Article Text

Download PDFPDF
A retrospective cross-sectional study on the association between tobacco smoking and incidence of ST-segment elevation myocardial infarction and cardiovascular risk factors
  1. Lloyd Steele1,
  2. Amelia Lloyd1,
  3. James Fotheringham2,
  4. Ayyaz Sultan3,
  5. Javaid Iqbal1,
  6. Ever D Grech3
  1. 1University of Sheffield, Sheffield, UK
  2. 2School of Health and Related Research (ScHARR), University of Sheffield, Sheffield, UK
  3. 3South Yorkshire Cardiothoracic Centre, Northern General Hospital, Sheffield UK
  1. Correspondence to Dr Ever D Grech, South Yorkshire Cardiothoracic Centre, Northern General Hospital, Herries Road, Sheffield S5 7AU, UK; Ever.Grech{at}


Background Cigarette smoking is a well-established risk factor for the development of coronary heart disease. However, the relationship between smoking and acute ST-segment elevation myocardial infarction (STEMI) is less well described.

Objective To determine the relative risk of acute STEMI in smokers and ex-smokers, compared with individuals who had never smoked.

Methods This observational study studied all patients with STEMI undergoing percutaneous coronary intervention (PCI) in South Yorkshire, UK from 1 January 2009 to 6 April 2012. Additional contemporary demographical data for the South Yorkshire population, supplied by the Office for National Statistics, allowed derivation of the incidence rate of STEMI in South Yorkshire—both overall and stratified by smoking status. Incidence rate ratios and population attributable risk (PAR) were calculated to quantify STEMI risk.

Results There were 1715 STEMIs in 1680 patients during the study period. Smoking status was obtained in 96.2% patients. The prevalence of smoking was 47.3% in patients with STEMI and 22.0% in the general population. In patients with STEMI, smokers were ∼10 years younger, mean (SD) 57.2 (11.1) years, than never-smokers, 66.4 (12.1) years, and ex-smokers, 67.9 (11.9) years. The age-standardised incident rate ratio of STEMI was 5.2 (4.5–6.1) for current smokers and 1.1 (1.0–1.3) for ex-smokers, with the reference group being never-smokers for both. Almost 50% of STEMIs were attributable to smoking (PAR=48.3%).

Conclusion Cigarette smoking is associated with a fivefold increased risk of STEMI. Smoking cessation reduced this risk to a level similar to never-smokers.

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


Coronary heart disease (CHD) is the leading cause of death globally.1 However, >90% of myocardial infarctions (MIs) are potentially avoidable and attributable to cardiovascular risk factors.2 One of the most important of these risk factors is cigarette smoking.2

In high-income countries, the prevalence of smoking has been falling for decades, and this fall has been cited as most important factor in the decline in CHD incidence observed over this time.3 Despite this improvement, smoking prevalence remains high in many of these countries.4 However, with a wealth of evidence showing that smoking cessation significantly reduces CHD risk,5–8 this suggests that continuing to promote smoking cessation in the future has the potential to yield significant reductions in CHD incidence.

The incidence of ST-segment elevation myocardial infarction (STEMI) also appears to be falling,9 ,10 but whereas the link between smoking and CHD has been extensively studied, the relationship between smoking and acute STEMI is less well described. Moreover, the current impact of smoking on STEMI incidence specifically is not known. Population attributable risks (PARs) from the INTERHEART study assessed patients with acute MI, rather than the current STEMI. designation.

In our study we therefore sought to delineate the relative risk of acute STEMI in both smokers and ex-smokers relative to never-smokers, and to quantify the impact of smoking in a contemporary population by calculating the PAR of cigarette smoking for STEMI. It was also hoped that this would contribute to publicising the dangers of smoking in a contemporary population. This is necessary as awareness appears to be low. Even though smoking remains the single greatest cause of preventable illness and premature death in countries such as the USA and the UK,11 ,12 <40% of patients undergoing coronary angiography recognise smoking as a risk factor for CHD,13 and more than one in three smokers believe the dangers of smoking to be exaggerated.14


Identification of patients admitted with STEMI

All patients admitted with STEMI undergoing primary percutaneous coronary intervention (PCI) in the South Yorkshire region from 1 January 2009 to 6 April 2012 were included in the study. The sample size used represented the time period of the study, which was based on the introduction of primary PCI in the South Yorkshire region.

