Impact of ACEI/ARB use on COVID-19 mortality in patients with ischaemic heart disease: insights from South Korean National health insurance service data

Abstract

Background

The COVID-19 pandemic has highlighted the importance of understanding the complex interactions between pre-existing conditions, such as ischaemic heart disease (IHD), and mortality risk. This study aims to investigate the associations among COVID-19 mortality, the presence of IHD, and the use of angiotensin receptor blockers (ARBs) and angiotensin-converting-enzyme inhibitors (ACEIs) by leveraging the South Korean National Health Insurance Service (NHIS) big data.

Methods

We conducted a retrospective cohort study of 30,056 patients with COVID-19 from the NHIS. The study population was stratified into four groups based on IHD status and the use of ACEI/ARBs. To assess their effects on mortality, we applied inverse probability of treatment weighting (IPTW) to balance baseline covariates, followed by a weighted logistic regression. The interaction between IHD and ACEI/ARB use was evaluated on both multiplicative (interaction term) and additive scales (Relative Excess Risk Due to Interaction, RERI), supplemented by an age-stratified analysis (≤ 70 vs. >70 years).

Results

Among 30,056 COVID-19 patients, we identified a significant interaction between IHD and ACEI/ARB use on mortality after adjusting for confounders. The main finding revealed that among patients with IHD, ACEI/ARB treatment was associated with 74% lower odds of mortality compared to non-treatment (Adjusted OR: 0.26, 95% CI: 0.08–0.92, p = 0.037). This protective interaction was strongly age-dependent, remaining statistically significant only in patients older than 70 (Interaction OR: 0.53, 95% CI: 0.27–0.99; p = 0.048). Furthermore, a significant antagonistic interaction on the additive scale (RERI: − 1.77, 95% CI − 3.04 to − 0.51; p = 0.006) confirmed a combined protective effect greater than the sum of the individual effects.

Conclusions

This study leverages the NHIS big data to provide a comprehensive analysis of the complex interactions among IHD, ACEI/ARB use, demographic factors, and comorbidities in the context of COVID-19 mortality. Our findings suggest a 45% reduction in the odds of mortality for IHD patients taking ACEI/ARB, particularly those over 70 years, but randomised controlled trials are needed before clinical guideline changes. The findings highlight the protective effect of ACEI/ARB in patients with IHD and emphasize the importance of personalized treatment approaches considering patient characteristics, pre-existing conditions, and medication use.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12879-025-11885-4.

Background

The COVID-19 pandemic has unveiled the intricacies of managing patients with pre-existing conditions, particularly ischaemic heart disease (IHD), which has been recognised as a significant risk factor for severe outcomes and increased mortality in those infected by SARS-CoV-2 [1, 2]. COVID-19 patients with cardiovascular disease demonstrate significantly elevated mortality rates, with ischemic heart disease associated with 2–3 fold increased risk of death compared to the general population (mortality rate: 15–20% vs. 2–5% in general population). This recognition has propelled a wave of research into potential therapeutic avenues, with a spotlight on angiotensin receptor blockers (ARBs) and angiotensin-converting enzyme inhibitors (ACEIs) for their potential to mitigate these adverse effects [3].

The literature to date presents a multifaceted view. Studies have explored how ARBs and ACEIs might affect the risk and severity of COVID-19, with some suggesting no significant increase in risk for those on these medications [3, 4]. Interestingly, evidence points to a potential protective effect of inpatient use of these drugs in reducing mortality rates among COVID-19 patients with hypertension, highlighting the nuanced interplay between cardiovascular disease management and COVID-19 outcomes [5, 6]. Recent systematic reviews have confirmed these protective associations across diverse populations, with effect sizes ranging from OR 0.7–0.9 for mortality reduction [7].

This elevated risk, combined with theoretical concerns about ACE2 receptor upregulation potentially facilitating viral entry, created clinical uncertainty about continuing these medications in COVID-19 patients with IHD. Despite the varied outcomes reported in studies on ACEIs and ARBs, these medications remain crucial components of the standard treatment regimen for IHD, which also includes aspirin, statins, and beta-blockers [6]. This established practice aligns with ongoing efforts to optimise care for patients with cardiovascular diseases in the context of the pandemic [7–10].

Building upon these insights, our study leverages the South Korean National Health Insurance Service (NHIS) database of 30,056 COVID-19 patients to elucidate the complex relationships among IHD, ACEI/ARB use, and COVID-19 mortality. This large-scale, population-based approach enables comprehensive analysis of interactions that were challenging to address in smaller studies, while addressing knowledge gaps regarding how demographic factors (age, sex, income) modulate these associations [11–14].

