Impact of Carotid Plaque Screening on Smoking Cessation and Other Cardiovascular Risk Factors:  A Randomized Controlled Trial FREE

Many studies have found that measurements of peripheral atherosclerosis improves cardiovascular risk prediction,^1- 2 but few randomized controlled trials (RCTs) have examined the clinical impact of atherosclerosis screening.³Quiz Ref IDEvidence from observational studies suggests that screening for peripheral atherosclerosis and knowing its results might help motivate patients to optimize their health behaviors,^4- 5 such as adherence to drug regimens,⁶ change in lifestyle,^7- 8 and increase in their perception of cardiovascular risk.^7,9 However, the interpretation of these observational data is limited by the lack of comparison with a randomized control group. Two RCTs assessing the impact of atherosclerosis screening yielded conflicting results.^10- 11

Although atherosclerosis screening with carotid ultrasonography (US) or a computed tomographic scan of coronary artery calcification (CAC) has passed into law in Texas¹² and is increasingly used in clinical practice,^3,5 the role of screening for atherosclerosis in the clinical setting is still controversial.^12- 13 Given the limited RCT data, it is debated whether imaging should be considered only in certain subgroups of patients¹⁴ or extended to healthy adults,¹⁵ leading to controversial recommendations.^14- 16

Because many smokers stop smoking only after an acute cardiovascular event,¹⁷we hypothesized that patients' knowledge of plaque screening results would enhance adherence to treatment regimens,^5,18 motivation for lifestyle changes, and subsequent cardiovascular risk factor (CVRF) improvement. Pictures of atherosclerotic plaques might have a potential educational/motivational role^4- 5 by capturing a “teachable moment.” However, it is unknown whether carotid plaque screening represents a teachable moment similar to acute cardiovascular events.¹⁷ To assess whether screening for carotid plaques helps motivate smokers to stop smoking and improve their other CVRFs after a 12-month follow-up, we randomized 536 regular smokers to carotid plaque screening by means of US vs no screening (control group)¹⁹ in addition to a comprehensive smoking cessation program for all participants.

We conducted and reported this RCT according to CONSORT (Consolidated Standards of Reporting Trials) guidelines for trials assessing nonpharmacologic treatments.²⁰

We recruited smokers from the general population in the French-speaking part of Switzerland by newspaper advertisements from November 1, 2007, through November 30, 2009. As previously described,²¹ eligible participants were women and men aged 40 to 70 years who smoked at least 10 cigarettes per day, with no periods of smoking abstinence of at least 3 months in the past year.²² We used questionnaires from a previous smoking cessation trial²³ and excluded smokers with cardiovascular, life-threatening, or unstable psychiatric conditions, a high level of alcohol consumption (≥6 standard units each day during the past week), any cannabis consumption in the past 3 months, or any other drug abuse; use of nicotine replacement therapy (NRT) or other medications to quit smoking; a carotid US examination in the past year; or difficulty obtaining good US imaging of the carotids because of past radiotherapy or major neck surgery. Smokers attended baseline and follow-up visits at the University of Lausanne, which has a large experience in smoking cessation trials.²³

The institutional review board approved the protocol, and all participants gave written informed consent. Our primary outcome assessed whether US carotid plaque screening improved smoking cessation rates at 12 months. Our second outcome tested whether carotid plaque screening improved CVRF levels and overall 10-year cardiovascular risk based on the Framingham risk score (FRS).²⁴

Smokers were randomly assigned to the intervention (US) group undergoing carotid plaque screening or the control group without screening (Figure 1), in addition to a comprehensive smoking cessation program for all. Carotid B-mode US (system 5; Vingmed/General Electric) was performed by a trained observer blinded to clinical information, according to a standard protocol.⁸ Carotid plaques were defined as a focal widening of at least 50% relative to an adjacent segment, as recommended.^25- 26 Although the choice of a best carotid measurement for risk prediction is controversial,²⁵ the presence/absence of carotid plaques is a dichotomized variable that might be easily understood by patients and potentially useful to motivate a behavioral change. Intraobserver variability for carotid plaques among 20 randomly selected subjects showed 95% agreement between the first and second readings (κ = 0.89; P < .001), similar to other studies.²⁶ Smokers with at least 1 carotid plaque received 2 pictures of their plaques followed by a 7-minute tutorial on the significance of atherosclerotic plaques, as previously tested.⁸ To ensure similar contact conditions, smokers without plaques and the control group received a 7-minute tutorial on smoking risks. Participants were asked to set a quit date within the week after US.

