Ten-year trends in cardiovascular risk factors among patients hospitalised with acute myocardial infarction in Estonia: a retrospective nationwide observational study

Abstract

Introduction

Cardiovascular risk-factor profiles among patients hospitalised with acute myocardial infarction (AMI) can inform surveillance of a high-risk clinical cohort, but do not directly represent risk-factor prevalence in the general population.

Methods

We analysed aggregate annual Estonian AMI registry data for hospitalised AMI patients during 2015–2024. AMI04 risk-factor trends were assessed using ordinary least squares regression, with sensitivity analyses for known status, age-sex standardisation to the pooled hospitalised AMI cohort, COVID-period exclusion, AMI case mix, and national hospitalised AMI burden.

Results

Documented smoking increased from 25.5% in 2015 to 28.1% in 2024 (β = 0.302% points/year; 95% CI 0.014 to 0.590; p = 0.042). The smoking trend remained significant in known-status analyses and after age-sex standardisation, although the crude observed trend became borderline after excluding 2020–2021. Documented dyslipidaemia increased descriptively, but full-period crude and age-sex-standardised trends were borderline, and unknown dyslipidaemia status decreased over time. Annual AMI attacks and national hospitalised AMI burden declined, with concurrent changes in sex and age composition.

Conclusion

Among hospitalised AMI patients in Estonia, documented smoking increased during 2015–2024 and remained consistent in known-status and age-sex-standardised sensitivity analyses. Dyslipidaemia trends require cautious interpretation because unknown status changed over time. These aggregate registry findings describe a selected hospitalised AMI cohort and should not be interpreted as direct evidence of national prevention-policy success or failure.

Graphical Abstract. Cardiovascular risk in patients with acute myocardial infarction in Estonia, 2015–2024

This graphical abstract summarises the key findings from a decade-long aggregate registry analysis of hospitalised AMI patients. Smoking increased in the AMI cohort and remained consistent in known-status and age-sex-standardised sensitivity analyses. Dyslipidaemia trends require cautious interpretation because unknown status changed over time. These findings describe hospitalised AMI patients and should not be interpreted as direct evidence of national prevention-policy success or failure

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1186/s12872-026-06223-8.

Introduction

Acute myocardial infarction (AMI) remains a leading cause of morbidity and mortality worldwide despite substantial advances in acute care and secondary prevention [1, 2]. Although short-term survival following AMI has improved considerably [3], recurrent cardiovascular events remain closely linked to the persistence and evolving distribution of modifiable risk factors, including hypertension, smoking, dyslipidaemia, diabetes, and excess body weight [4–6].

Understanding how documented risk factors change over time among patients presenting with AMI can help identify risk profiles within a high-risk hospitalised cohort and generate hypotheses about prevention, documentation, and case-mix patterns. However, such analyses cannot directly evaluate prevention effectiveness in the general population.

Across Europe, trends in cardiovascular mortality and risk factor burden have been heterogeneous. Western European countries have demonstrated sustained improvements in smoking reduction and lipid management [7, 8], whereas Eastern European and Baltic countries have historically experienced higher cardiovascular mortality and a greater burden of cardiometabolic risk factors [9]. These disparities likely reflect differences in socioeconomic development, healthcare system organisation, and implementation of preventive policies. However, longitudinal trends in modifiable cardiovascular risk factors among patients hospitalised with AMI have not been comprehensively evaluated at the national level in Baltic countries.

Estonia has undergone substantial healthcare reform and economic transition over the past two decades [10], accompanied by expanded preventive strategies and improvements in cardiovascular care. Whether risk-factor profiles among hospitalised AMI patients have changed during the most recent decade remains clinically relevant, particularly when interpreted alongside AMI denominator trends, national AMI burden, missingness, and population context.

Therefore, this study evaluated national temporal trends in major modifiable cardiovascular risk factors among patients hospitalised with AMI in Estonia from 2015 to 2024. We assessed crude observed trends, unknown-status and known-status analyses, age-sex-standardised trends, COVID-period sensitivity, AMI case mix, national hospitalised AMI burden, and contextual population indicators.