Our study objectives are to:

(1) investigate the association between ACEI/ARB use and COVID-19 mortality stratified by IHD status; (2) examine multiplicative and additive interaction effects between IHD and ACEI/ARB use; (3) assess age-dependent variations in the protective effects of ACEI/ARB, particularly focusing on whether benefits are confined to older patients; (4) evaluate sex- and hypertension-stratified interactions (as illustrated in Fig. 2), thereby providing evidence for more tailored treatment strategies; and (5) expand the discussion of potential biological mechanisms that may explain the observed protective associations, including anti-inflammatory, ACE2-modulating, and cardioprotective pathways.

Fig. 1.

Odds ratio plot: Factors influencing COVID-19 mortality. Forest plot showing the impact of demographic factors, comorbidities, and treatments on COVID-19 mortality. Odds ratios > 1 indicate increased risk, < 1 suggest protective effect. Key findings: higher risk with age and certain comorbidities; protective effect of ACEI/ARB in IHD patients

By addressing these objectives, we aim to contribute to the development of evidence-based treatment strategies for COVID-19 patients with cardiovascular comorbidities, particularly those with IHD.

Methods

Study design and data collection

In this retrospective cohort study, we leveraged the extensive population-based dataset from the NHIS to investigate the complex interactions among IHD, the use of ARBs or ACEIs, and COVID-19 mortality. To address the inherent limitation of lacking a randomized control group in retrospective studies, we applied inverse probability of treatment weighting (IPTW). IPTW was used to construct a weighted pseudo-population in which baseline covariates were balanced across the four study groups, thereby mimicking randomized allocation. This approach enhances comparability among groups and strengthens the causal interpretability of our results. Specifically, our study utilized a customized 10-million-person cohort, randomly sampled by the NHIS from a pool of approximately 20 million individuals who were diagnosed in 2020 with at least one of nine pre-specified conditions (COVID-19, diabetes, hypertension, asthma, cancer, cardiovascular disease, COPD, chronic kidney disease, or mental illness). The NHIS database contains comprehensive health records of nearly the entire South Korean population, providing a unique opportunity to conduct large-scale epidemiological studies. Previous validation studies of the NHIS database have demonstrated high accuracy for prescription data capture (>95% for chronic medications) and cardiovascular diagnoses (85–90% accuracy), making it suitable for pharmacoepidemiological research [14–16]. Recent studies have specifically validated ACEI/ARB prescription accuracy in Korean administrative databases, showing 94% concordance with clinical records.

Study population and follow-up

From the NHIS database, approximately 20 million individuals with either COVID-19 or at least one of 14 comorbidities were initially identified. The NHIS then generated a customized 10-million-person cohort through random sampling from this pool. From this sampled cohort, we identified patients with COVID-19 and subsequently applied predefined inclusion and exclusion criteria, resulting in a final study population of 30,056 patients.

Inclusion criteria: (1) Confirmed COVID-19 diagnosis (ICD-10: U07.1) between January 1-December 31, 2020; (2) Age ≥ 18 years; (3) Continuous NHIS enrollment ≥ 12 months before diagnosis; (4) Complete demographic and clinical data.

Exclusion criteria: (1) Previous COVID-19 diagnosis; (2) Death within 7 days of COVID-19 diagnosis; (3) Patients started on ACEI/ARB after COVID-19 diagnosis.

Follow-up period was from COVID-19 diagnosis until death or December 31, 2020 (median follow-up: 180 days, IQR: 120–240 days). Mortality was ascertained through linkage with national death certificates, with all-cause mortality as the primary outcome due to unavailability of cause-specific death data.

The study population was stratified into four groups based on the presence of IHD and the use of ACEI/ARB: (A) individuals without IHD not treated with ACEI/ARB; (B) those with IHD not receiving ACEI/ARB; (C) individuals without IHD but receiving ACEI/ARB; and (D) those with IHD who were also on ACEI/ARB treatment.

Variable definitions

IHD was identified using International Classification of Diseases, 10th revision (ICD-10) diagnostic codes I20-I25, requiring either: (1) ≥ 2 outpatient visits with IHD diagnostic codes, or (2) ≥ 1 hospitalization with IHD as primary diagnosis. Our definition did not include cardiac procedure codes (such as percutaneous coronary intervention [PCI], coronary artery bypass grafting [CABG], or coronary angiography), which may have resulted in underascertainment of IHD patients who underwent procedures without corresponding diagnostic codes. This represents a methodological limitation that could affect our effect estimates.

ACEI/ARB use was defined as prescription records for ≥ 30 consecutive days within 90 days preceding COVID-19 diagnosis, where consecutive was defined as allowing gaps of ≤ 7 days between prescriptions to account for typical refill patterns. Medications included: ACEIs (enalapril, lisinopril, ramipril, perindopril, captopril, quinapril, fosinopril) and ARBs (losartan, valsartan, candesartan, irbesartan, telmisartan, olmesartan, fimasartan). Combination medications containing ACEI/ARB were included. This timeframe was selected to capture patients on stable, ongoing therapy while excluding sporadic or discontinued use.