For both groups, participants underwent a comprehensive smoking cessation program for 1 year. As previously described,²¹ this program consisted of six 20-minute individual counseling sessions, 1 telephone call at 6 months, NRT patches tailored to individual needs, and brochures on smoking cessation. Interventions and visit schedules were standardized, and the study staff underwent a smoking cessation training for health care professionals²⁷ that was based on international guidelines. We also collected data on smoking history (using a standardized questionnaire²³), comorbid conditions, educational level, and medication use.

After attending the run-in visit, participants were randomly assigned to one of the 2 groups using a computer-generated randomization scheme in a 1:1 ratio without stratification or blocking (as recommended for non–fully blinded RCTs).²⁸ The random assignments were concealed using sealed, opaque, sequentially numbered envelopes until all eligibility criteria had been assessed and informed consent had been obtained.

Quiz Ref IDBased on standard criteria for smoking cessation RCTs,²⁹ self-reported smoking cessation was biochemically confirmed by exhaled carbon monoxide levels (Micro Smokerlyser; Bedfont Scientific Ltd) at each visit in addition to serum cotinine levels at 1 year using an enzyme-linked immunosorbent assay (Inspec II-Cotinine-EIA; Mahsan Diagnostika). We classified as current smokers those with carbon monoxide levels of at least 10 ppm or cotinine levels of at least 25 ng/mL (except if the participant was currently receiving NRT). The primary outcome was 1-week smoking abstinence (point prevalence) at 12 months. The secondary outcome was continuous abstinence from the quit date to study end.^22,29- 30 Smoking cessation ascertainment and tobacco counseling were performed by staff blinded to group allocation and US results, whereas participants could not be blinded, similar to previous screening studies,¹⁹ because the US result was the potential motivational tool. Participants' physicians were also blinded to group assignments and US results.

Secondary outcomes also included the following changes in CVRF levels after 12-month follow-up: low-density lipoprotein cholesterol (LDL-C) level, blood pressure, hemoglobin A_1c (HbA_1c) level in diabetic participants, high-sensitivity C-reactive protein (hs-CRP) level (smoking cessation has been suggested as a potential intervention for elevated hs-CRP levels³¹), and overall 10-year coronary heart disease risk by FRS.²⁴ We measured weight, height, blood pressure, and levels of HbA_1c, glucose, and lipids after overnight fasting using standard methods. Plasma hs-CRP levels were measured by means of a high-sensitivity latex-enhanced immunonephelometric assay (nephelometer; Dade Behring GmbH).³² Other outcomes included change in physical activity measured by the International Physical Activity Questionnaire previously validated in 12 countries³³ and change in cardiovascular drug regimen adherence by the Morisky medication adherence questionnaire³⁴ previously used in RCTs measuring drug regimen adherence.³⁵ To detect potential psychological harms of screening,¹¹ we assessed stress with the 4-item Perceived Stress Scale,³⁶ quality of life by the 36-Item Short Form Health Survey,³⁷ and depression scores using the Beck Depression Inventory.³⁸

The primary analysis examined the unadjusted rates of 12-month smoking cessation with χ² tests using an intention-to-screen allocation of groups. We conducted sensitivity analyses, adjusting for factors that differed at baseline. In prespecified subgroup analyses,²¹ we stratified the US group by the presence or absence of carotid plaque (their presence being the presumable motivating factor) and compared those US group participants with vs those without plaque and those participants with plaque in the US group vs the control group. We performed the same analyses for the other outcomes (CVRF and FRS) using 2-sample t tests. To assess the sensitivity of our results to missing 12-month CVRFs (78 participants with follow-up by telephone or letter; smoking cessation still validated by carbon monoxide level measurements), we used multiple imputation that included all baseline characteristics (except collinear variables) and 12-month smoking cessation.³⁹

Our power calculations have been previously described.²¹ Briefly, we expected 10% 12-month smoking cessation rates in the control group, based on trials of NRT,⁴⁰ and 20% rates in the US group. Using an α level of .05, a power of 80%, and an estimated 20% dropout rate, a difference would be observed with 265 smokers per group (with a total sample size of 530). No interim analyses were planned. For secondary outcomes, we could detect clinically meaningful CVRF changes with a power of more than 80% (changes in LDL-C level, <11.6 mg/dL [to convert to millimoles per liter, multiply by 0.0259]; in blood pressure, <5 mm Hg; in hs-CRP level, <1 mg/L [to convert to nanomoles per liter, multiply by 9.524]; and in FRS, <0.2% 10-year coronary heart disease risk).

