This retrospective cohort study was conducted at Imam Ali Hospital, a tertiary cardiovascular referral center located in Western Iran. We included all patients aged ≥18 years admitted with a confirmed diagnosis of either STEMI or NSTEMI between December 2019 and August 2020. STEMI and NSTEMI diagnoses were established based on standard clinical, electrocardiographic, and biochemical criteria. Patients with a history of prior myocardial infarction (MI) were excluded to minimize confounding from recurrent ischemic events and differences in baseline risk profiles, such as variations in treatment history or disease severity. Patients with incomplete clinical data or unavailable follow-up information at one year were also excluded from the analysis.

Demographic characteristics, medical history, and cardiovascular risk factors were collected at the time of hospital admission via structured interviews by trained nurses and validated against the hospital’s electronic medical records. Clinical data included prior myocardial infarction, revascularization history, presence of hypertension, diabetes mellitus, and hyperlipidemia. Vital signs, laboratory tests (lipid profile, serum creatinine), and body mass index (BMI) were recorded upon admission. Estimated glomerular filtration rate (eGFR) was estimated using the CKD-EPI formula. Blood pressure was categorized into three groups: ˂112 mmHg, 112–140 mmHg, and >140 mmHg based on prior literature.^[10]

Smoking burden was assessed using the pack-year index, calculated by multiplying the number of cigarette packs smoked per day by the number of years the patient had smoked. Patients were categorized as never-smokers (0 pack-years), moderate smokers (≤15 pack-years), or heavy smokers (>15 pack-years) based on prior literature.^[11] Smoking status was self-reported at admission due to the absence of biochemical verification.

The primary outcome was all-cause mortality within one year of AMI diagnosis, including in-hospital and post-discharge deaths. In-hospital mortality was documented via hospital records. Contact information for patients or their family members was recorded at admission to facilitate follow-up. Patients were contacted by phone at one year. For reported deaths, clinical and hospital records, including cause of death, were reviewed. Loss to follow-up was minimal (1.2%, 12/1,019), primarily due to outdated contact information. The follow-up period was defined as the time from AMI diagnosis to death, loss to follow-up, or 365 days, whichever occurred first.

To minimize bias, several measures were implemented. Selection bias was reduced by including all eligible patients during the study period and excluding those with prior MI to limit confounding from recurrent events. Information bias was addressed by validating self-reported data (e.g., smoking status, medical history) against electronic medical records where possible. For smoking status, structured interviews used standardized questions to reduce recall bias, though biochemical validation was not feasible. Missing data (<5% per variable) were handled via list wise deletion, with sensitivity analyses excluding cases with missing eGFR values to assess robustness. Multivariable adjustment in Cox models accounted for key confounders (age, sex, diabetes, hypertension, BMI, lipid profiles, eGFR, reperfusion therapy, SBP). No imputation was performed due to low missingness, but patterns of missing data were assessed to ensure they were not systematically related to outcomes.

All patients signed a written informed consent before enrolling in the study. The Research Ethics Committee at the Deputy of Research of the Kermanshah University of Medical Sciences approved the study protocol (Ethics registration code: IR.KUMS.REC.1400.252).

Continuous variables were presented as mean ± standard deviation (SD) and categorical variables as absolute value and percentage. Baseline characteristics were compared using chi-square and ANOVA as appropriate. Cox proportional hazards regression models were constructed separately for STEMI and NSTEMI to estimate hazard ratios (HRs) and 95% confidence intervals (CIs) for all-cause mortality. Three models were assessed: crude, age-adjusted, and fully adjusted models including all covariates. To address potential sources of bias, we used multivariable adjustment for confounders and excluded patients with prior MI. Missing data were minimal (<5% per variable) and handled via listwise deletion in analyses. No formal interaction tests were performed, but subtype-specific models served as subgroup analyses. Sensitivity analyses were conducted by excluding cases with missing GFR values, yielding similar results.

This study followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines.^[12]

Of 1,200 potentially eligible patients, 1,019 were included after excluding 100 with prior MI and 81 with incomplete data (missing follow-up or lab results) (Fig. 1). Loss to follow-up was 12 (1.2%). Among the 1,019 patients, 645 (63.3%) had STEMI and 374 (36.7%) had NSTEMI. Missing data were minimal (<2% for age/sex, 4% for eGFR).

