The present study showed that implementing evidence-based, low cost- and readily available methods for smoking cessation can have a significant impact on smoking habits after an MI. Implementation resulted in a higher proportion of abstinent patients at 2-month follow up; 64% vs 54% during the reference period, and including only patients counselled during admission smoking abstinence was 73% vs 54%.
To stop smoking following onset of CHD or after undergoing revascularization results in a considerably lower risk of death. Information and support for smoking cessation should be delivered routinely to all patients with a diagnosis of CHD (3, 9). This risk reduction is evident across strata by sex, age, index cardiac event, country, and year of study initiation (3).
According to SWEDEHEART data, being abstinent from smoking at 2-months after MI is associated with a 50% reduction in 10-year mortality risk (Fig. 4) (6). The population of patients as well as the structure of follow-up in SWEDEHEART is practically equivalent to the present study´s control period population. Hence, more abstinent patients 2-months after MI should be associated with a significant improvement regarding both long-term life expectancy.
At all six hospitals participating in the project, various forms of smoking cessation support were already in-place before project initiation. Thus, the implemented model was largely an implementation that strengthened and structured existing support. Focus was on assessing smoking status early, offering all present smokers brief verbal advice complemented with written material, and pharmacological treatment optimization. Implementation was performed without increasing personnel resources – and although it´s difficult to estimate the exact cost for society – there is little doubt that a nurse-managed smoking cessation program after acute MI is cost-effective. In a study by Quist-Paulsen, cost of a nurse-led smoking intervention program compared favourably to other treatment modalities for CHD patients, being approximately 1/25 the cost of statins and angiotensin-converting enzyme inhibitors (16). In another study by Krumholz et al that evaluated the cost-effectiveness of a similar nurse-led intervention, that included three hours of extra counselling time with a nurse, was cost-effective at an estimated $220/year of life saved (17). In the present study we estimate that the extra counselling time (patient identification, counselling during admission and more telephone contacts during follow-up) to be approximately 1–2 hours per patient. More time and resources for counselling in future studies and clinical praxis is most likely warranted, also since clinical studies show a positive relationship between number of counselling sessions and abstinence rate (18, 19). To achieve optimal effect, smoking cessation programmes should preferably include wide-ranging medical information, behavioral aspects, community-oriented methods and suitable pharmaco-therapeutic treatments (8, 18, 20). Moreover, it has been proposed that telephone counseling and text messaging is effective for outpatients after a MI (21).
Out of all current smokers at baseline less than 10% were concomitantly using wet snuff during the reference and implementation periods, respectively. Use of wet snuff is common in Sweden, with 18% of men and 5% of women using wet snuff daily (5) and using wet snuff as means to quit smoking cigarettes is quite common. The small number of current wet snuff users in our study, however, limited the possibilities for meaningful sub-analyses of this group.
As the CR nurses at the participating hospitals only work office hours Monday to Friday, lack of time to identify and subsequently counsel patients during admission was probably the major barrier to offering counselling during admission to more patients, but only 50% of the patients included in the current analysis received counselling. This in turn affected the CR nurses´ possibility to improve treatment during admission and at discharge. Also, if the patients were not identified during admission, they were not contacted by telephone during the first week after discharge. Instead, first contact was at the initial CR assessment visit (2–3 weeks post discharge), as it was during the reference period and which is the general rule in Sweden. The importance of identifying patients early is supported in our data by the large difference in odds for smoking cessation when only including patients who received counselling (per protocol, OR 2.50 [1.42–4.41], p = 0.002) vs all patients treated during the implementation period (intention to treat, OR 1.60 [1.04–2.48], p = 0.034). Given the potential benefits in gained lives and cost-effectiveness of a relatively simple intervention, increasing CR nurse resources to facilitate counselling even on weekends and during odd hours should be a priority.
Sweden has a low prevalence of daily smokers around 7% yet the proportion of MI patients that smoke is 23% (5, 6). Many European countries have significantly higher prevalence of smoking among MI patients, including Spain (41%), Turkey (42%), Serbia (46%) and Cyprus (57%) (22). Accordingly, simple interventions such as the one evaluated in the current study, in countries with a high proportion of active smokers, could have major effects on the general prognosis post-MI. We encourage healthcare authorities to facilitate implementation of structured multi-component smoking cessation methods, which should be offered to every smoker who suffers an MI.
Strengths and limitations
The study has a relatively large sample size (383 patients) and several hospitals, both university hospitals and rural hospitals, were involved. All currently smoking MI patients admitted at the participating hospitals were included, increasing study representativeness. Missing data was also minimal.
Smoking status was assessed by self-report and no biochemical validation was performed to verify abstinence. Nevertheless, self-reporting is considered accurate in most smoking cessation studies (23). The duration and intensity of smoking (i.e., n packages smoke per time unit, and years of smoking) is not available via SWEDEHEART or medical records. Some of the implemented routines were already in place at the participating hospitals before start of the project, also during the reference period. This likely reduced the difference between implementation and control periods. Although the association of each component of the new routines was assessed in the secondary analysis collinearity is plausible. The short follow-up of 2 months is a limitation. A longer follow-up of 6 or 12 months would have been preferred, as abstinence at 12 months is a good predictor for long-term abstinence (24). However, as previously stated the Covid19 pandemic made a longer follow-up period unfeasible. Other factors than the implemented routines, such as public health interventions or general attitudes in society might have changed during the periods and influenced the results. However, major confounding effects of this sort seem unlikely since SWEDEHEART data has shown similar results regarding smoking cessation after MI for ten years prior to and during the study period (6).