Methods

Data source and study population

This retrospective longitudinal analysis used aggregated annual data from the Estonian Health Statistics and Health Research Database [11]. The primary analytic cohort comprised hospitalised patients with acute myocardial infarction (AMI), not the general Estonian population. AMI risk-factor analyses were based on AMI04, which reports documented risk-factor status by calendar year, sex, and age group. AMI02 was used to describe annual AMI denominators, sex and age case mix, attack type, and AMI subtype. AMI01 was used to assess national hospitalised AMI burden, including hospitalised persons with AMI and morbidity rates per 100,000 inhabitants.

Outcome variable

The primary AMI04 outcomes were annual documented yes-category percentages for hypertension, smoking, dyslipidaemia, diabetes, and overweight. Because AMI04 also includes no and unknown status categories, we calculated unknown-category trends and known-status prevalence as yes / (yes + no) * 100. These measures describe documented risk-factor status among hospitalised AMI patients and should not be interpreted as risk-factor prevalence in the Estonian general population.

Registry risk-factor definitions

The analyses used the registry yes/no/unknown categories as reported in AMI04. Official risk-factor definitions were taken from the Estonian AMI registry metadata.

Hypertension: arterial hypertension diagnosed previously or during the current hospitalisation, based on at least two resting measurements on separate health care visits with systolic blood pressure > 140 mmHg and/or diastolic blood pressure > 90 mmHg.

Smoking: regular smoking of any type of tobacco during the month before the current hospitalisation.

Diabetes: diabetes diagnosed previously or during the current hospitalisation according to diagnostic criteria.

Dyslipidaemia: dyslipidaemia diagnosed previously or during the current hospitalisation and/or physician-treated dyslipidaemia, LDL-cholesterol ≥ 2.6 mmol/L, or continued use of lipid-lowering drugs regardless of LDL-cholesterol.

Overweight: BMI ≥ 25 kg/m² for persons younger than 65 years and BMI ≥ 30 kg/m² for persons aged 65 years or older. This age-dependent threshold was predefined by the registry metadata. Because the metadata did not provide a detailed clinical rationale for its selection, we treated it as an operational registry definition rather than a universally applicable clinical definition.

Data extraction

Data were extracted across predefined strata: national aggregate estimates (“Men and women”), sex-specific categories (“Men” and “Women”), and age-stratified groups (15–54, 55–64, 65–74, 75–84, and 85 years and older). Population context was assessed separately and descriptively using TKU50 for smoking among adults aged 16–64 years, TKU40 for BMI categories among adults aged 16–64 years, and ETU30 Health Interview Survey outputs for diabetes, hypertension/high blood pressure, and high cholesterol. These population datasets were used only for contextual comparison and were not treated as directly equivalent to the hospitalised AMI cohort.

Statistical analysis

Descriptive summaries, including the 2024 age- and sex-specific profile, were used to characterise contemporary risk-factor patterns. The primary inference was based on longitudinal AMI04 trend models. Annual trends were analysed using ordinary least squares (OLS) regression with calendar year as a continuous predictor. Slopes are reported as percentage-point changes per year with 95% confidence intervals and p-values.

The primary AMI04 analysis evaluated crude observed yes-category trends among men and women combined and all AMI age groups. Additional AMI04 analyses included sex-specific trends, age-specific trends, unknown-status trends, and known-status sensitivity analyses. Direct age-sex standardisation was performed using fixed pooled 2015–2024 hospitalised AMI cohort weights derived from AMI02. This was not standardisation to the Estonian general population; it was direct standardisation to the pooled hospitalised AMI cohort distribution.

COVID-period sensitivity analyses were performed by re-estimating AMI04 trend models after excluding 2020 and 2021.

Exploratory joinpoint analyses allowed a maximum of one joinpoint and required at least three observations per segment. Segments with three or fewer observations were explicitly flagged as having limited robustness. Because only ten annual observations were available, joinpoint findings were treated as hypothesis-generating and not confirmatory.

