Risk of Severe Hypoglycemia Associated With Concomitant Use of Sulfonylureas and Amiodarone Authors: Sungho Bea, Hyesung Lee, Young Min Cho, You‐Jung Choi, Ju‐Young Shin, Katsiaryna Bykov Published: 2026 ABSTRACT ABSTRACTAimsTo evaluate whether initiation of amiodarone while receiving sulfonylurea is associated with an increased risk of severe hypoglycemia compared with the initiation of other antiarrhythmic agents (flecainide, sotalol, or propafenone) while receiving sulfonylurea.Materials and MethodsWe conducted a population‐based cohort study using South Korean National Health Insurance Service claims data from 2010 to 2022. Adults aged 18 years or older with type 2 diabetes who initiated oral antiarrhythmic therapy (amiodarone or comparator agents) while on sulfonylurea treatment between 1 January 2011 and 31 December 2022, were included. The first dispensing date of an antiarrhythmic agent while on sulfonylureas was the index date. Patients with prior use of first‐generation sulfonylureas, meglitinides, insulin, or antiarrhythmics, or those initiating multiple antiarrhythmic drugs on the index date were excluded. The primary outcome was severe hypoglycemia, defined as hospital admission or emergency department visits with a primary diagnosis of hypoglycemia. Propensity score overlap weighting and weighted Cox models were applied.ResultsAmong 12 730 eligible patients, 5325 (41.8%) initiated amiodarone and 7405 (58.2%) initiated comparator antiarrhythmics. Mean age was higher in the amiodarone group (69.2 vs. 67.6 years). After weighting, amiodarone initiation was associated with a higher risk of severe hypoglycemia compared with initiation of comparator antiarrhythmics (weighted HR 1.92; 95% CI, 1.12–3.30). The increased risk persisted across multiple sensitivity and subgroup analyses.ConclusionsInitiation of amiodarone in patients receiving sulfonylureas was associated with a higher risk of hypoglycemia compared with other antiarrhythmics. Given the continued use of sulfonylureas in diabetes care, this finding highlights the need for caution and close monitoring during their concomitant use with amiodarone. Introduction Sulfonylureas are widely utilised medications in the management of type 2 diabetes due to their robust efficacy in achieving glycemic control [1, 2], established microvascular benefits [3, 4] and cost‐effectiveness [5]. However, the metabolism and elimination of sulfonylureas predominantly occur in the liver and are significantly affected by potent inhibitors of liver cytochrome P450 2C9 (CYP2C9) [6], such as amiodarone, an antiarrhythmic drug frequently prescribed for arrhythmia [7]. Despite the remarkable therapeutic effects of amiodarone in rhythm control, its inhibitory effect on CYP2C9 should be recognised, as it may precipitate CYP2C9‐mediated pharmacokinetic drug–drug interactions [8]. Type 2 diabetes is independently associated with arrhythmogenic substrates such as myocardial fibrosis and autonomic dysfunction [9], making concomitant use of sulfonylureas and amiodarone is common in clinical practice. Hypoglycemia is a serious adverse effect of sulfonylureas and has been implicated in the development of cardiac arrhythmias through QT interval prolongation, raising particular concern in patients with type 2 diabetes and preexisting arrhythmias [10, 11]. The inhibition of CYP2C9 by amiodarone further complicates this clinical context, potentially increasing the plasma concentration of sulfonylureas and, subsequently, the risk of hypoglycemia. Two observational studies have been conducted to assess the risk of hypoglycemic events in patients receiving both sulfonylurea and amiodarone [12, 13]. However, the first study adopted a case–control design, limiting the evaluation to relative measures of effect and requiring consideration of potential bias due to confounding by indication in the absence of an active comparator [12]. The second study evaluated CYP2C9 inhibitors as a group, among which amiodarone accounted for only 6.9% of CYP2C9 inhibitors [13]. Therefore, there is still a lack of robust evidence on the safety of the concomitant use of these two