Age was calculated as of COVID-19 diagnosis date in 2020. Income level was determined based on 2020 NHIS insurance premium quartiles, which correlate with household income in Korea’s universal healthcare system.

In addition to the primary exposure variables (IHD and ACEI/ARB use), we collected data on demographic characteristics (age, sex, and income level), comorbidities (hypertension, diabetes, chronic obstructive pulmonary disease, chronic kidney disease, and malignancy), and clinical outcomes including mortality [15].

Our study population (30,056 patients) represents approximately 50% of Korea’s total confirmed COVID-19 cases during 2020. This difference reflects: (1) drawing the sample from a 10-million-person NHIS dataset; (2) our inclusion criteria requiring continuous enrollment and complete data; (3) exclusion of pediatric cases; and (4) administrative processing delays.

Statistical analysis

Statistical power analysis indicated 80% power to detect an odds ratio of 0.75 or smaller for the primary interaction effect with our sample size of 30,056 patients, assuming 5% mortality rate and 20% ACEI/ARB exposure in IHD patients.

We performed chi-square tests to evaluate the association between demographic characteristics (sex, income, and age) and mortality. To assess the association between underlying conditions and mortality, Fisher’s exact test was employed due to the low prevalence of certain conditions and mortality rates.

Variance inflation factor (VIF) analysis was performed to assess multicollinearity, with all VIF values in our final model confirmed to be < 3.0, indicating no problematic multicollinearity between related conditions such as IHD and hypertension.

To evaluate the association between IHD, ACEI/ARB use, and COVID-19 mortality, we applied inverse probability of treatment weighting (IPTW) to balance baseline covariates, followed by weighted logistic regression [17]. Interaction effects between IHD and ACEI/ARB use were assessed on both multiplicative and additive scales [18–20]. For readability, we simplified the statistical descriptions in the main text; detailed formulas and methodological explanations (e.g., RERI calculation and statistical test specifications) are provided in the Supplementary Methods.

For all statistical tests, a two-sided p-value < 0.05 was considered statistically significant. All confidence intervals are reported as 95% CI in the format (lower bound-upper bound). P-values < 0.001 are reported as “<0.001” rather than scientific notation for readability. All statistical analyses were conducted using SAS version 9.4 (SAS Institute Inc., Cary, NC, USA) and R version 4.0.3 (R Foundation for Statistical Computing, Vienna, Austria).

Results

Demographic influences on COVID-19 mortality rates

Our study analysed 30,056 COVID-19 patients comprising 28,696 survivors and 1,360 deceased individuals (overall mortality rate: 4.5%). Males represented 44.47% of survivors but had a higher mortality rate of 2.31% compared to females, who had a marginally lower mortality rate of 2.22% despite making up a larger proportion of survivors (51%) (Table 1).

Table 1.

Demographic characteristics and COVID-19 mortality: association between sex, income, age, and COVID-19 mortality rates among 30,056 patients

Variables	Category	Alive (n = 28,696)		Deaths (n = 1,360)		p-value
		n	%	n	%
Sex	Male	13,367	44.47	694	2.31	0.001
	Female	15,329	51	666	2.22
Income	High	8,943	29.75	457	1.52	< 0.001
	Middle-high	6,799	22.62	272	0.9
	Middle-low	5,622	18.71	175	0.58
	Low	7,332	24.39	456	1.52
Age	< 51	16,365	54.45	65	0.22	< 0.001
	51–60	5,499	18.3	108	0.36
	61–70	4,107	13.66	225	0.75
	71–80	1,988	6.61	398	1.32
	81+	737	2.45	564	1.88

Age-stratified mortality rates demonstrated a clear gradient: <51 years (0.22%), 51–60 years (0.36%), 61–70 years (0.75%), 71–80 years (1.32%), and 81 + years (1.88%), representing a > 8-fold increase from youngest to oldest groups (p < 0.001, Table 1).

Income level also played a crucial role, with mortality rates varying across different income groups: high income (1.52%), middle-high (0.9%), middle-low (0.58%), and low income (1.52%), highlighting income-related disparities in COVID-19 outcomes (p < 0.001) (Table 1). Further subgroup analyses were performed to investigate the potential differences in the effects of ACEI/ARB use on COVID-19 mortality across different age groups, revealing age-dependent variations in the protective effect.

Baseline characteristics of study groups

The baseline demographic and clinical characteristics of the four study groups, before and after statistical weighting, are presented in Supplementary Table 1. In the unweighted (original) analysis, there were significant differences across the four groups for most baseline characteristics. For example, significant imbalances were observed for age (p < 0.001), income (p < 0.001), and nearly all comorbidities, including Hypertension (p < 0.001) and Diabetes (p < 0.001). This confirmed the presence of significant confounding by indication.