Model diagnostics for the overall observed yes-category OLS models included residual-versus-fitted plots, Q-Q plots, Durbin-Watson tests for residual autocorrelation, Shapiro-Wilk tests for residual normality, and Breusch-Pagan tests for heteroscedasticity. Diagnostic results for the overall observed yes-category OLS models are provided in Supplementary Table S1.

Analyses were conducted in R using reproducible scripts. Statistical tests were two-sided, and p < 0.05 was considered statistically significant.

Results

Overall risk-factor trends and 2024 age-sex profile

During 2015–2024, AMI02 recorded 26,469 hospitalised AMI attacks in Estonia. Annual AMI attacks declined from 2,825 in 2015 to 2,391 in 2024. The proportion of women among AMI attacks declined from 42.5% to 38.3%, and the age distribution shifted over time. Annual denominators, AMI burden, case mix, age composition, and AMI subtype distributions are provided in Supplementary Table S2. Overall crude observed AMI04 risk-factor trends are summarised in Table 1, with detailed 2024 age- and sex-specific values provided in Supplementary Table S3.

Table 1.

Overall crude observed AMI04 risk-factor trends among hospitalised AMI patients, 2015–2024

Risk factor	2015 (%)	2024 (%)	Absolute change, pp	Slope β, pp/year	95% CI	p-value
Hypertension	81.6	83.3	1.7	0.084	-0.159 to 0.327	0.447
Smoking	25.5	28.1	2.6	0.302	0.014 to 0.590	0.042*
Overweight	38.0	40.0	2.0	-0.030	-0.358 to 0.297	0.836
Dyslipidaemia	60.3	70.7	10.4	0.701	-0.049 to 1.452	0.063
Diabetes	23.6	25.8	2.2	0.059	-0.115 to 0.234	0.455

Values are documented AMI04 yes-category percentages among hospitalised AMI patients. β represents annual percentage-point change from ordinary least squares regression.

pp percentage points, CI confidence interval

*p < 0.05

Overweight prevalence peaked in middle-aged patients and declined in advanced age, reflecting the registry-specific age-dependent BMI definition. Dyslipidaemia and hypertension remained highly prevalent in 2024. Smoking was substantially more common in men than women and was concentrated in younger and middle-aged AMI patients.

Overall trends and sensitivity analyses

In crude overall AMI04 analyses, smoking was the only risk factor with a statistically significant full-period increase, rising from 25.5% in 2015 to 28.1% in 2024 (β = 0.302% points/year; 95% CI 0.014 to 0.590; p = 0.042). Hypertension, diabetes, overweight, and dyslipidaemia did not reach p < 0.05 in the main crude observed models. Dyslipidaemia increased descriptively from 60.3% to 70.7%, but the full-period trend was borderline (β = 0.701% points/year; 95% CI -0.049 to 1.45; p = 0.063).

Unknown status changed over time and varied by risk factor. Unknown smoking status increased from 12.9% in 2015 to 17.7% in 2024 (0.738% points/year; p = 0.0012), while unknown dyslipidaemia status decreased from 15.5% to 6.8% (-0.666% points/year; p = 0.0043). In known-status analyses, smoking increased from 29.2% to 34.1%, with an annual slope of 0.619% points (95% CI 0.284 to 0.954; p = 0.0028). Annual AMI04 yes/no/unknown risk-factor status and known-status prevalence values are provided in Supplementary Table S4.

After direct age-sex standardisation to the pooled 2015–2024 hospitalised AMI cohort distribution, smoking remained statistically significant. The standardised observed smoking trend was 0.309% points per year (95% CI 0.052 to 0.565; p = 0.024), and the standardised known-status smoking trend was 0.516% points per year (95% CI 0.245 to 0.787; p = 0.0023). Dyslipidaemia remained borderline after standardisation (0.742% points per year; 95% CI -0.004 to 1.49; p = 0.051).

