INTRODUCTION
Coronary heart disease (CHD) is a common cause of morbidity and mortality in Europe1. Although the European Society of Cardiology guidelines for secondary prevention of CHD are easily accessible and provide detailed recommendations, many patients still experience recurrent cardiovascular events2,3. Smoking is one of the major risk factors for CHD, and smoking cessation is potentially the most effective of all preventive measures, with substantial reductions in (recurrent) myocardial infarctions (MI) and death4. Quitting smoking is strongly recommended in the guidelines2. Diagnosis and/or treatment of CHD may be an important impetus for smoking cessation. Evidence-based guidelines recommended interventions for smoking cessation in secondary preventive follow-up after CHD events include: 1) advising the patient to stop smoking; 2) reiterating the benefits of quitting; and 3) agreeing on a specific plan with a follow-up arrangement2. Pharmaceutical support for stopping smoking should be considered in all smokers. However, persistent or re-uptake of smoking is common in patients with CHD5.
Smoking status is frequently assessed by self-report. Some patients falsely declare themselves to be non-smokers6,7. This can lead to an overestimation of smoking cessation rates. Objective measures of tobacco smoking may also be useful in improving clinical management and counseling of patients with difficulties quitting smoking.
The primary aim of this sub-analysis from a randomized controlled trial of hospital-based versus primary care-based follow-up after MI, percutaneous coronary intervention (PCI), and coronary artery bypass grafting (CABG), was to compare self-reported and objective measured [carbon monoxide (eCO)] smoking status. Second, we assessed the achievement of the other treatment targets for secondary prevention after MI, PCI, or CABG in patients with elevated eCO reporting smoking cessation compared to patients reporting smoking.
METHODS
Study design and study population
The data for this sub-analysis were retrieved from an open randomized controlled trial at Sorlandet Hospital, Arendal, Norway, in the period 2007–2022. The focus of the main study was to determine whether all-cause mortality and the composite endpoint of all-cause mortality, recurrence of non-fatal MI, new PCI/CABG, and non-fatal stroke differed between patients who received cardiovascular secondary preventive follow-up at the hospital and those within primary health care8. Patients who were admitted due to MI or scheduled PCI/CABG aged 18–80 years were randomized to follow-up at the hospital or within the primary health care. Lack of ability to cooperate, known alcohol or drug abuse, use of narcotics, pregnancy or breast-feeding, serious comorbidity with a life expectancy of <2 years, or participation in other secondary prevention studies were considered as exclusion criteria. Patients with a follow-up of <12 months were excluded from the analysis.
Intervention
Patients in the intervention group were followed up by trained nurses who consulted cardiologists if necessary. For this group, the first consultation was under hospitalization for the index event, followed by outpatient consultations after the discharge8. Patients were consulted regarding lifestyle measures and achievement of treatment targets evaluated at each consultation. Patients were asked to report smoking status, physical activity, diet, and use of medication. Blood pressure, body mass index, waist circumference, cardiac troponin T, lipid profile, HbA1c, and eCO were assessed8. Smokers were strongly advised to quit smoking during index hospitalization and at all outpatient visits. Nicotine replacement therapy (NRT) was offered during hospital admission, and continuation of NRT or a 12-week course of varenicline after discharge was advised8.
Patients randomized to the primary care group were followed up by family physicians with respect to the cardiovascular secondary preventive measures. The treatment targets for cardiovascular secondary prevention were described in discharge papers and sent to the patient’s family physician. Patients in this group had their outpatient appointments for data collection regularly after the discharge for index events, without interfering with follow-up visits with the family physician.
Treatment targets of secondary prevention
The secondary preventive treatment targets adhered to the latest ESC guidelines available and were as follows: no smoking, blood pressure <140/90 mmHg, LDL-cholesterol <1.8 mmol/L (<2.5 mmol/L until 2017, <1.4 mmol/L from 2020), body mass index (BMI) <25 kg/m2, daily use of lipid-lowering therapy, daily use of acetylsalicylic acid, and physical activity of moderate intensity ≥150 min weekly9-14.