medications. Thus, this nationwide cohort study aimed to evaluate the risk of severe hypoglycemia among patients with type 2 diabetes receiving sulfonylureas who initiated amiodarone versus other antiarrhythmic agents, including flecainide, sotalol and propafenone. Methods Study Design and Data Source An active‐comparator, new‐user cohort study was conducted to emulate a hypothetical target trial assessing the risk of severe hypoglycemia associated with amiodarone use among patients with type 2 diabetes receiving sulfonylureas [14]. The key design components and details of the target trial and its emulation are presented in Table S1. Nationwide claims data were obtained from the National Health Information Database (NHID) maintained by the National Health Insurance Service, covering the period from 2011 to 2022 (Figure S1). The NHID covers the entire population of South Korea (> 50 million) under a single‐payer system providing universal health insurance coverage. This database comprises comprehensive information on demographics, including socioeconomic status and healthcare utilisation (e.g., medical procedures, diagnostic codes and medications) from both inpatient and outpatient settings [15]. Study Population The study population included individuals aged 18 years or older with type 2 diabetes who were receiving sulfonylureas between 1 January 2011 and 31 December 2022. Among them, we identified patients who initiated oral antiarrhythmic medications, including amiodarone, flecainide, sotalol, or propafenone, in the outpatient setting during sulfonylurea treatment. The index date was defined as the date of initiation of antiarrhythmic medications during sulfonylureas use. Patients were excluded if they had received any antiarrhythmic medications, either oral or injectable, during the 1‐year period prior to the index date, as the study focused on identifying new users of antiarrhythmic medications. To ensure the accuracy of covariate adjustments, individuals with incomplete baseline information, such as missing data on age or sex, were excluded (n = 3). We also excluded patients who had used first‐generation sulfonylureas or meglitinides during the year before the index date, given their distinct pharmacokinetic properties and risk profiles compared to second‐generation sulfonylureas. Similarly, individuals with insulin use in the year prior to the index date were excluded to reduce heterogeneity related to diabetes severity and treatment intensity. Finally, we excluded patients who were prescribed more than one antiarrhythmic medications on the index date to avoid ambiguous exposure‐group assignment. Treatment Strategies Treatment strategies were defined by classifying study participants into two antiarrhythmic treatment groups based on whether they initiated amiodarone or other antiarrhythmics (flecainide, sotalol, or propafenone) while receiving sulfonylureas. Outcome and Follow‐Up Study outcome was severe hypoglycemia, which was defined as hospitalisation with hypoglycemia. We identified severe hypoglycemia events using hospital admission and emergency department records with a primary diagnosis coded for hypoglycemia, based on the International Statistical Classification of Diseases and Related Health Problems, 10th Revision (ICD‐10) (Table S2). This approach aligns with previous epidemiological studies using NHIS data to capture severe hypoglycemia events [16, 17]. Exposure was defined using an observational analog of the per‐protocol effect (as‐treated approach). Individuals were followed from the initiation of antiarrhythmic drugs until the earliest event, which included discontinuation of either sulfonylureas or antiarrhythmic drugs, switching to an alternative antiarrhythmic agent, death, or the end of the study period (31 December 2022) (Figure S1). Discontinuation for both medications, an antiarrhythmic or sulfonylurea, was the absence of a subsequent prescription within 30 days after the end of the previous prescription's supply period. Covariates We considered a broad range of potential confounders including demographic and socioeconomic characteristics (e.g., age, sex and income