After excluding 2020 and 2021 (COVID-period sensitivity analysis), the crude observed smoking trend became borderline (0.295% points per year; p = 0.086). However, known-status smoking remained statistically significant (0.606% points per year; p = 0.011), and age-sex-standardised known-status smoking also remained significant (0.516% points per year; p = 0.0097). Dyslipidaemia observed trends became statistically significant after excluding 2020–2021 in crude and standardised analyses, but these findings should be interpreted cautiously because dyslipidaemia unknown status changed substantially across the study period. Crude, known-status, age-sex-standardised, and COVID-excluded sensitivity analyses are summarised in Supplementary Table S5.

AMI burden and population context

Annual AMI attacks declined from 2,825 in 2015 to 2,391 in 2024. The proportion of women among AMI attacks declined from 42.5% to 38.3% and showed a statistically significant downward trend. Age distribution also changed, with a significant increase in the 65–74-year age group and significant decreases in the 55–64 and 75–84-year groups. STEMI and NSTEMI proportions did not change significantly over time, suggesting that broad STEMI/NSTEMI redistribution was unlikely to explain the observed AMI04 smoking trend.

AMI01 analyses showed that national hospitalised AMI burden declined during 2015–2024. The number of hospitalised persons with AMI decreased from 2,726 in 2015 to 2,327 in 2024. The crude morbidity rate decreased from 246.9 to 201.3 per 100,000 inhabitants (-4.22 per year; p < 0.001), and the standardised morbidity rate decreased from 210.0 to 161.6 (-4.62 per year; p < 0.001). Declines were observed in both men and women. For clinical context, crude observed smoking increased by 10.2%, known-status smoking increased by 16.8%, annual AMI attacks declined by 15.4%, and standardised AMI morbidity declined by 23.0% between the first and final study years.

In TKU50, current smoking in the general population aged 16–64 years decreased from 31.2% in 2014 to 19.9% in 2024, and daily smoking decreased from 23.3% to 13.3%. In contrast, among hospitalised AMI patients aged approximately 15–64 years, observed smoking remained high and did not show a comparable decline, changing from 56.5% in 2015 to 56.9% in 2024. These datasets are not directly comparable because TKU50 is a general-population survey and AMI04 captures documented smoking status among hospitalised AMI patients. These smoking patterns are summarised in Fig. 1

Fig. 1.

Smoking trends among hospitalised AMI patients and population context, 2014/2015–2024. Smoking prevalence among hospitalised AMI patients in Estonia is shown using crude observed AMI04 values, known-status AMI04 values, and observed AMI04 values among patients aged approximately 15–64 years. General-population current and daily smoking among adults aged 16–64 years are shown for contextual comparison using TKU50. Crude observed AMI04 values represent the documented yes-category percentage among all hospitalised AMI patients, whereas known-status values exclude unknown smoking status. The approximately 15–64-year AMI subgroup is shown to provide age-range context for comparison with TKU50 adults aged 16–64 years. Population survey data and AMI registry data are not directly comparable because they differ in population structure, ascertainment, and measurement

In TKU40, overweight or obesity among adults aged 16–64 years increased slightly from 50.9% in 2014 to 52.1% in 2024, without a statistically significant trend. Among hospitalised AMI patients aged approximately 15–64 years, observed overweight remained higher, changing from 71.0% in 2015 to 74.2% in 2024, also without a significant trend. ETU30 provided limited survey-year context for diabetes, hypertension/high blood pressure, and high cholesterol and was used descriptively only. General-population smoking, BMI, and limited cardiometabolic context values are provided in Supplementary Table S6.

Sex-stratified temporal trends

In sex-stratified crude observed analyses, smoking increased significantly among women (β = 0.33% points/year; p = 0.008), while no significant full-period crude smoking trend was observed among men. Dyslipidaemia also increased among women (β = 0.88% points/year; p = 0.024), but this finding should be interpreted cautiously because unknown dyslipidaemia status decreased substantially over time. In men, diabetes showed a borderline positive trend (β = 0.17% points/year; p = 0.075), while hypertension, smoking, dyslipidaemia, and overweight did not show significant full-period crude observed trends. Sex-specific estimates should be interpreted alongside the declining proportion of women among AMI attacks and age-distribution shifts observed in AMI02.

