.9% more likely to receive varenicline (RR, 1.159; 95% CI, 1.086–1.237), but 7.9% less likely to receive bupropion (RR, 0.921; 95% CI, .893–.950) relative to PWoH.

Rate of Receiving Smoking Cessation Medications Over Time by HIV Status

Over the study period, receipt of smoking cessation medications increased by 4.3% and 3.7% per year for PWH and PWoH, respectively (RR, 1.043; 95% CI, 1.040–1.047; RR, 1.037; 95% CI, 1.036–1.038). The rate difference among PWH was +0.6% greater relative to PWoH (RR, 1.006; 95% CI, 1.004–1.009) (Table 2).

Receipt of NRT increased at a similar rate of 5.2% per year across populations (RR, 1.052; 95% CI, 1.050–1.053).

The annual rate of receiving bupropion was not statistically different from zero for PWH (RR, 1.001; 95% CI, .994–1.010) and decreased 0.7% per year for PWoH (RR, 0.993; 95% CI, .991–.996). The rate difference among PWH was +0.8% greater relative to PWoH (RR, 1.008; 95% CI, 1.003–1.014).

Receipt of varenicline increased by 11.5% and 8.7% per year for PWH and PWoH, respectively (RR, 1.115; 95% CI, 1.097–1.133; RR, 1.087; 95% CI, 1.081–1.093). The rate difference among PWH was +2.8% greater relative to PWoH (RR, 1.028; 95% CI, 1.016–1.040).

DISCUSSION

In this national sample of VHA patients with and without HIV who reported current smoking, receipt of smoking cessation medications nearly doubled during the 2003–2018 study period for both PWH and PWoH. The majority of this increase appears to be attributable to improved rates of NRT receipt, in particular. Rates of receiving bupropion were similar over the study period for PWH and slightly decreased for PWoH, but we observed modest gains in receipt of varenicline for both populations. Nevertheless, despite efforts nationally and by the VHA specifically, fewer than 35% of all patients who smoke received these medications within 1 year of reporting current smoking. Although differences between PWH and PWoH in the rate of increases over time and ultimate rates of receipt were modest, PWH may be a group in need of specific targeting of smoking cessation medications given the enhanced dangers of smoking for PWH relative to PWoH.

The VHA has made important progress in implementing several system-wide interventions to improve patient access to evidence-based tobacco cessation care and pharmacotherapy. The marked increase of smoking cessation medication receipt among VHA patients is likely to reflect the adoption of population-based approaches to tobacco use, such as increased accessibility to NRT (2003), revision of clinical practice guidelines to integrate smoking cessation into primary care (2004), further integration into psychiatric settings (2006), and implementation of computerized clinical reminders to document smoking and performance measures (2009) [18]. Successful performance on VHA's performance measure requires annual documentation that each smoker be given brief counseling, offered medications, and provided a referral for additional services. At the end of the first year of the new performance measure, 75% of current smokers seen in mental health settings were offered smoking cessation medications to assist with quitting as were 71% of those in nonmental health settings [18]. These adaptations of policy and practice regarding smoking cessation have been shown to be effective in increasing access and receipt of smoking cessation medications, especially NRT. Our findings support that rates of receiving NRT improved most drastically between 2003 and 2009 as NRT became central to VHA smoking cessation efforts, which then expanded to primary care and mental healthcare settings.

Although considered another first-line therapy, bupropion receipt was similar over time among PWH and slightly fell among PWoH. Among PWH, there are several potential interactions between antiretroviral medications and bupropion. For example, ritonavir, lopinavir, tipranavir, and efavirenz can significantly alter plasma concentrations of bupropion [24]. In addition, outside of HIV, bupropion is contraindicated for patients experiencing or being treated for certain mental health and cardiovascular conditions [25, 26]. Therefore, it is possible that a sizeable portion of patients included in this study may experience barriers to receiving bupropion. Despite recent findings that documented mental health conditions were associated with greater likelihood of receiving pharmacotherapy among PWH and demographically similar PWoH [9, 27], it is unsurprising that NRT continues to be the cessation medication most received.

Certain VHA policies may also limit access of effective smoking cessation treatment in their formulary. This is especially true regarding varenicline and may contribute to its low rates of receipt. Specifically, our study demonstrates that varenicline begins to be received when the FDA approved it for treatment of tobacco use disorder. Over the study period, receipt of varenicline reached its peak in 2007 and then fell sharply in 2008 (4.5% among PWH and 4.8% among PWoH to 2.9% and 2.8%) after the FDA issued a warning for its use. The warning explicitly noted that changes in mood, behavior, and suicidal ideation had been reported in patients attempting to quit smoking while using varenicline [28]. For this reason and situations specifically involving veterans, the VHA previously considered varenicline a second-line treatment and had strictly controlled its use during the study period. Several studies show greater efficacy of varenicline over placebo, bupropion, and single as well as combination NRT in achieving smoking cessation and have demonstrated safety [29–31]. However, provider concerns about its neuropsychiatric effects may persist to the detriment of providing what is widely considered to be the most effective medication for tobacco use disorder.