level), healthcare utilisation (e.g., hospitalisation, physician visits and physician specialty), comorbidities (e.g., dyslipidemia, hypertension and atrial fibrillation), previous medication use (e.g., anticoagulants, antiplatelets and antihypertensives), antidiabetic medications, diabetic complications (e.g., retinopathy, neuropathy and nephropathy), and clinical prognostic scoring systems such as the Charlson Comorbidity Index (CCI) [18]. Demographic and socioeconomic variables were measured on the index date, and the other covariates were measured during the year prior to the index date. Details of the covariates are presented in Table S3. Statistical Analyses We summarised the distribution of covariates as means with standard deviations for continuous variables and frequencies with proportions for categorical variables. We calculated the propensity score (PS) using a multivariable logistic regression model that included all measured covariates. To account for potential confounding, we used PS‐based overlap weights [19]. Overlap weighting assigns weights to individuals based on the probability of receiving the opposite treatment, resulting in weights bounded between 0 and 1 thereby eliminating the need for truncation. This method achieves perfect balance of all covariates between two treatment groups [20]. The distribution of covariates between treatment groups (amiodarone vs. flecainide/sotalol/propafenone) was evaluated using the standardised mean difference, assuming good balance if the absolute value was less than 0.1. We calculated incidence rates (IRs) of severe hypoglycemia per 100 person‐years using weighted events and weighted person‐time with 95% confidence intervals (CIs) estimated from Poisson regression. For the primary analysis, we estimated hazard ratios (HRs) with 95% CIs using weighted Cox proportional hazards model. We also plotted cumulative incidence curves of severe hypoglycemia for the primary as‐treated analysis. We used SAS Enterprise Guide, version 7.1, for data management and analysis. Sensitivity and Subgroup Analyses We conducted various sensitivity analyses to assess the robustness of the study findings. First, we repeated the analysis after excluding individuals who had experienced hypoglycemia within the year prior to the index date, thereby ensuring that the anaylsis captured incident hypoglycemic outcomes after initiation of antiarrhythmic medications. Second, we conducted a competing risk analysis accounting for the risk of death using the Fine and Grey subdistribution hazards model. This approach was employed to provide more accurate risk estimates in the presence of death as a competing event that could preclude the occurrence of hypoglycemia [21]. Third, we broadened the definition of the outcome by including all hypoglycemia events regardless of care setting, such as inpatient admissions, emergency department visits and outpatient encounters. Fourth, we added the start date of insulin use as an additional censoring criterion to account for the increased risk of severe hypoglycemia associated with insulin therapy. Fifth, we applied an intention‐to‐treat approach as an alternative exposure definition, in which patients were followed for a fixed duration of 60 and 365 days from the index date regardless of changes in treatment status. Sixth, we repeated the analyses by varying the allowable gap between prescriptions when defining continuous exposure, using grace periods of 15 and 60 days. This sensitivity analysis assessed the robustness of our results to different assumptions regarding treatment continuity. Seventh, to assess the potential influence of concomitant medications affecting CYP2C9‐mediated metabolism, we conducted additional sensitivity analyses excluding patients who received potent CYP2C9 inducers, potent CYP2C9 inhibitors, or either class of medication (Table S4). We conducted several subgroup analyses to identify differences in the risk of severe hypoglycemia according to baseline characteristics. The following covariates were considered in the subgroup analyses: age group (< 65, ≥ 65 years), sex (male, female), presence of heart failure