Among 645 STEMI patients (356 never-smokers, 124 moderate smokers, 165 heavy smokers), smokers were younger (mean age: 55.53±11.81 years for ≤15 pack-years, 57.63±9.89 years for >15 pack-years vs. 62.38±13.33 years for never-smokers; p<0.001) and predominantly male (>95% vs. 63.2%; p<0.001). Smokers had lower BMI (25.50±4.28 kg/m² for >15 pack-years vs. 26.81±4.68 kg/m²; p=0.034), lower SBP and DBP (p=0.002), and lower prevalence of hypertension (22.4% vs. 47.9%; p<0.001) and diabetes (10.9% vs. 29.5%; p<0.001). eGFR was higher among smokers (p<0.001). Educational attainment was lower among smokers (p<0.001). One-year mortality was lower among smokers (7.3% for ≤15 pack-years, 6.1% for >15 pack-years vs. 13.2%; p=0.021). No differences were observed in income level or reperfusion therapy (Table 1).

Among 374 NSTEMI patients (253 never-smokers, 58 moderate smokers, 63 heavy smokers), age did not differ significantly (p=0.339). Male sex was more prevalent among smokers (87.9% for ≤15 pack-years, 100% for >15 pack-years vs. 56.1%; p<0.001). Hypertension (31.8% vs. 54.2%; p=0.003) and diabetes (7.9% vs. 26.9%; p=0.002) were less prevalent among smokers. eGFR was higher among smokers (p=0.029). Educational attainment was lower among smokers (p<0.001), but income level and reperfusion therapy did not differ. One-year mortality was similar across groups (p=0.761) (Table 2).

In the unadjusted model, patients with a smoking history >15 pack-years had a significantly lower risk of one-year mortality compared to never-smokers (HR: 0.46; 95% CI: 0.23–0.91). Those with <15 pack-years also showed a lower, though non-significant, risk (HR: 0.54; 95% CI: 0.26–1.19). After adjusting for age, the protective association was attenuated and no longer statistically significant (HR for >15 pack-years: 0.58; 95% CI: 0.29–1.16). In the full-adjusted model, which controlled for age, sex, diabetes, hypertension, BMI, lipid profiles, estimated eGFR, reperfusion therapy, and SBP, the association remained non-significant (HR: 0.75; 95% CI: 0.24–2.31 for >15 pack-years). Among the covariates, age was consistently associated with increased mortality risk across models (full-adjusted HR: 1.07; 95% CI: 1.02–1.12). Diabetes was a significant predictor in the crude and age-adjusted models but not in the full-adjusted model. eGFR was inversely associated with mortality in unadjusted (HR: 0.97; 95% CI: 0.96–0.98) and age-adjusted models (HR: 0.98; 95% CI: 0.96–0.99), though the association lost significance after full adjustment. SBP >140 mmHg was associated with reduced mortality in unadjusted (HR: 0.43; 95% CI: 0.23–0.78) and age-adjusted models (HR: 0.42; 95% CI: 0.23–0.77), but this effect was not observed in the full-adjusted model (HR: 1.31; 95% CI: 0.36–4.78). Reperfusion therapy was significantly associated with reduced mortality in the crude model (HR: 0.50; 95% CI: 0.29–0.85), but this association was not sustained after full adjustment (Table 3).

Among NSTEMI patients, no significant association was observed between smoking and one-year mortality across all models. In the crude model, the HR for <15 pack-years was 0.49 (95% CI: 0.11–2.15) and 1.00 (95% CI: 0.33–2.97) for >15 pack-years. These estimates remained non-significant after age adjustment (HR: 0.56 and 1.12, respectively) and in the full-adjusted model (HR: 0.14; 95% CI: 0.00–3.90 for <15 pack-years and HR: 1.10; 95% CI: 0.27–4.46 for >15 pack-years). Age was a significant predictor of mortality in the unadjusted model (HR: 1.06; 95% CI: 1.03–1.10), though this association weakened in the full-adjusted model (HR: 1.05; 95% CI: 0.98–1.14). eGFR remained a robust inverse predictor of mortality in all models, including the full-adjusted model (HR: 0.95; 95% CI: 0.90–0.99). Reperfusion therapy was associated with significantly reduced mortality in both the crude (HR: 0.12; 95% CI: 0.01–0.92) and age-adjusted (HR: 0.13; 95% CI: 0.017–0.96) models, though this effect was attenuated in the full-adjusted model (HR: 0.86; 95% CI: 0.08–8.41). Lower systolic blood pressure (<112 mmHg) was associated with increased mortality compared to higher SBP categories in unadjusted and age-adjusted models; however, these associations lost significance after full adjustment (Table 4).

Figs 2 and 3 show the survival curves for heavy and moderate smokers versus never-smokers based on the full-adjusted Cox regression model.