After applying IPTW, these imbalances were substantially reduced. In the weighted population, the distributions of most covariates became comparable across the groups. For instance, the p-value for the difference in income distribution increased from < 0.001 to 0.25, and for a history of Schizophrenia, the p-value increased from < 0.001 to 0.198. While some variables like age and hypertension remained statistically different, the overall balance was markedly improved, allowing for a more robust subsequent analysis.

Role of pre-existing conditions

In Table 2, showed that ten comorbidities—diabetes, hypertension, asthma, chronic obstructive lung disease, chronic kidney disease, cancer, ischaemic heart disease, cor pulmonale, cerebrovascular disease, and severe mental disorders—were each associated with higher COVID-19 mortality.

Table 2.

Comorbidities and COVID-19 mortality: Prevalence of various pre-existing medical conditions and their association with COVID-19 mortality outcomes in 30,056 patients

Variables	Category	Alive (n = 28,696)		Deaths (n = 1,360)		p-value
		n	%	n	%
Diabetes	No	27,669	92.06%	1,168	3.89%	< 0.001
	Yes	1,027	3.42%	192	0.64%
Hypertension	No	27,139	90.29%	1,044	3.47%	< 0.001
	Yes	1,557	5.18%	316	1.05%
Asthma	No	28,464	94.70%	1,270	4.23%	< 0.001
	Yes	232	0.77%	90	0.30%
Chronic Obstructive Lung Disease	No	26,655	88.68%	1,095	3.64%	< 0.001
	Yes	2,041	6.79%	265	0.88%
Chronic Kidney Disease	No	28,608	95.18%	1,301	4.33%	< 0.001
	Yes	88	0.29%	59	0.20%
Cancer	No	28,530	94.92%	1,270	4.23%	< 0.001
	Yes	166	0.55%	90	0.30%
Ischaemic Heart Disease	No	28,239	93.95%	1,229	4.09%	< 0.001
	Yes	457	1.52%	131	0.44%
Cor pulmonale	No	28,147	93.65%	1,277	4.25%	< 0.001
	Yes	549	1.83%	83	0.28%
Cerebrovascular Disease	No	28,516	94.88%	1,261	4.20%	< 0.001
	Yes	180	0.60%	99	0.33%
Organic Mental Disorders	No	28,378	94.42%	1,147	3.82%	< 0.001
	Yes	318	1.06%	213	0.71%
Substance-induced Mental Disorders	No	28,641	95.29%	1,349	4.49%	< 0.001
	Yes	55	0.18%	11	0.04%
Schizophrenia	No	28,512	94.86%	1,332	4.43%	< 0.001
	Yes	184	0.61%	28	0.09%
Mood Disorders	No	27,926	92.91%	1,224	4.07%	< 0.001
	Yes	770	2.56%	136	0.45%
Neurotic Disorders	No	28,198	93.82%	1,303	4.34%	< 0.001
	Yes	498	1.66%	57	0.19%

The odds ratio (OR) plot provides a visual representation of the relative impact of various factors on COVID-19 mortality (Fig. 1). ORs exceed 1 for patients aged > 70, those in lower-income brackets, and individuals with key comorbidities—including diabetes, hypertension, chronic kidney disease, cancer, and IHD—underscoring each factor’s independent contribution to mortality.

Odds ratio from contingency table with IHD and ACEI/ARB

Of 30,056 patients, 27,377 (91.1%) were not receiving ACEI/ARB at baseline. Overall crude mortality among non-users was 4.02% (1100/27,377). Within the non-IHD stratum, crude mortality among non-users was 3.66% (986/26,904). When the data were stratified by IHD status, two contrasting patterns emerged (Table 3):

Table 3.

Unweighted (crude) contingency table and odds ratios for IHD, ACEI/ARB use, and COVID-19 mortality: analysis of the relationship between IHD, use of aceis or ARBs, and COVID-19 mortality

		Death	Alive	Odds ratio (95% CI)	p-value
IHD	ACEI/ARB = 1	17	98	0.55 (0.31–0.95)	0.033
	ACEI/ARB = 0	114	359
No IHD	ACEI/ARB = 1	243	2321	2.75 (2.38–3.19)	< 0.001
	ACEI/ARB = 0	986	25,918
Overall	ACEI/ARB = 1	260	2419	2.57 (2.23–2.96)	< 0.001
	ACEI/ARB = 0	1100	26,277

IHD Ischaemic Heart Disease, ARBs Angiotensin Receptor Blockers, ACEIs Angiotensin-Converting Enzyme Inhibitors, CI Confidence Interval

• IHD present (benefit): among patients with IHD, ACEI/ARB use lowered mortality from 24.1% to 14.8%, corresponding to an OR of 0.55 (95% CI 0.31–0.95; p = 0.033).

