The Association of Smoking Status with Angina and Health-Related Quality of Life After Acute Myocardial Infarction

Abstract

Background

Smoking cessation after acute myocardial infarction (AMI) decreases the risk of recurrent AMI and mortality by 30–40%, but many patients continue to smoke. The association of smoking with angina and health-related quality of life (HRQOL) after AMI is unclear.

Methods and Results

Patients in 2 U.S. multicenter AMI registries (n=4003) were assessed for smoking and HRQOL at admission and 1, 6 and 12 months after AMI. Angina and HRQOL were measured with the Seattle Angina Questionnaire (SAQ) and Short Form-12 (SF-12) physical and mental component scales (PCS and MCS). At admission, 29% never had smoked, 34% were former smokers (quit before AMI) and 37% were active smokers, of whom 46% quit by 1 year (recent quitters). In hierarchical, multivariable, regression models that adjusted for sociodemographic, clinical and treatment factors, never and former smokers had similar and the best HRQOL in all domains. Recent quitters had intermediate HRQOL levels, with angina and SF-12 MCS scores similar to never smokers. Persistent smokers had worse HRQOL in all domains compared to never smokers and worse SF-12 MCS scores than recent quitters.

Conclusion

Smoking after AMI is associated with more angina and worse HRQOL in all domains, while smokers who quit after AMI have similar angina levels and mental health as never smokers. These observations may help encourage patients to stop smoking after AMI.

Smoking is common among patients presenting with acute myocardial infarction (AMI) and represents an important modifiable risk factor for recurrent events. Smoking cessation after AMI decreases the risk of recurrent MI and mortality by 30–50%.^1–4 Although efforts to improve smoking cessation after AMI have become important performance measures for both hospitals⁵ and outpatient clinics,⁶ a large percentage of smokers do not quit after their AMI.^{3, 4, 7} While current educational strategies focus upon the risks of continued smoking, patients may be concerned that smoking cessation will lead to worse quality of life (e.g., increased negative affect). Such concerns may lower patients’ motivation and success with quitting.⁸

Patients recovering from an AMI currently have no information about how smoking cessation might alter their health status (their symptoms, function and quality of life). Although it is well established that smoking after AMI is associated with a significantly greater risk of recurrent MI and mortality,^1–4 there are few studies describing how smoking relates to health-related quality of life (HRQOL) in cardiac patients,^9–15 despite HRQOL often being equally or more important to patients than longevity.^{16, 17} Better illuminating the health status impact of continued smoking after AMI might support improved counseling of patients to quit.

To address this current gap in knowledge, we sought to describe the association between smoking status and patient HRQOL in a cohort of 4003 post-AMI subjects using data from 2 large, prospective, observational registries. Understanding the association of smoking cessation with HRQOL could have important implications for smoking prevention and the treatment of patients who are actively smoking at the time of AMI.

METHODS

Study Population and Protocol

Our analytic cohort was derived from two consecutive multicenter, prospective cohort studies of patients hospitalized with AMI in the U.S. The Prospective Registry Evaluating Myocardial Infarction: Events and Recovery (PREMIER) study enrolled 2498 patients with AMI from 19 US hospitals between January 2003 and June 2004.¹⁸ Similarly, 4340 patients with AMI from 24 US hospitals were enrolled into the Translational Research Investigating Underlying disparities in acute Myocardial infarction Patients’ Health status (TRIUMPH) study between June 2005 and December 2008 (12 hospitals participated in both studies).¹⁹ Both studies were coordinated by Saint Luke’s Mid America Heart Institute and employed identical inclusion and exclusion criteria. Eligible patients had biomarker evidence of myocardial necrosis and additional clinical evidence supporting the diagnosis of an AMI (e.g., prolonged ischemic signs/symptoms or electrocardiographic ST changes during the initial 24 hours of admission).

Baseline data, including smoking status and health status data, were obtained through chart abstraction and a structured interview by trained research staff within 24 to 72 hours of admission. Detailed follow-up interviews were attempted on all survivors at 1, 6, and 12 months after AMI. To be eligible for the current study, patients had to survive to 1 year and not be discharged against medical advice, to another acute-care facility, or to hospice. Each participating hospital obtained Institutional Research Board approval, and all patients provided written informed consent for baseline and follow-up assessments.