Exploratory joinpoint analysis

Exploratory joinpoint analyses were rerun with a maximum of one joinpoint. Overall crude observed smoking did not select a joinpoint in the exploratory analysis; the full-period linear smoking trend remained positive. Some smoking known-status and standardised smoking models suggested post-2018 or post-2019 increases, but these analyses were exploratory.

Several terminal 2022–2024 segments, including dyslipidaemia and selected subgroup analyses, were based on only three observations and should be interpreted cautiously. These joinpoint findings are best regarded as hypothesis-generating rather than confirmatory evidence of abrupt epidemiological change. Exploratory joinpoint model results, including short-segment robustness flags, are provided in Supplementary Table S7.

Age-stratified temporal trends

Age-stratified crude observed analyses suggested heterogeneity across AMI age groups. Smoking increased significantly among patients aged 65–74 years (β = 0.75% points/year; p = 0.003) and 75–84 years (β = 0.54% points/year; p = 0.002). Hypertension increased among patients aged 75–84 years (β = 0.47% points/year; p = 0.005). Dyslipidaemia increased among patients aged ≥ 85 years (β = 1.31% points/year; p = 0.012), although dyslipidaemia subgroup findings should be interpreted cautiously because unknown dyslipidaemia status decreased over time. These age-stratified findings are exploratory because annual observations were limited and AMI case mix changed over time.

Discussion

Principal findings

In this national registry-based analysis, documented smoking among hospitalised AMI patients increased during 2015–2024 and persisted in known-status and age-sex-standardised analyses. Hypertension and dyslipidaemia remained highly prevalent, while diabetes and overweight showed relative temporal stability. Subgroup analyses suggested heterogeneity by sex and age, including increased smoking among women and older AMI patients, although these findings should be interpreted in the context of aggregate data, changing AMI case mix, and unknown-status trends.

The smoking finding was partly sensitive to COVID-period handling. When 2020 and 2021 were excluded, the crude observed smoking trend became borderline, whereas known-status smoking and age-sex-standardised known-status smoking remained significant. This suggests that the smoking signal was not fully explained by unknown status or age-sex composition, although aggregate annual data cannot establish causality.

Interpretation in the context of prior literature

Temporal trends in cardiovascular risk factor burden across Europe have been heterogeneous, reflecting differences in tobacco control policies, primary prevention strategies, socioeconomic transitions, and healthcare system performance [12–14]. Western European countries have generally reported sustained declines in smoking prevalence and improved lipid management, accompanied by reductions in cardiovascular mortality. In contrast, Eastern European and Baltic countries have historically experienced slower convergence toward Western prevention benchmarks [15, 16].

Our findings describe risk-factor patterns in a selected hospitalised AMI cohort, not the general population. National AMI burden declined significantly during the same period, with declines in hospitalised persons with AMI and in crude and standardised morbidity rates. Therefore, the AMI04 risk-factor trends occurred against a background of falling hospitalised AMI burden and should not be interpreted as evidence that national prevention policies succeeded or failed.

The contrast with population smoking trends is important. TKU50 showed that current and daily smoking declined substantially among adults aged 16–64 years in the general population from 2014 to 2024. By contrast, smoking among hospitalised AMI patients aged approximately 15–64 years did not show a comparable decline. This contextual contrast may indicate that patients presenting with AMI increasingly represent a subgroup with persistent or concentrated smoking-related risk despite broader population-level improvements. However, the comparison is descriptive only because TKU50 and AMI04 differ in population, age structure, measurement, and ascertainment.