and type of sulfonylureas on the index date (glimepiride, gliclazide, glyburide). We recalculated the PS and overlap weights within each subgroup, as described in the primary analysis. Results Overall, this study included 12 730 patients who initiated antiarrhythmic drugs while on sulfonylureas (Figure 1). Among these, 5325 (41.8%) patients initiated amiodarone (mean age 69.2 years; 60.6% male) whereas 7405 (58.2%) patients initiated flecainide/sotalol/propafenone (mean age 67.6 years; 59.8% male). Among sulfonylurea types, most patients in both groups were treated with glimepiride, accounting for 97.2% of prescriptions in the amiodarone group and 98.1% in the comparator group. Patients who initiated amiodarone had greater healthcare utilisation and higher prevalence of heart failure, previous percutaneous coronary intervention and coronary artery disease, with higher use of P2Y12 inhibitors and diuretics, likely reflecting a greater underlying cardiovascular burden (Table 1). However, diabetes‐related characteristics such as duration of sulfonylurea use, presence of diabetic complications and use of other antidiabetic medications were generally comparable between the two groups. After applying overlap weighting, we observed well‐balanced distribution across all baseline covariates, with all standardised mean differences below 0.1 (Table 1 and Table S5). Study flow diagram. TABLE 1Selected baseline characteristics of the study population before and after propensity score‐based overlap weighting.Before PS‐based overlap weightingAfter PS‐based overlap weightingAmiodarone (n = 5325)Flecainide, sotalol, or propafenone (n = 7405)SMDAmiodarone (n = 2742)Flecainide, sotalol, or propafenone (n = 2742)SMDMean age (SD), year69.2 (10.3)67.6 (10.1)0.1668.6 (10.0)68.6 (10.3)0.00Male sex, %3226 (60.6)4426 (59.8)0.021639 (59.8)1639 (59.8)0.00Income level, %0.080.00Q1 (0th‐25th percentile)1484 (27.9)1850 (25.0)725 (26.4)725 (26.4)Q2 (26th‐50th percentile)854 (16.0)1143 (15.4)437 (15.9)437 (15.9)Q3 (51st‐75th percentile)1114 (20.9)1601 (21.6)581 (21.2)581 (21.2)Q4 (76th‐100th percentile)1873 (35.2)2811 (38.0)999 (36.4)999 (36.4)No. of SU prescription, mean (SD)8.2 (4.6)8.0 (4.4)0.028.2 (4.4)8.2 (4.7)Type of SU0.060.00Glimepiride5177 (97.2)7265 (98.1)2675 (97.6)2675 (97.6)Gliclazide142 (2.7)129 (1.7)63 (2.3)63 (2.3)Glyburide6 (0.1)11 (0.2)3 (0.1)3 (0.1)Diabetic complications, %Retinopathy1245 (23.4)1909 (25.8)0.06662 (24.1)662 (24.1)0.00Neuropathy1035 (19.4)1368 (18.5)0.03526 (19.2)526 (19.2)0.00Nephropathy287 (5.4)376 (5.1)0.01145 (5.3)145 (5.3)0.00Antidiabetic medications, %Metformin4431 (83.2)6267 (84.6)0.042295 (83.7)2295 (83.7)0.00DPP4is2611 (49.0)3709 (50.1)0.021346 (49.1)1346 (49.1)0.00SGLT2is481 (9.0)598 (8.1)0.03245 (9.0)245 (9.0)0.00Comorbidities, %Atrial fibrillation1829 (34.4)2903 (39.2)0.10994 (36.3)994 (36.3)0.00Heart failure966 (18.1)655 (8.9)0.28347 (12.7)347 (12.7)0.00PCI371 (7.0)128 (1.7)0.2687 (3.2)87 (3.2)0.00Coronary artery disease1633 (30.7)1591 (21.5)0.21691 (25.2)691 (25.2)0.00Ischemic stroke435 (8.2)638 (8.6)0.02232 (8.5)232 (8.5)0.00Peripheral artery disease474 (8.9)677 (9.1)0.01246 (9.0)246 (9.0)0.00Other cardiac arrythmias585 (11.0)1337 (18.1)0.20366 (13.4)366 (13.4)0.00Other paroxysmal tachycardia146 (2.7)362 (4.9)0.1193 (3.4)93 (3.4)0.00Ventricular tachycardia108 (2.0)45 (0.6)0.1323 (0.8)23 (0.8)0.00Transient ischemic attack121 (2.3)183 (2.5)0.0168 (2.5)68 (2.5)0.00Comedications, %Warfarin618 (11.6)623 (8.4)0.11273 (10.0)273 (10.0)0.00Apixaban462 (8.7)523 (7.1)0.06211 (7.7)211 (7.7)0.00Edoxaban359 (6.7)506 (6.8)0.00185 (6.7)185 (6.7)0.00Rivaroxaban314 (5.9)454 (6.1)0.01165 (6.0)165 (6.0)0.00Aspirin2954 (55.5)3643 (49.2)0.131415 (51.6)1415 (51.6)0.00P2Y12 inhibitors1733 (32.5)1513 (20.4)0.28696 (25.4)696 (25.4)0.00RAS inhibitors3700 (69.5)4743 (64.1)0.121830 (66.8)1830 (66.8)0.00Non‐Dihydropyridine CCB1035 (19.4)1276 (17.2)0.06481 (17.5)481 (17.5)0.00Dihydropyridine CCB2584 (48.5)3548 (47.9)0.011333 (48.6)1333 (48.6)0.00β‐blockers3054 (57.4)3877 (52.4)0.101463 (53.4)1463 (53.4)0.00Diuretics3006 (56.5)2946 (39.8)0.341329 (48.5)1329 (48.5)0.00Statins3541 (66.5)4811 (65.0)0.031773 (64.7)1773 (64.7)0.00Healthcare utilisationNo. of hospitalisation, %0.240.0002678 (50.3)4518 (61.0)1564 (57.0)1564 (57.0)1–22143 (40.2)2439 (32.9)977 (35.6)977 (35.6)≥ 3504 (9.5)448 (6.1)201 (7.3)201 (7.3)No. of physician visits, %0.140.000–228 (0.5)23 (0.3)12 (0.4)12 (0.4)3–553 (1.0)56 (0.8)24 (0.9)24 (0.9)≥ 65244 (98.5)7326 (98.9)2705 (98.7)2705 (98.7)Abbreviations: CCB, calcium channel blockers; DPP4is, dipeptidyl peptidase 4 inhibitors; PCI, percutaneous coronary intervention; PS, propensity score; SD, standard deviation; SGLT2is, sodium‐glucose cotransporter 2 inhibitors; SMD, standardised mean difference; SU, sulfonylureas. Selected baseline characteristics of the study population before and after propensity score‐based overlap weighting. Abbreviations: CCB, calcium channel blockers; DPP4is, dipeptidyl peptidase 4 inhibitors; PCI, percutaneous coronary intervention; PS, propensity score; SD, standard deviation; SGLT2is, sodium‐glucose cotransporter 2 inhibitors; SMD, standardised mean difference; SU, sulfonylureas. Amiodarone was associated with a higher incidence of severe hypoglycemia compared to other antiarrhythmics, both before (2.18 vs. 0.85 per 100 person‐years) and after overlap weighting (1.98 vs. 0.96 per 100 person‐years) (Table 2). Slightly more than half of the study individuals, 52.5%, were censored due to discontinuation of the antiarrhythmic agent, while 28.4% were censored due to discontinuation of sulfonylureas (Table S6). After balancing covariates using PS‐based overlap weighting, the median follow‐up duration was 90 days for amiodarone users and 101 days for flecainide/sotalol/propafenone users (Table S7). The weighted HR for severe hypoglycemia comparing amiodarone with flecainide/sotalol/propafenone was 1.92 (95% CI 1.12–3.30). The cumulative incidence of severe hypoglycemia was higher in the amiodarone group throughout the one‐year follow‐up, as shown in the cumulative incidence curves (Figure 2). TABLE 2Risk of severe hypoglycemia associated with the concomitant use of sulfonylurea and amiodarone.Amiodarone and sulfonylureasFlecainide, sotalol, or propafenone and sulfonylureasHazard ratio (95% CI)No. of patientsNo. of eventsIR per 1000 PYs (95% CI)No. of patientsNo. of eventsIR per 1000 PYs (95% CI)Primary analysisBefore PS‐based overlap weighting5325742.18 (1.74–2.74)7405530.85 (0.65–1.12)2.35 (1.65–3.36)After PS‐based overlap weighting2742341.98 (1.42–2.77)2742220.96 (0.62–1.44)1.92 (1.12–3.30)Abbreviations: CI, confidence interval; IR, incidence rate; PS, propensity score; PY, person‐year. Risk of severe hypoglycemia associated with the concomitant use of sulfonylurea and amiodarone. Abbreviations: CI, confidence interval; IR, incidence rate; PS, propensity score; PY, person‐year. Cumulative incidence curve of severe hypoglycemia comparing amiodarone versus flecainide, sotalol, or propafenone in the primary as‐treated analysisAbbreviation: SU, sulfonylureas *Individuals were followed from the initiation of antiarrhythmic drugs until the earliest event, which included discontinuation of either sulfonylureas or antiarrhythmic drugs, switching to an alternative antiarrhythmic agent, death, or the end of the study period (31 December 2022). Sensitivity analyses yielded estimates consistent with the primary analysis, supporting the robustness of our findings (Figure 3). Excluding patients with a previous diagnosis of hypoglycemia within 1 year before the index date yielded a weighted HR of 1.86 (95% CI 1.04–3.30). Similar estimates were observed when accounting for the competing risk of all‐cause mortality, identifying hypoglycemia in any care setting, applying intention‐to‐treat approaches with maximum follow‐up of 60 and 365 days, additionally censoring at insulin initiation and varying the grace period to 15 and 60 days. Excluding patients with prior use of either CYP2C9 inducers or inhibitors resulted in a weighted HR of 1.73 (95% CI 1.03–2.93) (Figure 3 and Table S8). Sensitivity analyses for the risk of severe hypoglycemia associated with the concomitant use of sulfonylurea and amiodarone after propensity score‐based overlap weighting. CI, confidence