Smoking Status

Smoking was assessed at baseline and each follow-up time point using the following options: (1) I have never smoked, not even a puff, (2) I have smoked in the past but less than 100 cigarettes total, (3) I stopped smoking more than 1 year ago, (4) I stopped smoking between 1 month and 1 year ago, and (5) I have smoked (even a puff) in the past 30 days.” At baseline, responses (1) and (2) were categorized as “never smokers”. Responses (3) and (4) were categorized as “former smokers”, and response (5) was categorized as “current smokers.” Those who were “current smokers” at baseline were then reclassified based on their 1-year follow-up interview responses. At 1 year, those who responded that they had quit in the past year were designated “recent quitters” while those who continued to smoke were designated “persistent smokers.” Patients with inconsistent answers at baseline and 1 year (e.g., reported being a former smoker at baseline and never smoker at 1 year) and patients who started smoking during the year (e.g., reported being a former smoker at baseline and a current smoker at 1 year) were excluded from the analyses.

Health-Related Quality of Life

Disease-specific and generic HRQOL were measured using the Seattle Angina Questionnaire (SAQ)²⁰ and the Medical Outcomes Study 12-item Short Form (SF-12).²¹ The SAQ is a reliable, responsive, and valid 19-item questionnaire that assesses the symptoms, function, and quality of life (QOL) of patients with coronary artery disease and has been shown to be associated with mortality, admission for acute coronary syndrome, and healthcare costs.^{22, 23} For this study, we focused on the SAQ angina frequency and QOL domains. Scores range from 0 to 100, with higher scores indicating less disease burden. Based on prior work, SAQ angina frequency was categorized into the clinically interpretable framework of no (score 100) vs. any (score <100) angina.²⁴ The SF-12 is a reliable and valid measure of generic health status that provides summary component scales for overall physical (PCS) and mental (MCS) health.²⁵ Scores are standardized using norm-based methods to a mean of 50 and a standard deviation of 10 with higher scores indicating better health status.²¹

Statistical Analysis

The purpose of these analyses was to describe the association between smoking status and patient HRQOL 1 year after AMI. Baseline characteristics and post- AMI treatment of the 4 smoking groups (defined at the 12-month time point as ‘never smokers’, ‘former smokers’,’ recent quitters’, and ‘persistent smokers’) were compared using the chi-square test for categorical variables and one-way ANOVA for continuous variables. To describe the HRQOL profile over the course of 1-year after MI across the full spectrum of smoking status categories, unadjusted baseline and 1-, 6-, and 12-month follow-up HRQOL scores were compared among smoking status groups using ANOVA.

We then constructed multivariable regression models to assess the independent association of 1-year smoking status with HRQOL at one year following AMI. All models were hierarchical with site entered as a random effect to account for clustering of patients within sites. Models were logistic (SAQ angina) or linear (SAQ QOL, SF-12), as appropriate. For all models, never smokers were used as the reference group and then recent quitters were also compared to persistent smokers. As smoking might be associated with several factors that may also affect HRQOL, we performed sequential multivariable adjustment to better examine the effect of adjustment on the independent association between smoking and HRQOL. We serially adjusted for factors in the order of basic demographics (age, sex, race), clinical comorbidities (hypertension, diabetes, prior MI, prior angioplasty, prior bypass graft surgery, heart failure, ST-elevations at presentation, LV systolic dysfunction [ejection fraction <40%], lung disease, hypercholesterolemia, and history of renal failure), treatment during the index hospitalization (angioplasty, bypass graft surgery, and % of quality of care measures received [e.g., timely reperfusion for ST-elevations, beta-blocker at discharge, etc.]), post-MI treatment during the year following the index hospitalization (revascularization within 1 year; whether the patient was prescribed antiplatelet, beta blocker, statin, or angiotensin converting enzyme [ACE]/angiotensin II receptor blocker [ARB] therapy and whether the patient reported taking these medications at follow-up; and patient report of participation in cardiovascular rehabilitation), sociodemographic characteristics (marital status, education, insurance status, self-reported monthly financial reserve, and self-reported avoiding medication due to costs), and psychosocial characteristics of social support (assessed with the Enhancing Recovery in Coronary Heart Disease [ENRICHD] Social Support Inventory)²⁶ and symptoms of depression (assessed with the 9-item Patient Health Questionnaire [PHQ-9]²⁷).