• IHD absent (higher risk): among patients without IHD, ACEI/ARB users had 9.48% (243/2,564) mortality vs. 3.66% (986/26,904) in non-users (OR 2.75, 95% CI 2.38–3.19; p < 0.001.

These stratum-specific contrasts are visualised in Fig. 2A, where the red (no ACEI/ARB) and blue (ACEI/ARB) lines cross, highlighting that the direction of the ACEI/ARB association flips according to underlying IHD. As shown in Fig. 2B, the trend observed in men was not statistically significant (p = 0.472). In contrast, the crossing pattern is significantly steeper in patients with hypertension (Fig. 2C), a finding supported by our interaction test (p < 0.001).

Fig. 2.

Interaction effects of ACEI/ARB use and IHD status on COVID-19 mortality rate. Three panels showing: (A) Effect of ACEI/ARB on mortality in patients with/without IHD (B) Sex differences in ACEI/ARB and IHD interaction (C) Impact of hypertension on ACEI/ARB and IHD interaction Red lines: non-ACEI/ARB users; Blue dashed lines: ACEI/ARB users. Line convergence/divergence indicates interaction effects

Table 3 therefore highlights that ACEI/ARB use is associated with a clinically meaningful 45% reduction in the odds of mortality only when IHD is present, whereas the higher mortality observed in ACEI/ARB-treated patients without IHD likely reflects confounding by indication: these individuals more often carry other high-risk cardiovascular conditions (e.g., hypertension, heart failure) that prompted initiation of ACEI/ARB therapy. Accordingly, a dedicated interaction model that incorporates an ACEI/ARB × IHD product term—while simultaneously adjusting for other cardiovascular comorbidities that drive treatment selection—is essential.

Multiplicative interaction between IHD and ACEI/ARB

We first evaluated whether adding the ACEI/ARB × IHD term improved model fit in a multiple logistic-regression model that already contained both main effects and was adjusted for age, sex, income, and 12 additional comorbidities. Two different statistical tests confirmed that the interaction was statistically significant. The Wald test rejected the null hypothesis of no interaction (F = 4.40, p = 0.036), and this was supported by the Rao-Scott adjusted likelihood-ratio test (LR = 5.12, p = 0.025). Thus, including the interaction term significantly improves the model and indicates that ACEI/ARB therapy modifies the association between IHD and COVID-19 mortality.

Given the significant interaction between IHD status and ACEI/ARB use, we analyzed the stratum-specific odds ratios (ORs) to understand their combined effect on mortality (Table 4).

Table 4.

Odds ratios of interaction terms: comparison of COVID-19 mortality risk for patients with different combinations of IHD status and ACEI/ARB use, relative to the reference group and subgroups

Reference (IHD+, ACEI/ARBs+)vs.	Estimate	Std. Error	Odds ratio (95% CI)	p-value
Group (IHD -, ACEI/ARBs -)	−0.4314	0.6030	0.65 (0.20–2.11)	0.474
Group (IHD +, ACEI/ARBs -)	−1.3329	0.6373	0.26 (0.08–0.92)	0.037
Group (IHD -, ACEI/ARBs +)	−0.3912	0.6124	0.68 (0.20–2.25)	0.523

IHD Ischaemic Heart Disease, ARBs Angiotensin Receptor Blockers, ACEIs Angiotensin-Converting Enzyme Inhibitors, CI Confidence Interval

• Among patients with IHD, those treated with an ACEI/ARB had significantly lower odds of death compared to those not treated with an ACEI/ARB (OR = 0.26, 95% CI 0.08–0.92; p = 0.037). This suggests a strong protective association for ACEI/ARB therapy within this high-risk group.

• Among patients receiving ACEI/ARB therapy, there was no statistically significant difference in mortality between those with IHD and those without IHD (OR = 0.68, 95% CI 0.20–2.25; p = 0.523).

• Furthermore, the group with neither IHD nor ACEI/ARB use showed no significant difference in mortality when compared to the reference group of patients with both IHD and ACEI/ARB use (OR = 0.65, 95% CI 0.20–2.11; p = 0.474).

Collectively, these results demonstrate a notable protective association of ACEI/ARB therapy that is statistically evident specifically among patients with IHD. The therapy appears to reduce the mortality risk in IHD patients to a level that is statistically indistinguishable from that of patients without IHD.

Additive interaction RERI between IHD and ACEI/ARB

Additive-interaction analysis yielded a RERI of − 1.77 (95% CI − 3.04 to − 0.51; p = 0.006). A negative RERI indicates antagonism on the additive scale, meaning that the absolute risk of COVID-19 death in patients with both IHD and ACEI/ARB therapy was 1.77% points lower than the sum of the risks associated with IHD alone and ACEI/ARB use alone. Clinically, this antagonistic interaction implies that ACEI/ARB treatment confers an additional absolute mortality reduction in IHD patients, beyond what would be expected if the effects of IHD and ACEI/ARB were simply additive.