During the study period, Estonia strengthened its tobacco-control framework. Amendments to the Tobacco Act implemented in 2019 prohibited the display of tobacco products and related products at points of sale, restricted remote sales, and extended several tobacco-control provisions to electronic cigarettes. The regulatory framework also prohibited flavours and fragrances other than tobacco in electronic-cigarette liquids [17–19]. These measures occurred alongside substantial declines in current and daily smoking in the Estonian general population. However, the present aggregate registry analysis was not designed to evaluate the implementation or causal effectiveness of individual tobacco-control policies. The absence of a comparable decline among hospitalised patients with AMI may reflect the concentration of smoking-related risk within a selected high-risk cohort, as well as differences in age structure, cumulative exposure, documentation, and case mix.

Dyslipidaemia should be interpreted cautiously. Although observed dyslipidaemia increased numerically, the full-period crude and age-sex-standardised trends were borderline. Unknown dyslipidaemia status decreased significantly, which may reflect improved documentation, testing, or registry completeness rather than a true increase in prevalence alone. COVID-excluded and exploratory joinpoint analyses suggested stronger increases, including in selected subgroups, but these should be treated as sensitivity or hypothesis-generating findings rather than confirmatory evidence.

The persistence of high hypertension and dyslipidaemia prevalence among AMI patients is consistent with contemporary European ACS cohorts showing that traditional cardiometabolic risk factors remain common despite advances in prevention and treatment [20, 21]. Smoking remains one of the most important modifiable drivers of myocardial infarction risk [22].

Persistent cardiometabolic risk among AMI patients may reflect concentration of residual risk among individuals who experience AMI despite broader prevention efforts. Potential contributors include smoking exposure, lipid-related residual risk, inflammation, lipoprotein(a), diabetes control, treatment intensity, socioeconomic patterning, and health care access [23–27]. These mechanisms were not measured in the aggregate registry exports and should therefore be discussed as plausible explanations rather than analysed pathways.

Demographic Heterogeneity and Sex-Specific Patterns

Sex-specific analyses should be interpreted in light of changing AMI case mix. The proportion of women among AMI attacks declined over time, and age distribution shifted. These changes could influence crude risk-factor estimates because risk-factor prevalence varies by sex and age. Nevertheless, the smoking trend persisted after age-sex standardisation, supporting continued sex-aware surveillance, particularly given prior evidence that cardiovascular prevention and outcomes differ by sex and that smoking may confer substantial cardiovascular risk in women [28–31]. STEMI/NSTEMI proportions did not change significantly, suggesting that broad AMI subtype redistribution is unlikely to fully explain the smoking trend.

Clinical and public health implications

The findings support continued attention to systematic smoking assessment and cessation support among patients hospitalised with AMI, particularly because the smoking increase remained evident in known-status and age-sex-standardised analyses. They also highlight the importance of transparent registry documentation, including unknown-status reporting, especially for dyslipidaemia. Targeted secondary-prevention strategies after AMI, including risk-factor reassessment, lipid management, medication adherence support, and smoking cessation, remain clinically important [32–35]. However, these aggregate data do not allow direct evaluation of tobacco policy effectiveness or identification of the individual-level causes of persistent smoking among AMI patients.

Strengths and limitations

This study has several strengths. It used national registry data over a continuous ten-year period and incorporated AMI denominators, case mix, national AMI burden, unknown-status trends, known-status sensitivity analyses, age-sex standardisation, COVID-period sensitivity, and contextual population survey data.

Several limitations should be acknowledged. First, the study used aggregate annual registry and survey data, which prevented individual-level adjustment, causal inference, and direct evaluation of within-person changes [36, 37]. The primary AMI04 cohort consisted of hospitalised AMI patients, not the general Estonian population; therefore, the risk-factor estimates describe documented risk factors among patients admitted with AMI and cannot be interpreted as population prevalence or as direct evidence of national prevention-policy success or failure.

Second, risk-factor documentation changed over time. Unknown status increased for smoking and overweight and decreased for dyslipidaemia and hypertension. These changes may reflect documentation practices, diagnostic testing, registry completeness, or case ascertainment, in addition to true changes in patient risk-factor profiles [38]. Known-status analyses partly addressed this issue but cannot fully remove documentation bias. This is particularly relevant for dyslipidaemia, because aggregate AMI04 data do not distinguish whether changes in recorded dyslipidaemia reflect lipid levels, treatment, testing intensity, or documentation completeness.