Baseline data were fairly complete, with only 22% missing 1 covariate and 13% missing more than 1 and an average of 0.68 covariates missing per patient. Missing covariate data were imputed using multiple imputation incorporating all baseline and outcome variables. To examine the possibility of selection bias, we compared the baseline characteristics, including baseline smoking status, of those included in the analysis with those excluded due to missing or inconsistent 1-year follow-up data.

All statistical analyses were conducted using SAS Version 9.3 (SAS Institute, Cary, NC), IVEware (University of Michigan, MI), and R Version 2.11.1 (Free Software Foundation, Boston, MA).²⁸ All tests for statistical significance were 2-tailed and were evaluated at a significance level of 0.05.

RESULTS

Patient Population

Among the 6838 patients included in TRIUMPH and PREMIER, baseline smoking status was obtained on 6780. We excluded 148 patients who were not discharged alive to home and 422 patients who did not survive to the 1-year interview. Additionally, we excluded 1585 (23%) patients who did not participate in the 1-year interview, 167 patients who participated in the interview but did not report 1-year smoking status or HRQOL, and 455 patients who gave inconsistent smoking status answers (n=393; e.g., reported being a former smoker at baseline but at 1 year reported having never smoked) or who started smoking during the year following AMI (n=62). As such, the final analytic cohort was comprised of 4003 patients (Figure 1).

Figure 1.

Flowchart of patient inclusion and exclusion in the study

Patients who were excluded due to missing or inconsistent smoking status data (n=2207), as compared with those included in the analysis (n=4003), were more likely to be younger, non-Caucasian, have poorer socioeconomic status, have cardiac and non-cardiac comorbidities, undergo less revascularization, and take less cardiac medications (Supplemental Table 1). However, there was no difference in baseline smoking status between those with and without missing/inconsistent data (p=0.116).

Description of Smoking Groups

At the time of admission for AMI, 29% of patients reported that they were life-long non-smokers, 34% were former smokers (i.e., quit prior to AMI), and 37% had smoked within 30 days of the AMI. Of those currently smoking at the time of their AMI, 46% quit smoking during the following year and were classified as recent quitters, while 54% were persistent smokers 1 year after their AMI. The baseline characteristics and post-AMI treatment of never smokers, former smokers, recent quitters, and persistent smokers are presented in Table 1. There were numerous differences among the 4 smoking status groups. For example, those who were smoking at the time of their MI, compared to those who were not, tended to be about 10 years younger, unmarried, less educated, of lower socioeconomic status, have higher depression scores, and have less comordities. They were also more likely to present with an ST-elevation MI and be treated with percutaneous coronary intervention during their index hospitalization, but less likely to participate in cardiac rehabilitation after their AMI.

Table 1.

Comparison of Baseline Characteristics and Post-AMI Treatment by 1-Year Smoking Status