Age-Stratified interaction effect age as an effect modifier in the IHD and ACEI/ARB interaction

The seemingly paradoxical protective association of IHD + ACEI/ARB versus no exposure prompted a focused examination of age, since older patients are both more likely to have IHD and to receive ACEI/ARB (Table 4). In a sensitivity analysis, removing age from the fully adjusted multiplicative-interaction model rendered the ACEI/ARB × IHD coefficient nonsignificant (OR = 0.92, 95% CI 0.50–1.71; p = 0.79), whereas re‐including age restored a significant protective interaction (adjusted OR = 0.52, 95% CI 0.28–0.95; p = 0.048). We then stratified the cohort at age 70 years (Table 5).

Table 5.

Age-stratified analysis of interaction effects: Odds ratios for the interaction between IHD and ACEI/ARB use on COVID-19 mortality, stratified by age groups (>70 years and ≤ 70 years)

Age Group	Group	Death	Alive	Odds ratio (95% CI)	Logistic Regression
					Odds ratio (95% CI)	p-value
Older group (> 70)	IHD +, ACEI/ARBs +	14	49	0.82 (0.45–1.50)	0.53 (0.27–0.99)	0.048
	IHD -, ACEI/ARBs -	674	1943
Younger group (≤70)	IHD +, ACEI/ARBs +	3	49	4.70 (1.46–15.17)	1.44 (0.37–5.57)	0.594
	IHD -, ACEI/ARBs -	312	23,975

IHD Ischaemic Heart Disease, ARBs Angiotensin Receptor Blockers, ACEIs Angiotensin-Converting Enzyme Inhibitors, CI Confidence Interval

When we stratified the cohort at 70 years, the protective interaction persisted only in older patients: those > 70 years had an interaction OR of 0.53 (95% CI 0.27–0.99; p = 0.048), indicating a borderline protective effect of ACEI/ARB in the presence of IHD, while patients ≤ 70 years showed non-significant increase in the odds of death (OR 1.44, 95% CI 0.38–5.57; q = 0.594), which does not suggest protection. These results demonstrate that age modifies the ACEI/ARB × IHD interaction: the protective association emerges only when age is accounted for and is confined to the older stratum. Accurate evaluation of ACEI/ARB efficacy in IHD patients therefore requires treating age as both a covariate and an effect modifier.

Discussion

In summary, our research elucidates the multifaceted impact of demographic factors, medication use, pre-existing conditions, and interaction effects on COVID-19 mortality rates. These findings not only reinforce the importance of comprehensive patient management strategies but also underscore the need for a nuanced understanding of the variables influencing COVID-19 outcomes. The RERI analysis, in particular, provides valuable insights into the complex interplay between IHD, ACEI/ARB use, and COVID-19 mortality, suggesting potential avenues for targeted interventions and personalised treatment approaches.

The present study sheds light on the complex interplay among IHD, the use of ARBs or ACEIs, and COVID-19 mortality rates. Our findings underscore the significant impact of demographic factors, pre-existing medical conditions, and medication use on patient outcomes during the pandemic. A notable strength of our study is the utilisation of the NHIS big data, which covers nearly the entire South Korean population of over 50 million individuals. From this extensive database, we analysed data from 30,056 COVID-19 patients out of approximately 10 million accessible health records. This large-scale, population-based approach enhances the robustness and generalisability of our findings to the broader South Korean population. By leveraging the NHIS dataset, which covers approximately 10 million individuals including 30,056 COVID-19 patients, we comprehensively analysed factors influencing COVID-19 mortality rates. This large-scale, population-based approach allowed us to examine intricate interactions that were challenging to address in prior studies [21]. The use of NHIS data enhances the generalisability of our findings and minimises sample selection bias [14], providing crucial insights for evidence-based decision-making in COVID-19 patient management.

Furthermore, our study identified significant interaction effects between ACEI/ARB use and IHD status, highlighting the importance of understanding the complex interplay among these factors beyond their individual impacts. Notably, the modulating effects of sex and hypertension status on the interaction between ACEI/ARB use and IHD status provide valuable insights for personalized patient management. These findings underscore the need for tailored approaches in the care of COVID-19 patients and generate novel hypotheses for future research [22]. While our findings are consistent in some aspects with previous meta-analyses, such as that of Baral et al. [23], our study uniquely observed age-dependent effects. This difference may be attributed to our large cohort size of 30,056 COVID-19 patients and the comprehensive clinical information available in the NHIS database, which allowed for such nuanced analysis. The larger sample size increased our statistical power to detect age-dependent effects that might have been missed in smaller studies.