Patients were identified from the departmental coronary intervention database. Patient records were used to obtain: the date of STEMI; age at time of STEMI; gender; postcode, which was used as a proxy for socioeconomic status (SES) and smoking status (smoker, ex-smoker, never-smoker), which was defined as documented in the patient notes at the time of STEMI. We also recorded the presence of hypertension, dyslipidaemia, diabetes mellitus, a family history of CHD, previous CHD, chronic kidney disease stage IV or stage V, previous transient ischaemic attack or stroke, peripheral vascular disease, body mass index and drug use on admission (aspirin, P2Y12 inhibitor, statin, ACE inhibitor or angiotensin receptor blocker and β-blocker). These additional risk factors were not adjusted for in risk analyses due to the nature of the analysis, but this information did allow us to compare the risk factor profile of smokers, ex-smokers and never-smokers.

Population data

The South Yorkshire region was defined based on borders provided by the Office for National Statistics (ONS). The population at risk of STEMI in this region was determined by population estimates provided by the ONS based on annual Integrated Household Survey (IHS) data. These data were available stratified by smoking status.

What is the IHS?15

▸ The largest social survey collected by the ONS, which uses information from both the ‘Annual Population Survey’ and ‘Living Cost and Food Survey’.

▸ It provides estimates from ∼340 000 individual respondents—the biggest pool of UK social data after the census. It is conducted in the form of a questionnaire with responses collected anonymously.

▸ It is designed to gather information for a representative of households. Individual responses are weighted according to population counts for given areas to represent the overall population.

▸ Topics covered by the IHS include perceived general health, smoking prevalence, education, housing and employment.

Study region: South Yorkshire

Population was approximated at 1.4 million in 2013.16

Acute hospital for primary PCI service: South Yorkshire Cardiothoracic Centre, Northern General Hospital, Sheffield.

Data analysis

Continuous variables are expressed as a mean and SD and categorical variables are shown as counts and proportions. The main variable of interest was smoking status.

Patient postcodes were queried against the ONS postcode service to derive index of multiple deprivation (IMD) values for each patient. IMD values are frequently used in describing and monitoring socioeconomic inequalities in health as they are unified measure of SES.17

Using IHS data, we calculated the population at risk of STEMI in the South Yorkshire region by calculating person-time at risk. Person-time at risk was employed to account for varying population counts and the expanding catchment area for the PCI service during the period studied. We excluded patients aged <18 years from the incidence analysis as IHS data provided information on only those aged ≥18 years. This excluded one patient.

Statistical analyses

We calculated the incidence rate of acute STEMI both overall and by smoking status per 100 000 person-years. Incidence rates calculated were age-standardised by direct standardisation of the frequencies to the overall South Yorkshire age distribution with a three-level age stratification (<50 years, 50–65 years, >65 years). Person-time at risk was employed to account for varying population counts and the expanding catchment area for the PCI service during the period studied, and was calculated from annual IHS responses with weighting applicable for the year in question. Missing data on smoking status were 3.8% for both the STEMI group and South Yorkshire population and so was ignored.

Incidence rates were used to calculate incidence rate ratios, a recognised measure of relative risk in epidemiological studies, using the equation below.18 Incidence rate ratios were age-standardised with indirect standardisation, again with a three-level age stratification (<50 years, 50–65 years, >65 years). Embedded Image

PAR was calculated to show the percentage of STEMIs that could be attributed to smoking using the following equation.18Embedded Image where ‘p’ is the proportion of the total population exposed to the risk factor’ (ie, prevalence of the risk factor) and incidence rate ratio is a measure of relative risk.

To assess the differences in risk factors between smoking groups, multinomial logistic regression was used for categorical variables, and one-way analysis of variance was used for continuous variables.

SPSS V. (SPSS, Chicago, Illinois, USA) was used for all statistical analyses with the exception of the calculation of incidence rate ratios, performed using STATA V.12 (StataCorp LP, College Station, Texas, USA).

Ethical committee position

At the time of study set up, National Health Service (NHS) Research and Ethics Committee approval was not required for research involving previously collected data extracted from hospital records and rendered non-identifiable by the direct care team before being used for research purposes. NHS permission was gained for this study via the Sheffield Teaching Hospitals Research and Development Department to ensure the study's compliance with the Data Protection Act and to protect patient confidentiality.



A total of 1715 acute STEMIs were treated with PCI in 1680 patients from 1 January 2009 to 6 April 2012. A total of 1239 STEMIs (72.2%) occurred in men, and 476 (27.8%) occurred in women. The mean age was 62.6 years. A total of 114 patients (6.6%) were deceased at 1-year post-STEMI. Smoking status was obtained in 96.2% of cases (1650 of 1715 cases).

Cardiovascular risk factors

On admission, 47.3% of patients with STEMI were smokers, compared with 22.0% in the general South Yorkshire population. For ex-smokers, the corresponding prevalences were 26.1% and 29.6% in the STEMI cohort and the general population, respectively. For never-smokers, these were 22.8% and 44.7% (table 1).