	Never Smoked n = 1145	Former Smokers n = 1374	Recent Quitters n = 683	Persistent Smokers n = 801	P-Value
Basic Demographics
Age	62.8 ± 13.3	64.1 ± 10.7	54.6 ± 9.8	54.6 ± 9.8	< 0.001
Male	55.4%	74.9%	71.2%	66.7%	< 0.001
Caucasian	75.7%	84.0%	72.7%	69.8%	< 0.001
Sociodemographics
Marriage	63.1%	69.5%	59.7%	45.9%	< 0.001
Less than high school education	41.3%	44.9%	50.4%	57.5%	< 0.001
Insurance coverage for medications	81.0%	82.8%	72.6%	66.5%	< 0.001
Monthly financial situation Not enough to make ends meet	12.6%	10.3%	18.6%	27.7%	< 0.001
Avoided health care due to cost	14.7%	13.8%	25.7%	34.6%	< 0.001
Psychosocial factors
ENRICHD Social Support Score	22.7 ± 3.7	22.7 ± 3.7	22.1 ± 4.1	21.2 ± 5.0	< 0.001
PHQ Depression Score	4.5 ± 4.9	4.7 ± 4.9	5.2 ± 5.3	6.4 ± 5.8	< 0.001
Clinical Co-morbidities
Hypertension	65.5%	68.2%	52.4%	58.8%	< 0.001
Diabetes	30.9%	30.9%	20.5%	20.2%	< 0.001
Prior MI	17.4%	22.6%	12.3%	20.3%	< 0.001
Prior PCI	15.5%	23.3%	11.7%	17.4%	< 0.001
Prior CABG	12.1%	17.8%	5.4%	6.9%	< 0.001
Congestive Heart Failure	7.8%	8.9%	2.9%	6.5%	< 0.001
STEMI	41.5%	42.3%	58.7%	49.7%	< 0.001
LV systolic dysfunction	18.8%	20.4%	20.1%	17.8%	0.425
Hypercholesterolemia	51.9%	57.0%	41.0%	44.6%	< 0.001
Chronic Renal Failure	7.7%	8.2%	2.5%	4.1%	< 0.001
Chronic Lung Disease	5.7%	10.1%	7.0%	10.4%	< 0.001
Treatment during index hospitalization
In-hospital PCI	63.2%	66.7%	75.8%	72.7%	< 0.001
In-hospital CABG	10.7%	13.0%	12.2%	6.2%	< 0.001
<100% of Quality of Care Measures Received	27.3%	29.3%	41.3%	40.1%	< 0.001
Post-MI treatment within 1 year of index hospitalization
Revascularization	7.2%	8.2%	7.9%	6.4%	0.459
Antiplatelet use	83.1%	87.8%	90.7%	85.3%	< 0.001
Beta blocker use	71.0%	73.8%	75.4%	68.8%	0.022
Statin use	64.8%	69.0%	71.5%	63.7%	0.003
ACE/ARB use	50.5%	51.0%	56.7%	53.1%	0.061
Cardiac Rehabilitation	54.3%	54.2%	50.8%	35.0%	< 0.001

Continuous variables compared using one-way analysis of variance. Categorical variables compared using chi-square or Fisher's exact test.

Unadjusted Analyses at Baseline, 1-, 6-, and 12-months

Overall, all groups showed improvement from baseline to 1 year in SAQ angina, SAQ QOL, and SF-12 MCS, but little or no improvement in SF-12 PCS. In unadjusted comparisons, there were statistically significant differences (p < 0.001) across the 4 smoking status groups for each HRQOL domain at baseline and 1, 6, and 12 months after AMI. At each separate time-point, there was a gradation of more angina and worse disease-specific and generic HRQOL across the 4 smoking status groups from never smokers having the best health status and persistent smokers the worst (Supplemental Table 2 and Supplemental Figure 1).

Multivariable Analyses of 1-year Smoking Status and 1-year HRQOL Domains

The associations between smoking status and each domain of 1-year HRQOL were somewhat attenuated at each step of sequential adjustment, but no particular step drastically changed the step-wise associations between smoking status and 1-year HRQOL, indicating that no particular factor category accounted for a greater portion of variance in the sequential analysis than other categories (Supplemental Figure 2). Thus, the final fully-adjusted models are shown as the primary results (Figure 2a–d). In the fully-adjusted models with never smokers as the reference group, former smokers were statistically similar to never smokers for all HRQOL domains. Recent quitters varied in the degree they were similar to and different from never smokers, depending upon the HRQOL domain. There were no statistically significant differences between recent quitters and never smokers in the odds of having angina at 1 year (Figure 2a) and in SF-12 MCS scores (Figure 2d). Recent quitters had worse SAQ QOL and SF-12 PCS scores compared with never smokers, having a 2.7-points lower score on the SAQ QOL scale (95% CI -4.5 to −0.9, p = 0.003) and a 2.3-points lower score on the SF-12 PCS (95% CI −3.4 to −1.2, p < 0.001) (Figures 2b and c). Persistent smokers had worse HRQOL than never smokers for all domains. Persistent smokers had a 1.5-fold increased odds of having angina at 1 year as compared with never smokers (95% CI 1.1 to 1.9, p = 0.003), 3.5-points lower scores on the SAQ QOL scale (95% CI -5.2 to −1.7, p < 0.001), 1.6 points less on the SF-12 PCS (95% CI −2.7 to -0.5, p = 0.003), and 2.3 points less on the SF-12 MCS (95% CI −3.3 to −1.4, p < 0.001) (Figures 2a–d).