Several potential biological mechanisms could explain the protective effect of ACEI/ARB on COVID-19 mortality. First, these drugs may reduce inflammatory responses by inhibiting the action of angiotensin II. Given that COVID-19 severity is associated with excessive inflammatory reactions, this anti-inflammatory effect could be beneficial [1, 2]. Second, ACEI/ARB might increase ACE2 expression in the lungs, which could paradoxically be protective. While SARS-CoV-2 uses ACE2 for cell entry, increased ACE2 might also enhance the conversion of angiotensin II to angiotensin- [1–7], which has anti-inflammatory and anti-fibrotic properties [3, 5]. Third, these medications may help maintain cardiovascular homeostasis, which is crucial for patients with pre-existing conditions like IHD when facing the additional stress of COVID-19 infection. Fourth, ACEI/ARB may prevent COVID-19-associated cardiac complications including myocardial injury, arrhythmias, and acute heart failure exacerbation, which are major contributors to mortality in patients with pre-existing cardiovascular disease. This cardioprotective effect may be particularly important in IHD patients who are already at high risk for these complications. However, these mechanisms remain speculative and require further investigation [7, 8].

The analysis of demographic influences on COVID-19 mortality rates revealed notable disparities based on sex, income, and age. The observed higher COVID-19 mortality in males may be attributed to several factors: [1] higher ACE2 receptor expression in males; [2] differences in immune response with males showing more pronounced inflammatory reactions; [3] higher prevalence of cardiovascular risk factors and smoking in Korean males; and [4] potential differences in healthcare-seeking behavior [24]. Furthermore, the observed income-related disparities in mortality rates highlight the need for targeted interventions to address socioeconomic determinants of health in the context of the pandemic [25].

The increased mortality risk observed with comorbidities such as diabetes, hypertension, and chronic kidney disease reflects the complex interplay between COVID-19 and pre-existing organ dysfunction [26]. These conditions are associated with chronic inflammation, endothelial dysfunction, and compromised immune responses, creating a vulnerable substrate for severe COVID-19 outcomes [27].

Our study also emphasizes the critical role of pre-existing medical conditions in determining COVID-19 outcomes. The significantly increased mortality risk among patients with conditions such as diabetes, hypertension, asthma, chronic obstructive lung disease (COLD), chronic kidney disease (CKD), cancer, IHD, cor pulmonale, cerebrovascular disease, and mental health disorders underscores the importance of tailored management strategies for these vulnerable populations [28]. These findings are consistent with prior research indicating that comorbidities are associated with poorer prognoses and increased mortality in COVID-19 patients [29].

The quantitative analysis of the relative excess risk due to interaction (RERI) provides valuable insights into the synergistic effects of the studied variables on health outcomes. The significant negative RERI estimate suggests that targeted interventions leveraging these interaction effects could substantially reduce the risk of adverse outcomes in COVID-19 patients [23]. This finding underscores the need for personalized treatment approaches that take into account the complex interplay between patient characteristics, pre-existing conditions, and medication use. The significant interaction effects between ACEI/ARB use and IHD status further highlight the importance of considering medication use in the context of pre-existing conditions when managing COVID-19 patients [30].

Our findings have important clinical implications, suggesting that continuation of ACEI/ARB in patients with IHD during COVID-19 infection might be beneficial, particularly in patients over 70 years of age. However, before this can be implemented as a clinical recommendation, randomised controlled trials are necessary to confirm these observational findings. Clinicians should consider these findings when making treatment decisions for IHD patients with COVID-19, while also taking into account individual patient characteristics and potential contraindications. Our results are largely consistent with previous meta-analyses, such as the one by Baral et al. [23], which also found a protective effect of ACEI/ARB in COVID-19 patients. However, our study uniquely observed age-dependent effects, which may be attributed to our large cohort size and detailed clinical information. This highlights the importance of considering age as a modifying factor in the relationship between ACEI/ARB use and COVID-19 outcomes.