Despite great progress over the last 16 years, fewer than 35% of current smokers in our study received smoking cessation medications in any given year, and approximately 50% received them at least once over the entire study period. Although adoption of population health approaches to address tobacco use disorder have shown some success, there is considerable room for improvement. Furthermore, efforts may additionally be needed to target PWH due to their increased susceptibility to tobacco-related illnesses. It is notable that PWoH contributed more observations to the study data than PWH, suggesting that they either had more frequent clinic visits or were asked their smoking status through clinical reminder more often than PWH, despite being otherwise demographically comparable. Continued efforts addressing current smoking among PWH are needed. People with HIV represent a vulnerable population at higher “risk of risks” due to smoking because of compounded effects of health factors, putting them at high risk of disease, and social factors, placing them in positions that perpetuate disparities [32, 33]. These include but are not limited to discrimination, stigmatization, and loneliness [34], as well as low income and education levels [35, 36]. In fact, despite comparable and sometimes greater utilization of cessation treatment, higher rates of smoking persist among PWH [27, 36]. In July 2012, the VHA created a tobacco cessation handbook specifically for their HIV care providers [26], although further action is needed that will focus interventions and resources on PWH.

This study has several limitations. Specifically, this cohort comprised and describes veterans receiving care in the VHA, who are predominantly male and may have tobacco use behaviors that differ from a civilian population. This study also used data from VACS, which enrolls PWoH matched on age, race/ethnicity, and site of care to PWH. Thus, results may not be fully generalizable to patients without HIV in the VHA or patients receiving care in other healthcare systems. In addition, clinical reminder smoking questions and responses, which enabled us to identify veterans with and without HIV who smoke, varied widely by site and over time until October 2018. Despite this variation in implementation, clinical reminders have been found to have substantial agreement with smoking status reported on survey [21].

Smoking cessation medications procured outside the VHA represent another potential opportunity for misclassification. Even though NRT is available over-the-counter and prescriptions can be filled at other pharmacies, veterans are less likely to obtain their medications elsewhere because of low copayments within the VHA [37–39]. Furthermore, the VHA prescription mail order service is highly rated and provides approximately 80% of VHA outpatient medications [40]. Finally, it is not clear whether nonreceipt of smoking cessation medications was due to patient unwillingness to quit, preference against pharmacotherapy, or missed opportunity for the provider to encourage cessation treatment because this information was not available.

Despite these limitations, our study has several strengths. Our study examines the extent to which PWH receive smoking cessation medications compared with demographically similar PWoH in the largest system of healthcare and the largest provider of HIV care in the United States. Previous studies have been limited by the number of PWH receiving care in their health systems; however, at least 25%–51% of veterans infected with HIV in care are current smokers [20], allowing for a large evaluation of pharmacotherapy receipt. In addition, we were able examine changes in smoking cessation medication receipt over a 16-year period and chart the potential impacts of several VHA-wide interventions to address tobacco use disorder.

CONCLUSIONS

In summary, as pharmacotherapy became standard care in treating tobacco use disorders, rates of receiving key medications nearly doubled among PWH and PWoH in the VHA. Much of this improvement is likely driven by population-based policies that have made NRT more accessible, as rates of other medications exhibited modest changes. This study provides evidence that PWH appear to be provided the opportunity to receive effective treatment at rates similar to PWoH, although efforts to improve pharmacotherapy receipt should continue for both groups given the known dangers of smoking. Finally, strategies must include a tailored, vulnerable population approach in addressing tobacco use among PWH who represent a subgroup of veterans at higher risk of risks due to smoking relative to PWoH.

Appendix 1. Characteristics Among VACS Participants Who Currently Smoke by Year and HIV Status (N = 92 632), Between 2003 and 2018