Exhaled carbon monoxide (eCO)
The eCO concentration (ppm) was measured at each study visit with a CO tester (Smokerlyzer, Bedfont, UK) and rounded to the nearest integer, with values 0–5 ppm interpreted as no smoking and ≥6 ppm suggestive of smoking15. For eCO measurement, subjects were asked to hold their breath for 15 s and then exhale smoothly into the disposable mouthpiece of the monitor. The device was calibrated annually. The manufacturer reports ±2 ppm (5%) accuracy and a range of 0–150 ppm16.
Outcomes
The primary endpoint was self-reported smoking status in patients with eCO ≥6 ppm one year after the index coronary heart event (MI or scheduled CABG/PCI). The secondary endpoint was the achievement of treatment targets of secondary prevention for cardiovascular risk factors, and medication use one year after the index coronary heart event for patients with eCO ≥6 ppm and reporting non-smoking versus smoking.
Statistical analysis
Continuous variables are reported as mean ± SD (standard deviation), and differences between groups analyzed using independent samples t-tests. Categorical variables are presented as frequencies and percentages, while between-group differences are analyzed by the chi-squared test. Missing values are reported, and categorical variables are reported as the proportion of non-missing values. Complete-case analysis was used to manage missing data. A p<0.05 was regarded as statistically significant. The statistical analyses were performed using STATA, version 17 (StataCorp, 4905 Lakeway Dr, College Station, TX 77845, USA).
Ethics The study was approved by Regional Committee for Medical and Health Research Ethics (REK, 2.2007.248). All study participants signed a written consent form, and principles of Declaration of Helsinki were respected. The data collection, storage and processing were approved by Norwegian Social Science Data Services (NSD, 2007/17221). The study is registered at ClinicalTrials.gov (16.05.2008, NCT00679237).
RESULTS
A total of 1540 patients were included in the main study during the inclusion period from 2007 to 2017 (Figure 1). Patients with missing information regarding eCO or self-reported smoking status one year after the index event (n=249; 16%) were excluded from further analysis. Among the remaining 1291 patients, 358 (28%) patients reported current smoking, and 511 (40%) patients reported previously smoking at the index event (MI, PCI, or CABG). After one year, 127 (35%) initial smokers reported smoking cessation. A total of 258 (20%) patients reported current smoking.
Figure 1
Study flow chart of participants included in a sub-analysis of self-report smoking habits from a randomized controlled secondary preventive trial after myocardial infarction (MI) or after scheduled percutaneous coronary intervention (PCI) or coronary artery bypass grafting (CABG) at Sorlandet Hospital Arendal, Norway, 2007–2022
Elevated exhaled CO and self-reported smoking status
The concentration of eCO was ≥6 ppm one year after the index event in 285 (22%) patients, with a median value of 11 ppm (interquartile range, IQR: 7–20). A total of 72 (25%) of the patients with eCO ≥6 ppm reported non-smoking. This group constituted 7% of the participants reporting non-smoking. Among participants with eCO ≥6 ppm, median values were 7 ppm (IQR: 6–9) and 15 ppm (IQR: 9–22) in patients reporting non-smoking and smoking, respectively (Figure 2). A total of 113 (32%) initial smokers reported smoking cessation and had eCO <6 ppm one year after the index coronary heart event.
Figure 2
Exhaled carbon monoxide (eCO) concentration in patients self-reported smoking versus non-smoking in a sub-analysis from a randomized controlled secondary preventive trial at Sorlandet Hospital, Arendal, Norway, 2007–2022, with 1540 patients hospitalized due to myocardial infarction (MI) or after scheduled percutaneous coronary intervention (PCI) or coronary artery bypass grafting (CABG)
Baseline characteristics
Baseline characteristics at the index events (MI, PCI, or CABG) of patients with eCO ≥6 ppm reporting non-smoking and patients with eCO ≥6 ppm reporting smoking one year after the index event are described in Table 1.