Several important limitations must be acknowledged. First, as an observational study, we cannot establish causal relationships. Although the absence of a randomized control group is an unavoidable limitation of our retrospective design, our use of IPTW and multivariable adjustment substantially improved comparability between groups. This weighting strategy allowed us to approximate randomized allocation and mitigate confounding, thereby enhancing the validity of our findings despite the non-randomized nature of the data. The associations observed require validation through experimental studies. Second, our IHD definition may have led to under-detection due to omission of procedure codes, and our 90-day ACEI/ARB window, though commonly used in Korean pharmacoepidemiology, may not capture all treatment patterns. Future studies incorporating procedure data and alternative exposure definitions are warranted. Third, while we used logistic regression due to its clear interpretation of interaction effects through odds ratios, it would be more informative to use Cox proportional hazards (PH) models. Note that Cox PH model has an advantage of accounting for differential follow-up times or provide hazard ratios over time. Future studies should employ time-to-event analyses to better understand the temporal dynamics of ACEI/ARB protection, particularly whether the protective effect varies during early versus late COVID-19 infection periods. Future studies should employ time-to-event analyses to better understand the temporal dynamics of ACEI/ARB protection, particularly whether the protective effect varies during early versus late COVID-19 infection periods. Fourth, the NHIS database lacks detailed medication dosage and adherence information, limiting our ability to assess dose-response relationships or medication compliance effects. Fifth, we cannot distinguish COVID-19-specific deaths from other causes, which limits mechanistic interpretation of our findings. Sixth, our 90-day medication exposure window was empirically chosen rather than literature-based due to lack of established standards in the field. Seventh, unmeasured confounders such as BMI and smoking status were unavailable and could influence both medication prescribing and COVID-19 outcomes. Eighth, we did not calculate composite comorbidity scores such as the Charlson Comorbidity Index, which could provide more comprehensive risk adjustment than individual comorbidity variables. While we adjusted for major comorbidities individually, a composite score might better capture overall disease burden and its interaction with ACEI/ARB effects. Finally, generalizability to other populations requires caution given our Korean-specific cohort.

Future research should focus on several key areas to further elucidate the relationship between ACEI/ARB use and COVID-19 outcomes in IHD patients. Randomised controlled trials are urgently needed to establish causal relationships and determine optimal management strategies. Mechanistic studies could help elucidate the exact biological pathways through which ACEI/ARB may confer protection against severe COVID-19. Additionally, investigations into optimal dosing and timing of ACEI/ARB administration in the context of COVID-19 infection, as well as exploration of potential differences in effects between various types of ARBs and ACEIs at different dosage levels, would provide valuable insights. Studies incorporating detailed medication adherence data, and dose-response relationships are particularly needed. Finally, studies in diverse populations are crucial to assess the generalizability of these findings across different ethnicities and healthcare systems.

Conclusion

This study leverages the comprehensive NHIS database to provide valuable insights into the complex interactions among IHD, ACEI/ARB use, demographic factors, and comorbidities in the context of COVID-19 mortality. Our findings suggest a clinically meaningful 45% reduction in the odds of mortality for IHD patients taking ACEI/ARB, particularly those over 70 years of age. The antagonistic interaction demonstrated by RERI analysis indicates biological synergy that warrants further investigation. These results contribute to the growing body of evidence guiding clinical decision-making during the ongoing pandemic and underscore the need for further research to elucidate the underlying mechanisms of the observed effects.

Abbreviations

ACEI

Angiotensin-Converting Enzyme Inhibitor

ARB

Angiotensin Receptor Blocker

COPD

Chronic Obstructive Pulmonary Disease

CKD

Chronic Kidney Disease

COVID

19-Coronavirus Disease 2019

FDR

False Discovery Rate

IHD

Ischaemic Heart Disease

ICD-10

International Classification of Diseases, 10th Revision

LRT

Likelihood Ratio Test

NHIS

National Health Insurance Service

OR

Odds Ratio

RERI

Relative Excess Risk Due to Interaction

SARS

CoV-2-Severe Acute Respiratory Syndrome Coronavirus 2

SAS

Statistical Analysis System

Authors’ contributions

TP and SWK: study conceptualization and design. TG and SN: data processing, statistical analysis, and visualization. BO and SL: statistical and clinical result review and interpretation. TG and SWK: manuscript preparation. TP: study supervision. All authors contributed to the article and approved the submitted version.

Funding

This research was supported by a research grant from the Ministry of Science and ICT, South Korea (No. 2021M3E5E3081425), and by the Mid-Career Bridging Program through Seoul National University. The research was supported by a grant (RS-2023 00227944) of the National Research Foundation, funded by the Ministry of Science, Technology, and Telecommunication of the Government of the Republic of Korea.

Data availability

The data supporting the findings of this study are available from the South Korean National Health Insurance System (NHIS); however, access to these data is restricted due to licensing agreements and is not publicly available. Nevertheless, the data can be obtained from the authors upon reasonable request and with permission from the NHIS (https://nhiss.nhis.or.kr/).

Declarations

Ethics approval and consent to participate

This study was approved by the Seoul National University Institutional Review Board (IRB No. E2112/001–003). The IRB determined that the requirement for informed consent was waived because the analyses were conducted retrospectively using anonymized data derived from the South Korean National Health Insurance Service (NHIS) database. The use of NHIS data was additionally approved by the Health Insurance Review and Assessment Service (NHIS-2022-1-549). All procedures of this study were conducted in accordance with the principles of the Declaration of Helsinki.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.