interval; HR, hazard ratios; IR, incidence rate; PS, propensity score; PY, person‐year. The increased risk associated with amiodarone was consistent across subgroups, though a higher incidence was observed in older adults and females. The weighted HR was 1.23 (95% CI, 0.22–6.92) for those under 65 years and 2.09 (95% CI, 1.18–3.71) for those aged 65 years or older. By sex, the weighted HR was 2.66 (95% CI, 1.17–6.02) in females and 1.48 (95% CI, 0.71–3.08) in males (Table S9). Discussion In this population‐based cohort study, initiation of amiodarone during sulfonylurea therapy was associated with a clinically meaningful increase in the risk of severe hypoglycemia compared with other antiarrhythmic agents. This association persisted after covariate adjustment using PS‐based overlap weighting and remained consistent across multiple sensitivity analyses. These findings support the robustness of the observed association and suggest the potential for clinically relevant drug–drug interaction between amiodarone and sulfonylureas. Previous preclinical studies have primarily focused on the pharmacokinetic interaction involving amiodarone, particularly its inhibitory effects on CYP2C9 and P‐glycoprotein. These inhibitory properties support our hypothesis that amiodarone may increase the risk of hypoglycemia by impairing sulfonylurea metabolism and transport via CYP2C9 and P‐glycoprotein pathways, thereby increasing sulfonylurea concentrations [8, 22, 23]. However, few clinical studies have directly assessed the effects of concomitant use of amiodarone and sulfonylureas. Only one population‐based nested case–control study, using Taiwan's national claims data from 2008 to 2013, has examined this interaction. It reported an odds ratio of 1.56 (95% CI, 0.98–2.46) for severe hypoglycemia associated with concomitant use of amiodarone and sulfonylureas, demonstrating a finding consistent with our results [12]. However, several limiations should be considered. As is inherent in nested case–control designs, potential bias related to control selection cannot be entirely ruled out [24] and absolute measures such as incidence rates were not provided [25]. Furthermore, comparing sulfonylurea users receiving amiodarone with those not receiving amiodarone (i.e., sulfonylurea plus amiodarone vs. sulfonylurea alone) may be influenced by underlying differences in clinical characteristics between groups, raising the possibility of confounding by indication [26, 27]. Our active‐comparator cohort design, comparing initiators of amiodarone with initiators of other antiarrhythmic agents among sulfonylurea users, was intended to reduce such confounding by aligning the clinical indication for antiarrhythmic treatment across groups. Sulfonylureas are primarily metabolised by the CYP2C9 enzyme in the liver [28]. Amiodarone and its active metabolite form, desethylamiodarone, are potent inhibitors of CYP2C9 and can significantly reduce its metabolic activity [8]. Concomitant use may therefore delay sulfonylurea clearance, increase plasma concentrations and prolong hypoglycemic effects. In addition, genetic polymorphisms in CYP2C9 may be further influence individual susceptibility to this interaction [28]. Reduced‐function CYP2C9 alleles may further increase susceptibility to the sulfonylurea‐amiodarone interaction. In vitro kinetic and pharmacogenetic studies have shown that both CYP2C9*2 and CYP2C9*3 reduce metabolic activity, with CYP2C9*3 generally exerting a substantially stronger functional effect [29]. Although the allele frequency of CYP2C9*3 in East Asian populations is modest, approximately 3.4%, it remains a relevant reduced‐function variant to consider when considering variability in susceptibility to the sulfonylurea‐amiodarone interaction [30]. Moreover, amiodarone has been reported to inhibit P‐glycoprotein function [23, 31]. Although transporter‐mediated pathways have been described for some sulfonylureas, there relevance of this mechanism to our findings remains uncertain [32, 33]. In our study population, more than 97% of sulfonylurea use was attributable to glimepiride, for which direct evidence as a P‐glycoprotein