Third, the COVID-19 period may have affected hospital admission patterns, case ascertainment, documentation, and health care use [39]. Excluding 2020 and 2021 attenuated the crude observed smoking trend, although known-status smoking remained significant. AMI case mix also changed during the study period, including declines in total AMI attacks, a lower proportion of women, and shifts in age distribution. Although age-sex standardisation addressed some compositional change, residual differences in AMI phenotype, severity, comorbidity, treatment pathways, and survival to hospitalisation may remain.

Fourth, the registry includes hospitalised AMI patients and may be affected by survivor bias and AMI phenotype selection. Patients who died before hospitalisation, were not admitted, or were managed outside the captured hospitalised AMI pathway were not represented. Individual-level data on treatment intensity, medication adherence, socioeconomic position, ethnicity, education, income, region, smoking intensity, smoking duration, and health care access were also unavailable. Diagnostic checks identified possible residual autocorrelation in the hypertension model, although the short annual time series limited formal diagnostic interpretation.

Finally, exploratory joinpoint analyses were constrained by only ten annual observations. Although analyses allowed a maximum of one joinpoint and flagged segments with three or fewer observations, late 2022–2024 segments remain statistically fragile and should not be interpreted as confirmatory evidence of abrupt trend changes [40]. Because multiple subgroup and sensitivity analyses were conducted, statistically significant subgroup findings should be interpreted as hypothesis-generating rather than confirmatory [41]. The registry exports also did not include inflammatory biomarkers, lipoprotein(a), detailed lipid fractions, treatment targets, or other residual cardiovascular risk markers; these factors may be relevant to AMI risk and secondary prevention but could only be discussed conceptually [23–27].

Conclusion

In this national aggregate analysis of hospitalised AMI patients in Estonia, documented smoking increased during 2015–2024 and remained consistent in known-status and age-sex-standardised sensitivity analyses, although the crude observed trend was attenuated after excluding 2020–2021. Dyslipidaemia increased descriptively but requires cautious interpretation because unknown status changed substantially and full-period standardised trends were borderline. These trends occurred alongside declining national hospitalised AMI burden, fewer annual AMI attacks, a declining proportion of women among AMI attacks, and shifts in age distribution. Population context showed declining general-population smoking, whereas smoking among hospitalised AMI patients aged approximately 15–64 years did not show a comparable decline. Overall, the findings describe changing risk-factor profiles within the hospitalised AMI population and should not be interpreted as direct evidence of national prevention-policy success or failure.

Authors' contributions

AA led the conceptualization, investigation, and methodology, drafted the original manuscript, and contributed to formal analysis, visualization and review and editing of the manuscript. MD led the formal analysis and contributed to data curation, visualization, writing of the original manuscript, and review and editing of the manuscript. DV and SA contributed to investigation, validation, project administration, and review and editing of the manuscript. MN contributed to supervision, validation, and review and editing of the manuscript. All authors read and approved the final manuscript.

Funding

No financial support was received for this study.

Data availability

The datasets supporting the conclusions of this article are publicly available in the Estonian Health Statistics and Health Research Database (National Institute for Health Development, Estonia). The analysis used AMI01, AMI02, AMI04, TKU50, TKU40, and ETU30 tables. Available at: https://statistika.tai.ee/pxweb/en/Andmebaas/. Accessed 3 June 2026.

Declarations

Ethics approval and consent to participate

This study utilised publicly available, anonymised, aggregated data obtained from the Estonian Health Statistics and Health Research Database (National Institute for Health Development, Estonia), including AMI01, AMI02, AMI04, TKU50, TKU40, and ETU30 tables. No individual-level patient identifiers were available to the investigators. In accordance with national regulations governing the use of publicly accessible aggregated health statistics data, formal ethics committee approval and individual informed consent to participate were not required. All methods were carried out in accordance with the Declaration of Helsinki.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.