Table 1

Incidence rate of ST-segment elevation myocardial infarction (STEMI) by smoking status in South Yorkshire between 2009 and 2012

In patients with STEMI, smokers were ∼10 years younger, mean (SD) 57.2 (11.1) years, than never-smokers, 66.4 (12.1) years, and ex-smokers, 67.9 (11.9) years. Risk factors were generally of a similar or lower prevalence in current smokers compared with ex-smokers and never-smokers. The notable exception was SES, with smokers being significantly more deprived on average than ex-smokers and never-smokers (table 2). The use of preventive medication was generally higher in ex-smokers than smokers or never-smokers (table 3).

Table 2

Cardiovascular risk factor prevalence in STEMI cohort by smoking status

Table 3

Drug use at time of admission for STEMI

Relationship between smoking status and STEMI risk

The incident rate ratio of STEMI (95% CI) was 5.2 (4.5 to 6.1) for current smokers and 1.1 (1.0 to 1.3) for ex-smokers. There was no significant difference in risk when incidence rate ratios (95% CI) were calculated for each gender (male 3.0 (2.7 to 3.3), female 3.5 (2.9 to 4.2)).

The PAR of smoking was 48.3%, suggesting that almost 50% of STEMIs were attributable to smoking. The PAR of ex-smoking was 4.7%.


Although there is a wealth of information on CHD and smoking, this is one of only a few studies focusing specifically on acute STEMI. Our study indicates that smokers are over five times more likely to suffer an acute STEMI than never-smokers, whereas ex-smokers have a risk similar to never-smokers.

This study carries an important message to those smokers who may be unaware of the marked cardiovascular benefit of smoking cessation. This is important as smoking remains a serious problem in cardiovascular medicine, with nearly half of all patients with STEMI being current smokers (47.3%) in our study, which was more than double the prevalence of the South Yorkshire population (22.0%). Other studies have also reported a similarly high prevalence of smoking in STEMI cohorts (43%–46%),19–21 although others have found a lower prevalence (31%–35%).22 ,23 One study conducted in Michigan was especially similar to ours and also assessed smoking prevalence both in the STEMI cohort (46.4%) and the general population (20.5%).21

The PAR of smoking in this study suggests that almost 50% of STEMIs were attributable to smoking (PAR 48.3%). This is higher than the PAR calculated from INTERHEART data (38%).24 The main reasons for this discrepancy are the INTERHEART definition of acute MI, which differs from the current STEMI classification; differences in the gender ratio, with the INTERHEART study finding a large variation in PAR between men (PAR=44%) and women (PAR=6%) and differences in our analysis, with our PAR unadjusted for other risk factors.

The relative risk reported for smoking in this study (5.2; 95% CI 4.5 to 6.1) is above the range quoted in a systematic review by Critchley and Capewell for CHD (1.5–3.0).25 This could be due to a number of reasons. First, this study focused specifically on patients with STEMI rather than CHD more broadly, and thus our finding could suggest that smoking is a greater risk factor for STEMI than other CHD presentations. The only other STEMI-specific study identified, by Larsen et al,21 also reported a relative risk of STEMI in smokers (3.4; 95% CI 3.3 to 3.4) that was higher than the range quoted for CHD. This seems biologically plausible as non-STEMI is generally associated with platelet-rich, partially occlusive thrombus, whereas STEMI is associated with fibrin-rich, more-occlusive thrombus.26 It is possible that the antifibrinolytic effect of smoking is a prominent one, thus leading to larger, fibrin-rich thrombi,27 resulting in a greater likelihood of complete vessel occlusion and STEMI.

Second, the South Yorkshire population may smoke more heavily relative to other populations, with a dose–response effect well recognised with smoking—where relative risk increases with both the number of cigarettes smoked per day and the number of smoking years.24 ,28

Finally, this study was conducted in a more contemporary population. The significance of this is that it has recently been reported that the relative risk of CHD for smoking has increased with time, as judged by differences in large population studies from 1959 to 1965, 1982 to 1988 and 2000 to 2010.29 It has been suggested that this may be because women started smoking later than men, and thus compared with previous generations, women in contemporary CHD cohorts would have taken up smoking earlier and in greater numbers. With the aforementioned dose–response effect of smoking, their risk of CHD would be higher than the risk of CHD in women in antecedent studies.