Table 1
Baseline clinical characteristics at hospitalization in patients with exhaled carbon monoxide concentration (CO) ≥6 ppm reporting non-smoking and patients with exhaled CO ≥6 ppm reporting smoking one year after myocardial infarction (MI), scheduled percutaneous coronary intervention (PCI) or coronary artery bypass grafting (CABG), Sorlandet Hospital, Arendal, Norway, 2007–2022
| Characteristics | Missing values | Self-reported non-smoking | Self-reported smoking | pc |
|---|---|---|---|---|
| (N=285) n (%) | (N=72) n (%) | (N=213) n (%) | ||
| Demographic parameters | ||||
| Age (years), mean (SD) | 0 (0) | 63 (10) | 59 (10) | 0.002 |
| Men | 0 (0) | 60 (83) | 161 (76) | 0.17 |
| Married/cohabiting | 3 (1) | 55 (76) | 144 (69) | 0.21 |
| Higher educationa | 21 (7) | 20 (30) | 40 (20) | 0.11 |
| Workingb | 0 (0) | 32 (46) | 89 (42) | 0.58 |
| Smoking history | ||||
| Self-reported previous smoking | 0 (0) | 38 (53) | 17 (8) | <0.001 |
| Self-reported smoking at baseline | 0 (0) | 14 (19) | 194 (91) | <0.001 |
| Cigarettes/week, mean (SD) | 122 (43) | 69 (43) | 111 (56) | 0.03 |
| Pharmaceutical therapy | ||||
| Lipid lowering therapy | 14 (5) | 32 (47) | 89 (44) | 0.64 |
| Antihypertensive therapy | 8 (3) | 28 (40) | 87 (42) | 0.77 |
| Medical history/comorbidity | ||||
| BMI (kg/m2), mean (SD) | 34 (12) | 28 (4) | 28 (5) | 0.85 |
| Diabetes | 1 (0) | 15 (21) | 36 (17) | 0.46 |
| Previous myocardial infarction | 3 (1) | 7 (9) | 33 (16) | 0.23 |
| Previous percutaneous coronary intervention | 1 (0) | 12 (17) | 34 (16) | 0.90 |
| Previous coronary artery bypass grafting | 1 (0) | 2 (3) | 8 (4) | 0.69 |
| Previous stroke | 2 (0) | 6 (8) | 12 (6) | 0.43 |
| Renal failure | 3 (1) | 1 (1) | 5 (2) | 0.62 |
| Acute myocardial infarction as index event | 4 (1) | 43 (61) | 87 (41) | 0.005 |
| Randomization result | ||||
| Hospital based follow-up | 0 (0) | 36 (50) | 106 (50) | 0.97 |
Participants reporting non-smoking despite having an eCO ≥6 ppm were older than the patients reporting smoking. Most patients with elevated eCO reporting non-smoking one year after index event were prior smokers (n=52; 72%). Half of the patients were offered hospital-based follow-up. Greater proportion of self-reported non-smokers presented with MI as a qualifying index event.
Achievement of treatment targets for secondary prevention
Fewer patients with elevated eCO reporting non-smoking achieved the treatment target for blood pressure [38 (53%) vs 144 (68%), p=0.02]. We found no other differences in the achievement of the targets for secondary prevention between patients with elevated eCO reporting non-smoking versus those reporting smoking (Table 2).
Table 2
Secondary preventive target achievement for cardiovascular risk factors and medication use in patients with exhaled carbon monoxide (CO) ≥6 ppm reporting non-smoking and patients with exhaled CO ≥6 ppm reporting smoking one year after myocardial infarction (MI), scheduled percutaneous coronary intervention (PCI) or coronary artery bypass grafting (CABG), Sorlandet Hospital, Arendal, Norway, 2007–2022
| Target achievementa | Missing values | Self-reported non-smoking | Self-reported smoking | pb |
|---|---|---|---|---|
| (N=285) | (N=72) | (N=213) | ||
| n (%) | n (%) | n (%) | ||
| Blood pressure | 0 (0) | 38 (53) | 144 (68) | 0.02 |
| LDL-cholesterol | 8 (3) | 58 (81) | 141 (69) | 0.06 |
| Body mass index | 2 (1) | 20 (28) | 58 (27) | 0.96 |
| Lipid lowering therapy | 5 (2) | 69 (99) | 206 (98) | 0.79 |
| Acetylsalicylic acid | 1 (0) | 65 (90) | 204 (96) | 0.05 |
| Physical activity | 0 (0) | 41 (57) | 103 (48) | 0.21 |
DISCUSSION
In this sub-analysis from a randomized controlled secondary preventive trial of patients hospitalized due to MI or after scheduled PCI/CABG, an elevated eCO was found in 1 in 5 participants one year after the index event. A quarter of these patients reported non-smoking.