We enrolled 536 smokers, including 267 randomly assigned to the carotid plaque screening group and 269 to the control group (Figure 1). A similar number of patients withdrew or were lost to follow-up among the randomized groups. Overall, 1 patient died in the US group and 2 in the control group; these deaths were unrelated to cardiovascular disease (suicide, drowning, and gastrointestinal tract bleeding). Fourteen participants were lost to follow-up, and 7 who had withdrawn their participation were classified as current smokers. The 3 deceased participants and 1 who had moved to an untraceable address were excluded from abstinence analyses, as recommended.²⁹ At baseline, participants (mean age, 51.1 years; 45.0% women) smoked an average of 20 cigarettes per day, with a median duration of smoking of 31.7 years^33,41- 42 (Table 1). Baseline characteristics did not differ markedly between the randomized groups (Table 1). The US group had a 57.9% prevalence of carotid plaques. Most participants (n = 461) intended to quit smoking within 30 days (86.0%), 65 (12.1%) intended to quit within 6 months, 8 (1.5%) were already trying to quit, and 2 (0.4%) were indeterminate.

At 12 months, biochemically validated smoking cessation rates were higher than 20%. Quiz Ref IDAlthough there was a nonsignificant pattern of higher cessation rates at 1 and 6 months in the US group, the cessation rate did not differ at 12 months between both groups in intention-to-screen analysis for point prevalence abstinence (24.9% [95% CI, 19.7%-30.1%] vs 22.1% [95% CI, 17.1%-27.1%]; P = .45) and continuous smoking abstinence (20.4% [95% CI, 15.5%-25.2%] vs 20.2% [95% CI, 15.4%-25.0%]; P = .97) (Figure 2). Very few participants did not submit samples for biochemical verification of cessation, and the frequency of the few positive sample findings did not differ by group (Table 2).

The 12-month change in blood pressure and levels of LDL-C, hemoglobin HbA_1c, and hs-CRP did not differ between the 2 groups (Table 2). The mean change in 10-year FRS was +0.6% in the US group vs +0.3% in the control group (P = .56). Change in CVRF treatment did not differ except for antihypertensives in the US group (+2.1% vs 0% [P = .03]) but without a difference in controlled hypertension. The proportion of participants reporting potential harms did not differ between groups. No participant underwent carotid revascularization during the 1-year follow-up, 1 had coronary artery stenting for stable angina, and 3 had bypass procedures or dilation for lower extremity arteriopathy. Moreover, 1 participant had a cerebellar ischemic stroke and 1 had a myocardial infarction.

In the preplanned subgroup analysis among the US group, cessation rates did not significantly differ among participants with and without carotid plaques (point prevalence abstinence, 25.7% [95% CI, 18.7%-32.6%] vs 24.1% [95% CI, 16.2%-32.0%] if no plaques; P = .77), and changes in CVRFs were also similar (Table 3) except for a larger decrease in LDL-C levels in those with vs without plaques (−8.1 vs +3.9 mg/dL; P = .004), which persisted after removing participants who were treated or who initiated treatment with medications to lower lipid levels. Rates of controlled hypertension were slightly higher among those with plaques, which was likely related to the small increase in antihypertensive use. The rates of smoking cessation and 12-month CVRF change and control were similar in sensitivity analyses adjusting for the unbalanced distribution of some baseline characteristics or after multiple imputations for missing CVRFs.

Point prevalence abstinence was significantly associated with male sex (odds ratio, 1.68; 95% CI, 1.09-2.57), older age (1.87 for a 10-year increase; 95% CI, 1.22-1.88), lower number of cigarettes per day (0.76 for 10 more cigarettes; 95% CI, 0.60-0.96), and lower duration of smoking history (0.57 for a 10-year increase; 95% CI, 0.38-0.83) in multivariate analysis adjusting for randomized group assignment and baseline characteristics selected by backward deletion using P ≤ .20.

Quiz Ref IDIn this RCT of 536 middle-aged smokers, carotid plaque screening by means of US was not associated with increased smoking cessation or better control of CVRFs after 12 months when added to comprehensive individual counseling and use of NRT products. In both groups, smoking cessation rates were more than 20% at 12 months. Furthermore, in the US group, smoking cessation rates did not differ according to the presence or absence of carotid plaques.