Our relative risk for ex-smoking (1.1; 95% CI 1.0 to 1.3) is consistent with most studies for MI or CHD, which report a risk between 1.0 and 1.3,6 ,7 ,30 although a minority have reported a higher risk of >1.5.28 ,31 ,32 The low risk observed is biologically plausible, with smoking-induced changes (such as endothelial function, coagulation parameters and inflammatory biomarkers) declining within weeks of cessation.33 Our risk for ex-smokers is possibly an underestimate due to the probability of reverse causality, whereby smokers may have ceased smoking due to previous CHD symptoms. This is suggested by the high prevalence of previous CHD in ex-smokers and has also been observed in other studies, such as the 50-year follow-up of British doctors, which reported that 29.2% of ex-smokers quit because of vascular disease.34 However, in addition to an observed high prevalence of previous CHD in ex-smokers, it was also observed that the use of preventative medications in ex-smokers was significantly higher than in never-smokers. This may in turn have reduced their acute STEMI risk.

The main limitation of this study was the reliance on the accuracy and validity of routine data. Smoking statuses were dependent on patients’ self-reporting, and there is evidence that patients under-report smoking, or deny smoking altogether, when asked about a smoking-related illness.35 However, not all studies have identified under-reporting.36 ,37 Moreover, more detailed smoking history—namely smoking years, number of cigarettes smoked per day and, if applicable, time of smoking cessation—was also reliant on patient self-reporting and was often not available in patient notes. Due to this, there was no minimum time used for defining an ex-smoker. The impact of this is not clear due to differing estimates of the temporal effect of smoking cessation, with some studies reporting than CHD risk falls quickly,8 ,27 ,33 ,38 while others have reported that it remains elevated for >10–20 years.28 ,30 The reliance on routinely recorded information, together with the retrospective nature of this study, also made it difficult to determine the indication for cardiovascular drugs used on admission, as well as introducing the possibility of observer bias, with the presence of cardiovascular risk factors at time of STEMI recorded by many different individuals.

The population data provided by ONS may have been prone to selection bias as it was based on a survey that was not compulsory. However, ONS employ a number of techniques to minimise the impact of non-response.15 Other issues with these data include inaccurate population estimates and the possibility of patients in the STEMI database (incident cases) also being repeated in ONS data (prevalent cases).

The results of our study should be interpreted with the following caveats. This study did not include acute patients with STEMI who did not undergo PCI. These may include patients who died in transit, were too frail or who had severe comorbidities. However, it is likely that failure to include these patients did not affect the results to any material extent as the relation of smoking status to fatal disease appears quite similar to its relation to non-fatal illness.39 Additionally, as we have conducted multiple analyses on a single data set, this introduces the possibility of an α-inflation error. Finally, we did not adjust for other cardiovascular risk factors in our relative risk and PAR calculations. Although the pathogenesis of STEMI is often multifactorial and risk factors can act synergistically, the magnitude of these values suggests that doing so would not entirely attenuate the relationship between smoking and STEMI.

Our study provides evidence that smoking acts as an important risk factor in the development of STEMI in a contemporary population. In future, larger studies are needed to help delineate the differences in risk of STEMI by smoking status in males and females, and with adjustment for other cardiovascular risk factors. The role of SES and its association with STEMI and smoking also needs to be investigated, as this will be important in elucidating how best to reduce smoking prevalence and thus STEMI incidence.

Main messages

  • The risk of STEMI is over five times higher in smokers relative to non-smokers.

  • Smoking cessation significantly reduces acute STEMI risk to a similar level to never-smokers.

  • Almost 50% of STEMIs in the South Yorkshire region could be avoided if the population did not smoke.

Current research questions

  • Does the relationship between smoking and STEMI differ from the well-established relationship between smoking and coronary heart disease?

  • Are new targeted smoking interventions needed to reduce smoking prevalence?

  • What is the temporal relationship between smoking cessation and STEMI risk?


The authors thank the Office for National Statistics for providing data to support this study, and Katherine Everard (Department of Health) for advice on appropriate data sources. The authors also thank Dr Dawn Teare for providing statistical advice.



  • Contributors EDG conceived the idea of the study and all authors contributed to its design. JF was responsible for obtaining and analysing South Yorkshire population data from the Office for National Statistics. LS and AL were responsible for collecting data regarding patients with STEMI. LS and JF analysed the data, with interpretation from LS, JF, JI and EDG. LS, JI and EDG contributed to drafting of the article. EDG is guarantor. All authors had access to all of the data in the study and can take responsibility for the integrity of the data and the accuracy of the data analysis. The guarantor affirms that the manuscript is an honest, accurate and transparent account of the study being reported, that no important aspects have been omitted and that any discrepancies from the study as planned have been explained.

  • Competing interests None declared.

  • Ethics approval At the time of study set up NHS REC approval was not required for research involving previously collected data extracted from hospital records and rendered non-identifiable by the direct care team before being used for research purposes. NHS permission was gained for this study via the Sheffield Teaching Hospitals Research and Development Department to ensure the study's compliance with the data protection act and protect patient confidentiality.

  • Provenance and peer review Not commissioned; externally peer-reviewed.