Characteristic	2003		2004		2005		2006		2007		2008		2009		2010
	HIV+ N = 9367	HIV− N = 23 209	HIV+ N = 6786	HIV− N = 16 633	HIV+ N = 6941	HIV− N = 17 557	HIV+ N = 6854	HIV− N = 17 549	HIV+ N = 7272	HIV− N = 18 499	HIV+ N = 7233	HIV− N = 18 151	HIV+ N = 7205	HIV− N = 17 916	HIV+ N = 7137	HIV− N = 17 812
HIV+, %	28.8		29.0		28.3		28.1		28.2		28.5		28.7		28.6
Age, mean (SD)	49.2 (8.2)	49.6 (8.3)	49.5 (8.4)	49.7 (8.8)	50.0 (8.7)	50.3 (9.2)	50.8 (8.8)	51.1 (9.4)	51.2 (8.9)	51.5 (9.3)	51.6 (9.0)	51.9 (9.6)	52.1 (9.2)	52.5 (9.5)	52.6 (9.6)	53.2 (9.8)
Male, %	96.9	97.7	97.2	97.5	97.2	97.4	96.6	97.5	96.9	97.5	97.1	97.5	97.0	97.4	96.8	97.4
Race, % Black	53.0	50.4	53.2	50.8	54.1	51.5	52.4	51.0	52.6	51.0	52.5	50.7	52.8	51.1	52.8	50.4
Hispanic	7.8	6.7	7.1	6.7	6.6	6.4	7.1	6.9	7.0	6.4	6.8	6.8	6.7	6.6	6.1	6.6
Other	2.0	2.2	2.1	2.5	2.2	2.5	2.0	2.5	2.5	2.6	2.3	2.5	2.3	2.6	2.4	2.6
White	37.2	40.8	37.6	40.0	37.1	39.7	38.5	39.7	37.9	40.0	38.4	40.0	38.3	39.7	38.7	40.5
Received medications, %	15.2	17.0	17.4	18.1	18.6	19.1	23.2	22.9	28.4	29.9	30.1	31.0	32.1	31.5	30.2	30.3
NRT	9.5	9.7	11.6	11.5	13.2	12.7	16.6	16.1	19.5	20.0	21.8	21.7	24.8	23.5	23.0	22.6
Bupropion	7.6	9.5	8.1	9.4	8.3	9.3	9.4	10.0	9.7	11.7	10.5	12.2	10.7	11.7	10.3	10.9
Varenicline	0	0	0	0	0	0	0.8	1.0	4.5	4.8	2.9	2.8	1.7	1.7	1.7	1.6

	2011		2012		2013		2014		2015		2016		2017		2018
	HIV+ N = 6775	HIV− N = 17 041	HIV+ N = 6713	HIV− N = 16 740	HIV+ N = 6427	HIV− N = 16 164	HIV+ N = 6445	HIV− N = 15 840	HIV+ N = 6229	HIV− N = 15 114	HIV+ N = 6081	HIV− N = 14 694	HIV+ N = 6130	HIV− N = 14 869	HIV+ N = 5878	HIV− N = 14 376
HIV+, %	28.5		28.6		28.5		28.9		29.2		29.3		29.2		29.0
Age, mean (SD)	53.0 (9.7)	53.7 (9.9)	53.4 (10.1)	54.6 (9.8)	53.8 (10.2)	55.3 (9.9)	54.3 (10.5)	56.0 (10.0)	54.7 (10.8)	56.7 (10.1)	55.3 (10.9)	57.3 (10.1)	56.3 (11.0)	58.3 (10.2)	56.9 (11.2)	59.3 (10.0)
Male, %	96.6	97.4	96.7	97.3	96.9	97.5	96.6	97.3	96.7	97.3	96.4	97.4	96.5	97.3	96.6	97.3
Race, % Black	52.8	50.8	51.2	51.4	51.7	50.9	50.9	50.8	50.8	52.7	50.6	49.9	50.3	51.4	50.0	51.4
Hispanic	6.7	6.4	6.6	6.2	6.2	6.3	6.2	6.4	7.0	6.2	6.5	7.0	7.0	7.0	7.0	6.8
Other	2.5	2.8	3.0	2.6	2.8	2.7	3.3	2.7	3.6	2.5	3.1	2.9	3.7	2.8	3.7	2.6
White	38.1	40.1	39.2	39.8	39.3	40.1	39.6	40.1	38.6	38.6	39.7	40.3	39.1	38.8	39.3	39.2
Received medications, %	31.1	30.7	31.3	30.1	32.1	30.1	31.2	30.9	32.5	31.6	34.1	32.0	33.5	31.2	33.8	32.0
NRT	23.9	23.3	24.5	22.9	24.3	23.3	24.7	23.7	26.0	25.5	27.3	25.3	26.3	24.6	26.2	25.2
Bupropion	10.3	10.8	10.0	10.3	10.7	10.0	8.9	10.3	9.6	9.2	9.4	9.1	8.4	8.5	8.1	8.4
Varenicline	1.8	1.3	1.4	1.1	1.6	1.3	1.9	1.2	1.7	1.3	2.7	2.2	3.6	2.5	4.1	3.4

Abbreviations: HIV, human immunodeficiency virus; NRT, nicotine replacement therapy; SD, standard deviation; VACS, Veterans Aging Cohort Study.