substrate is limited. Therefore, transporter‐mediated effects may represent a biologically plausible secondary pathway but should be interpreted cautiously rather than regarded as an established mechanism underlying our findings. Sulfonylurea‐induced hypoglycemia may elevate arrhythmia risk in patients with type 2 diabetes. Several studies have shown that hypoglycemia caused by sulfonylureas can alter cardiac electrophysiology [34], leading to QTc prolongation [35], ventricular arrhythmias [10] and an increased risk of sudden cardiac death [36]. Concomitant use of sulfonylureas and amiodarone may further increase this risk, warranting close monitoring and an individualised therapeutic approach. Moreover, amiodarone contains a high iodine content (approximately 39% by its molecular weight), which exceeds daily recommended intake and can lead to thyroid dysfunction [37], a known risk factor for hypoglycemia [38]. Taken together, careful therapeutic decision‐making is warranted to balance the benefits of antiarrhythmic treatment against the potential metabolic risks in this high‐risk population. In the 60‐day intention‐to‐treat sensitivity analysis, the estimate was in the same direction as the primary analysis, although it was less precise than the estimates from the primary as‐treated analysis and the 365‐day intention‐to‐treat analysis. This lower precision is likely explained by the shorter fixed follow‐up window, which captured substantially fewer severe hypoglycemia events; the weighted number of events was less than half of that in the primary analysis. Thus, the 60‐day analysis is consistent with the main finding, while providing a less precise estimate because of the smaller number of events. In our study, the incidence rate of severe hypoglycemia was substantially higher among patients aged 65 years and older than among those younger than 65 years. This finding suggests that clinicians should be cautious when initiating amiodarone in older adults receiving sulfonylureas, as well as in patients with impaired renal or hepatic function, in whom reduced drug clearance may heighten the risk of hypoglycemia [39]. To mitigate this complication, clinicians may consider sulfonylurea dose reduction, closer blood glucose monitoring, and patient education on recognising hypoglycemic symptoms. Current diabetes guidelines emphasise hypoglycemia avoidance, particularly in older adults and patients at high risk for adverse drug events [40, 41]. When modification of glucose‐lowering medication is appropriate, DPP‐4 inhibitors or GLP‐1 receptor agonists or SGLT2 inhibitors may be considered because they have a lower intrinsic risk of hypoglycemia and are less dependent on CYP2C9 metabolism [2, 28, 40, 41]. Given the high prevalence of polypharmacy among patients with type 2 diabetes and arrhythmia, proactive identification and management of metabolic drug interactions are essential to improving medication safety. Several limitations should be considered. First, although we applied PS weighting to achieve balance across measured covariates, the potential for residual confounding from unmeasured factors remains. These may include arrhythmia subtype and severity, behavioural factors such as alcohol consumption and smoking and physiological characteristics such as atrial fibrosis, heart rate, electrocardiogram abnormalities and systemic inflammation markers (e.g., C‐reactive protein). Moreover, the lack of genetic information, particularly regarding CYP2C9 polymorphisms, may have influenced the observed associations in our study and precluded us from assessing effect modification by CYP2C9 polymorphism. To mitigate these limitations, we measured a wide range of variables associated with hypoglycemia to serve as proxies for potential unmeasured confounders [42]. Second, claims data reflect prescription dispensing but do not verify actual medication intake. Consequently, we could not be certain that sulfonylureas and amiodarone were used concomitantly during follow‐up. To address these limitations, we performed several sensitivity analyses using alternative definitions of drug exposure. Third, to minimise