Registration of smoking habits in clinical studies may be based on self-report. Self-assessment of smoking status may lead to inaccurate measures of smoking exposure due to smoking denial or difficulty in recalling the quantity and duration of smoking. This misclassification potentially leads to an overestimation of the effect of smoking intervention and an underestimation of the biological effects of smoking cessation. Veit et al.7 have described a particularly high number of false negative self-reports of tobacco use among lung transplant candidates. This might be explained by the patient’s awareness of the potentially life-threatening consequences, including exclusion from transplantation. Similarly, Middleton and Morice17 described false negative self-reports among those with eCO >6 ppm. However, false declaration of non-smoking has no direct or immediate negative or positive consequences in secondary preventive follow-up after a CHD event. The reasons why some patients probably lie about their tobacco use are unclear. Embarrassment at not being able to stop smoking can be one of several possible explanations.
Higher age among the patients reporting non-smoking despite elevated eCO may possibly indicate age itself as a factor influencing eCO levels. Adjustment of a cut-off value according to the age group, with a cut-off value of 4 ppm and 5 ppm for those aged 16–25 years and 26–70 years, respectively, has been suggested18. In case higher age contributes to elevated eCO levels, it could probably explain lower median eCO levels among these patients compared to those who reported smoking, as well as contribute to higher blood pressure in this group and more advanced coronary artery disease resulting in greater proportion patients presenting with MI as an index event. The association of age and eCO might be a subject for further studies.
Except for blood pressure targets, we found no differences in the achievement of treatment targets for secondary prevention in patients with elevated eCO reporting non-smoking versus patients reporting smoking. A lower proportion of participants who achieved blood pressure treatment targets among self-reported non-smokers with elevated eCO might also support results recently described by Bradley et al.19, where higher eCO was associated with poorer cardiovascular health both for smokers and non-smokers. Although smoking markedly increases the risk of atherosclerotic cardiovascular events, data are inconsistent regarding the impact of smoking on the incidence of hypertension20-22.
Air pollution in large cities has been shown to increase dramatically eCO levels23. We assume this as a less possible reason for elevated eCO among self-reported non-smokers in our study, as there are no large cities within the geographical area covered.
In our opinion, this study indicates the potential usefulness of measuring eCO in secondary preventive follow-up after MI, PCI and/or CABG. Measurements of eCO may identify patients at high risk who benefit from optimalization of secondary preventive therapy, including pharmaceutic support for smoking cessation.
Limitations
The present study has certain limitations. We did not obtain patient histories regarding passive smoking, the last cigarette smoked, smoking patterns, or the presence of chronic obstructive pulmonary disease or asthma. Those factors can influence eCO levels and could decrease the sensitivity and specificity of CO monitoring. In addition, we did not collect demographic data related to ethnic or racial characteristics. The concentration of exhaled CO reflects smoking exposure over a limited time interval and may not provide accurate estimates if there is a break in smoking. The manufacturer reports ±2 ppm (5%) accuracy, which might contribute to the incorrect classification of non-smokers who have borderline (near 6 ppm) eCO values. Statistical analyses were not adjusted for multiple testing. Furthermore, the study was limited to one hospital and a limited number of participants.
CONCLUSIONS
Self-reported smoking status must be interpreted with caution. Some patients appear not to report their smoking status correctly. We did not delve into the reasons for this, though embarrassment or fear may be important. While not all studies of smoking cessation may be able to include objective measures of cessation, interpretation of the results of such studies will require taking the limitations of self-report into consideration. This study underlines a need for objective measures for smoking cessation both in clinical studies and in clinical practice.