Figure 2.

a-d. Multivariable analyses of the association of 1-year smoking status with HRQOL domains at 1 year following AMI.

In fully-adjusted models comparing recent quitters with persistent smokers, recent quitters were statistically similar to persistent smokers for angina (p = 0.298), SAQ QOL (p = 0.411), and SF-12 PCS (p = 0.220) (Figures 2a–c). However, recent quitters had better SF-12 MCS scores compared with persistent smokers (p < 0.0001) (Figure 2d).

DISCUSSION

In a large, multicenter cohort of AMI patients, we found a gradation in the association of smoking status with angina and HRQOL. Those who had never smoked had the least angina and the best disease-specific and general HRQOL. Based on adjusted analyses, those who quit prior to their AMI (former smokers) were very similar to those who had never smoked in all HRQOL domains. In contrast, those who continued to smoke after their AMI (persistent smokers) were significantly worse than those who had never smoked in all HRQOL domains. Those who quit smoking after their AMI (recent quitters) had intermediate levels of HRQOL, with their experience of angina and mental health status being similar to those who never smoked. Furthermore, recent quitters had markedly better general mental health than persistent smokers.

Our findings support and extend prior studies that have reported smoking to be associated with poorer HRQOL;^{9, 10, 12} while other studies have reported that smoking is not associated with HRQOL¹³ and that smoking cessation may, in fact, be associated with poorer HRQOL for some patients.¹⁴ In terms of disease-specific HRQOL, our findings complement those from studies of patients undergoing PCI, which demonstrated that never and former smokers have similar levels of angina and angina-related quality of life that are significantly better than that of those who persist in smoking.^{10, 11} Jang et al. further found that patients who quit smoking within 1 year after PCI had less angina than those who continued to smoke.¹¹ In our study, although recent quitters and persistent smokers still had similar levels of angina, recent quitters did not have more angina compared to never smokers whereas persistent smokers had more angina than never smokers.

There are several potential mechanisms underlying the association between smoking and angina. Exposure to toxic compounds contained in cigarette smoke causes free-radical mediated oxidative stress and decreased nitric oxide bioavailability resulting in decreased endothelium-dependent vasodilatation.^29–32 Studies have shown that smoking increases mean platelet volume, platelet activation, platelet aggregation and non-responsiveness to clopidogrel.^33–36 Smoking also causes endothelial damage and activation of inflammation pathways resulting in increased expression of pro-inflammatory cytokines, adhesion molecules and other prothrombotic factors.^37–39 Damage to the endothelium allows the subendothelium to be exposed to these factors as well as platelets with increased activity and the overall result can promote thrombus formation. Moreover, in patients with coronary artery disease, smoking has been associated with increased adrenergic tone and increased coronary vasospasm.^{37, 40}

In terms of general HRQOL, our results are also consistent with those of Stafford et al.¹⁵ who looked at this association in a broad group of patients with coronary artery disease and showed a stronger inverse association between smoking status and general mental health than with general physical health. In our study, even after adjusting for social support and depressive symptoms, the general mental health status of both former and recent quitters was very similar to that of never smokers. Furthermore, recent quitters had substantially better general mental health status than persistent smokers. Our findings are also consistent with a recent systematic review and meta-analysis that concluded that smoking cessation is associated with improved mental health compared with continued smoking in both general and clinical populations . ⁴¹ There are likely to be complex mechanisms underlying the association of smoking and poor mental health that involve biological, behavioral, and environmental factors. For example, there is evidence of shared biological pathways of both smoking and depression with coronary artery disease as smoking has been associated with oxidative stress⁴² and increases in C-reactive protein induced inflammation⁴³ which have been depicted in the pathophysiology of both depression and cardiovascular disorders.^44–47