confounding due to indication and reduce bias, we employed an active comparator new‐user design using other rhythm control antiarrhythmic agents such as flecainide, sotalol and propafenone. Nevertheless, residual difference in clinical indications may remain, because amiodarone is often prescribed for patients with structural heart disease, whereas the comparator drugs, are typically used in less complex cardiovascular cases [43]. However, such differences are unlikely to have substantially influenced the risk of hypoglycemia during the relatively short follow‐up period (median 100 days) in our study. Moreover, we used PS‐based overlap weighting to address potential baseline imbalances and achieve exact covariate balance in the weighted population [19]. Fourth, the study was not sufficiently powered to detect effect modification across subgroups defined by age, sex and clinically relevant comorbidities. Therefore, subgroup results should be viewed as exploratory and interpreted with consideration of these limitations. Finally, because our study was limited to Asian patients, the generalizability of our findings to other populations remains uncertain. For example, CYP2C9 allele distributions vary by ancestry; CYP2C9*3 has been reported to have an allele frequency of 3.4% in East Asians and 5.6% in Europeans. Further studies in diverse populations are warranted to evaluate whether the observed association differs across populations. In conclusion, initiation of amiodarone while receiving sulfonylureas was associated with an increased risk of severe hypoglycemia compared with initiation of other antiarrhythmic medications, including flecainide, sotalol and propafenone. Given the high prevalence of arrhythmia among patients with type 2 diabetes and the continued use of sulfonylureas as a cost‐effective glucose‐lowering therapy, heightened clinical awareness and close monitoring for hypoglycemia are warranted when these agents are used concomitantly. Author Contributions S.B., H.L., J.‐Y.S. and K.B. designed the study and H.L. performed statistical analysis. All authors interpreted the data. S.B. and H.L. drafted the manuscript. Y.M.C. and Y.‐J.C. critically revised the manuscript. J.‐Y.S., as guarantors of the study, accepts full responsibility for the results of this study, had access to the data and controlled the decision to publish. Funding K.B. was supported by National Institute of Health grant K01AG068365. JYS was supported by a grant (22183MFDS431) from the Ministry of Food and Drug Safety, South Korea, from 2022 to 2025. S.B. was supported by a grant from the American Diabetes Association (grant No. 1‐25‐PDF‐66). Ethics Statement Ethical approval was obtained from the Institutional Review Board of Sungkyunkwan University, where the requirement of informed consent was waived as this study used anonymized administrative data (IRB No. SKKU 2021–08‐003). Conflicts of Interest J.‐Y.S. received grants from the Ministry of Food and Drug Safety, the Ministry of Health and Welfare, the National Research Foundation of Korea and pharmaceutical companies, including LG Chem, UCB, Pfizer, Celltrion and SK Bioscience, outside the submitted work. The remaining authors declare no conflicts of interest. Supporting information Figure S1: Graphical depiction of the study design. Table S1: Specification and emulation of target trial assessing the risk of severe hypoglycemia associated with concomitant use of sulfonylurea and amiodarone using administrative claims data. Table S2: Definition of outcome. Table S3: Definitions for comorbidities and previous medication use. Table S4: Cytochrome P450 2C9 inducer and inhibitor medications included in the sensitivity analyses. Table S5: Baseline characteristics of the study population before and after propensity score‐based overlap weighting. Table S6: Censoring events. Table S7: Median follow‐up time. Table S8: Sensitivity analyses for risk of severe hypoglycemia associated with the concomitant use of sulfonylurea and amiodarone. Table S9: Subgroup analyses for the risk of severe hypoglycemia associated with the concomitant use of sulfonylurea and amiodarone.