Our study adds important information to the few trials that have assessed the impact of atherosclerosis screening.¹³ In contrast to previous RCTs that were limited by a low number of participants with atherosclerosis¹¹ or the lack of biochemical validation of smoking cessation,^10,43 our RCT included the largest sample of smokers and smoking cessation was assessed objectively. A recent meta-analysis has found limited evidence of an effect of cardiovascular imaging on health behavior in primary prevention and called for higher-quality studies because most previous RCTs were relatively small (most trials with <200 participants) and included few smokers and because several did not record medications used during follow-up.⁴⁴ Our results are consistent with a previous RCT that found no difference in CVRF control among 450 military personnel aged 39 to 45 years who underwent CAC screening vs no screening.¹¹ However, the prevalence of smoking and other modifiable risk factors was low in that RCT, and only 15% of their participants had abnormal CAC scores⁵ compared with our higher-risk population with a 57.9% prevalence of carotid plaques. Our results were also consistent with the findings of a previous study⁴³ that detected no difference for 4-year smoking cessation rates (P = .64) and physical activity (P = .77) among 2137 volunteers randomized to CAC screening vs no scanning. However, the previous study included a small number of smokers (111 participants [5.2%]), and participants underwent only a single counseling session on CVRFs compared with 6 sessions during the 12 months in our study. Although CVRFs partially improved in their CAC group compared with the no-scanning group, the absolute changes were very small (systolic blood pressure difference, −2 mm Hg; LDL-C level difference, −6 mg/dL). Potential explanations for these differences were that these changes occurred in highly educated participants (postgraduate education, 91% vs 21.9% in our RCT) and were concurrent with the introduction of antihypertensives (P = .02; 18%-24% vs 0%-2.1% in our RCT) and medication to lower lipid levels (P = .06; 25%-29% vs 3.9%-5.3% in our RCT). Whether these modest changes would reduce cardiovascular events remains to be demonstrated because myocardial infarction rates did not differ between participants who received scanning and those who did not.

For smoking cessation, a recent Cochrane review found little evidence of increased smoking cessation rates by providing feedback on the biomedical effects of smoking⁴⁵ except a significant benefit of lung age feedback after spirometry (relative risk, 2.12; 95% CI, 1.24-3.62) and US plaque screening in an RCT conducted in the Seychelles islands. In the study by Bovet et al,¹⁰ smokers who were provided with pictures of their own atherosclerotic plaque had a higher smoking cessation rate (17.6% in the screening group [22.2% in those who had plaque] vs 6.3% in the group without screening). However, that study included mostly light smokers (mean, 10 cigarettes per day), biochemical validation of smoking cessation was not performed, NRT was not provided, and smokers were provided with one brief smoking counseling session. Compared with previous studies, smoking abstinence at 6 months was far higher in the control group of our RCT (36.4% vs 6%), likely owing to the intensive program of smoking cessation counseling; we obtained similar high rates of smoking cessation among the control group in another RCT using the same program.²³ We therefore cannot exclude carotid US as having no effect on smoking cessation because of the high intensity of our smoking cessation program per se. Whether plaque screening would have an effect on smoking cessation if counseling was less intensive (eg, in primary care settings) remains to be examined.

Quiz Ref IDOur study has several limitations. We included only smokers motivated to quit, and volunteers are generally healthier than random smokers, potentially explaining the high cessation rates in our study. Imaging results were discussed only once, which might have limited the impact. However, previous trials of the effect of subclinical cardiovascular disease screening on patient behavior also provided information on screening results only once^10- 11,43; we also found that knowledge retention about atherosclerosis⁸ remained very high at 12 months (94%) in our RCT. Carotid US screening was not repeated during follow-up, and we could not assess the potential motivational impact of plaque progression/regression. Because 21.1% of participants had missing hs-CRP measurements at 12 months, we used multiple imputation to assess the sensitivity of our results to these missing data and found similar results. The significantly larger decrease in LDL-C levels in participants with plaques vs those without plaques should be interpreted with caution. This difference persisted after removing participants treated with medications to lower lipid levels and may be related to better diet or better drug regimen adherence or may have occurred by chance in the context of multiple analyses. We did not examine clinical cardiovascular events because such analysis would have likely required more than 10 000 study subjects.¹⁸

Our RCT has several strengths. Despite the fact that many smokers discontinue RCTs in case of smoking relapse, we had remarkably few dropouts (3.9%) compared with previous smoking cessation RCTs.^22,30,46 Our dropout rate was also lower than those in several previous large studies of atherosclerosis screening^11,43- 44 and lower than planned in our power calculations (20%). Compared with previous similar RCTs,^10- 11 we included a higher-risk population with a high prevalence of CVRFs and carotid plaques (57.9%) and performed a biochemical validation of smoking cessation.

In summary, our RCT demonstrates that the addition of carotid plaque screening to a thorough smoking cessation program is not associated with improved smoking cessation or better control of CVRFs. These findings are important, given the increasing use of such screening in clinical practice^5,18 and the growing trend of insurance company reimbursement for this screening.¹² Future trials should examine the effect of atherosclerosis screening on cardiovascular events,³ although some have argued that such large and expensive trials may not be undertaken.^15,47 Moreover, it should be evaluated whether atherosclerosis screening may be helpful among smokers who undergo less intense smoking counseling or to refine cardiovascular risk estimates to better target preventive medications.⁴ Until such data become available, the benefit of atherosclerosis screening should be demonstrated in RCTs⁴ before its use can be generally recommended.⁴⁸