Similar to the approach of Smedt et al.¹² and Jang et al.,¹¹ our study extends the insights from prior studies on smoking status and HRQOL in coronary heart disease patients by comparing HRQOL across 4 smoking status categories that address the full smoking status spectrum: never smokers, former smokers (quit before AMI), recent quitters (quit after AMI), and persistent smokers (continued to smoke after AMI). However, in contrast to previous studies, our study included only subjects with AMI and is therefore directly relevant to post-AMI patients. Furthermore, our study used HRQOL measures specific to coronary artery disease, which greatly strengthens our findings and conclusions as they are specific to AMI patients who smoke. In fact, our study is the largest known study to examine the disease-specific manifestations of continued smoking after AMI.

It is important to interpret our findings through a clinical framework. In our attempt to isolate the effect of smoking with HRQOL, we knowingly adjusted for multiple covariates that are associated with both smoking and HRQOL (e.g., depression, socioeconomic status). As such, the small group differences in the multivariable analyses are potentially due to overadjustment and the complex inter-relationships between smoking and other variables, such as depressive symptoms, may underestimate the HRQOL benefits of quitting smoking (e.g., patients who quit may also become less depressed). In addition, it should be noted that the HRQOL mean differences are reported at the population level rather than at a patient level. As such, our findings likely indicate that smoking cessation is associated with a large HRQOL improvement in some patients while others may have no benefit. Finally, noting that former smokers were statistically similar to never smokers indicates that there may be a time-dependent effect of smoking on health status. Patients categorized as recent quitters could have quit just over 30 days prior to their 1-year follow-up interview after their AMI and yet it could be that longer durations of smoking cessation are associated with better HRQOL. Understanding the long-term effect of smoking cessation on HRQOL and the time-dependence of this effect will require longer study. However, even in the absence of large improvements in HRQOL with smoking cessation, we have, at a minimum, shown that smoking cessation is not associated with worse HRQOL. In addition, our results—showing better mental health for those who quit smoking—may provide helpful information for those concerned that smoking cessation may have a detrimental effect on mental health.

Our results must be interpreted in light of several study limitations. First, patient self-report was used to determine smoking status and these reports were not accompanied with biochemical measures for confirmation. However, research shows that self-report is a valid indicator of current smoking, especially when there are no strong incentives to deceive.^{48, 49} Second, some studies have shown that decreasing intensity (pack-years and/or number of cigarettes per day) can improve HRQOL.³ As we did not assess intensity of smoking we cannot comment on whether there were differences in angina and HRQOL within the ‘current smoker’ group associated with different smoking intensity subgroups. Third, we excluded a large number of subjects due to missing or inconsistent data. This could limit study generalizabilty and introduce possible selection bias as those excluded for missing/inconsistent data did differ from those included on several baseline characteristics and treatment factors that could be associated with smoking cessation and HRQOL. However, those excluded for missing/inconsistent data did not differ in baseline smoking status as compared with those we included. Finally, while we adjusted for many clinical and sociodemographic factors that may confound the association between HRQOL and smoking, there is always the potential for unmeasured confounding in any observational study.

In conclusion, we observed, in a large multicenter cohort of post-AMI patients, that persistent smoking after the AMI is associated with significantly more angina and worse disease-specific and general HRQOL. Those who quit smoking prior to their AMI were similar to those who had never smoked in all HRQOL domains. Within one year, the odds of having angina and the general mental health status of those who quit smoking after their AMI was also similar to those who had never smoked. Further study of the time-dependent effect of longer-term smoking cessation on HRQOL might determine if those who quit smoking after their AMI might see benefits in additional areas of HRQOL with increased time of abstinence. Our findings provide unique information to post-AMI patients and should provide strong support for counseling patients about how smoking cessation not only reduces the risk of MI and death, but is also associated with better health status over time. As such, these observations may offer current smokers increased incentive and motivation for quitting.