Adverse drug reactions (ADRs) account for 3–6% of all hospital admissions and occur in 10–15% of hospitalized patients, contributing to increased morbidity, prolonged hospitalization, and mortality. Most ADRs are type A reactions, whichare predictable and dose-dependent, based on the drug’s pharmacological properties. In contrast, drug hypersensitivity reactions (DHRs) are type B reactions that are unpredictable, dose-independent, and comprise approximately 6–10% of all ADRs [1]. Although less common, DHRs are clinically important because they can occur with minimal drug exposure, are prone to recurrence upon reexposure, and may present with severe manifestations such as anaphylaxis or severe cutaneous adverse reactions (SCARs). Therefore, early recognition and accurate diagnosis of DHRs are crucial for the management of patients with ADRs.

Several studies have examined the epidemiology of ADRs; however, many did not differentiate subtypes, such as hypersensitivity reactions, and diagnoses were often made in the emergency department (ED) without subsequent confirmation by an allergist. Consequently, the true incidence and clinical features of DHRs in real-world settings remain unclear [2]. In particular, while most medications are prescribed in outpatient settings, it is challenging to monitor adverse events that occur post-prescription, making the comprehensive evaluation of DHRs outside hospital settings challenging.

Although EDs are not dedicated outpatient settings, they can serve as important sites for the initial evaluation of drug-related adverse events that occur outside the hospital, particularly in acute presentations where temporal relationships are more apparent [3,4]. This is especially relevant when patients present soon after drug exposure, which allows clinicians to assess causality with fewer confounding factors.

Although previous studies have examined drug-related adverse events in the ED, one study focused on ADRs in general [5], while others addressed DHRs without allergist confirmation, raising concerns about diagnostic accuracy [6–9]. Given the broad clinical spectrum of DHRs, specialized evaluation using skin tests, serum-specific immunoglobulin E (IgE) assays, or drug provocation tests are often required; otherwise, both over- and underdiagnosis may occur [10]. Referral to allergists is therefore recommended; however, no study has investigated post-ED evaluation to confirm or refine the initial diagnosis.

To address this gap, this study aimed to investigate the clinical features of DHRs identified in the ED of a university hospital in Korea. By incorporating follow-up evaluations performed by allergists, this study sought to improve the diagnostic accuracy and provide a more comprehensive understanding of drug hypersensitivity in real-world practice.

This single-center retrospective study characterized DHRs presenting to the ED and evaluated culprit drug identification through allergy outpatient follow-ups. Among the 668 patients, 63.9% had immediate-type reactions and 36.1% had delayed-type reactions. Cutaneous manifestations were common in both immediate- and delayed-type reactions; however, respiratory, cardiovascular, and neurological symptoms were more frequently observed in immediate-type reactions. Of the 427 patients with immediate-type reactions, 33.5% met the criteria for anaphylaxis, which was associated with a higher prevalence of underlying allergic diseases and an epinephrine administration rate of 45.5%. Radiocontrast media, NSAIDs, and β-lactams are the most common culprit drug classes, with reactions involving multiple suspected drugs occurring frequently. Only 12.9% of patients had outpatient follow-up, but among those who did, 59.3% had a confirmed culprit drug, most commonly cefaclor, amoxicillin/amoxicillin-clavulanate, NSAIDs, or eperisone. In patients with multiple suspected drugs, the NSAID plus β-lactam combination was the most common, with more than half having the culprit drug within this combination. Building on these methodological strengths, several notable findings emerged in our cohort.

Previous large-scale studies in Spain and Croatia identified β-lactam antibiotics as the most common culprits of ED visits in DHRs [6,8]. However, these studies relied mainly on ED records or diagnoses without allergist confirmation, limiting the diagnostic accuracy. Follow-up rates were also low, with < 10% in one Lebanese study [7] and as low as 2.3% in a pediatric ED series from Australia [13], and confirmatory testing was rarely performed.

In contrast, our study directly linked ED cases to allergist evaluation, enabling a definitive diagnosis through a multimodal approach that included medical history, ImmunoCAP, skin testing, and drug provocation testing. By analyzing both immediate- and delayed-type reactions, we also compared phenotypes, including anaphylaxis and non-anaphylaxis, and assessed diagnostic yield across multiple drug exposures. Taken together, these features allow us to identify several notable findings. First, radiocontrast media were the leading culprit in our cohort, likely reflecting their high utilization in Korea’s diagnostic and therapeutic settings, with national surveillance data showing more than a tenfold rise in related ADR reports between 2009 and 2019 [14]. Second, the unexpectedly high proportion of delayed reactions suggests that such events can be sufficiently severe or distressing to prompt ED visits, underscoring the underrecognized clinical burden [15]. We further delineated delayed phenotypes, showing that most delayed ED presentations manifested as MPE, whereas SCAR was uncommon and limited to a single case clinically suspected of SJS. This distribution is consistent with the expectation that benign exanthems predominate in ED-identified delayed reactions, while true SCARs are infrequent but high-stakes and warrant targeted evaluation and early referral. Third, among anaphylactic patients, comorbid asthma and allergic rhinitis were more prevalent, consistent with known risk factors [16,17], and elevated serum tryptase (≥ 11.4 ng/mL) was significantly more frequent, supporting its diagnostic value [18,19]. Fourth, although the epinephrine administration rate of 45.5% in anaphylaxis exceeded that reported in some previous studies [13, 20], this rate primarily reflects administration in Brown grade 3 cases and may therefore not indicate underuse. Nevertheless, this underscores the ongoing need for education on timely recognition and administration in clinical practice [21]. Finally, the high confirmation rate of culprit diagnoses among follow-up patients, an observation not reported in prior ED studies, demonstrates the clinical value of post-ED allergy evaluation for securing accurate diagnoses and preventing reexposure. This study underscores the need to strengthen the continuum of care for patients presenting to the ED with DHRs. First, establishing a direct pathway for allergy specialist assessment, either during the ED visit or through structured referral and appointment systems at discharge, could improve diagnostic accuracy and patient safety. Second, preventive strategies should focus on high-risk drugs, such as radiocontrast media, NSAIDs, and β-lactam antibiotics, with particular attention to patients exposed to multiple agents, where standardized diagnostic algorithms could facilitate culprit identification. Third, greater awareness of delayed-type reactions is warranted, as these account for a substantial proportion of ED presentations and can significantly affect patient well-being. Finally, improving adherence to guideline-recommended epinephrine use for anaphylaxis will require targeted clinician education and streamlined protocols to ensure its timely administration.

This study has several limitations. First, as this was a single-center and single-nationality retrospective study, the generalizability of the findings is limited, and there is potential for selection and information bias. Second, the proportion of patients who underwent outpatient follow-up after their ED visit was low, potentially leading to an underestimation of the actual confirmation rate. Some patients who did not attend follow-up might have still had their culprit drug identified if evaluated. Third, specific clinical outcomes and severity assessments relied on medical records, which may have been subject to inaccuracies or incomplete documentation. Fourth, we classified reaction timing as immediate (≤ 1 h) and delayed (> 1 h), a simplified scheme compared with the commonly used classifications of immediate, accelerated (1–72 h), and delayed (> 72 h). This difference should be considered when compared to other studies. Fifth, owing to limited clinical information and incomplete outpatient follow-up, all NSAID hypersensitivity reactions could not be accurately classified into established subtypes. Nevertheless, among the nine patients who were fully evaluated in the outpatient clinic, seven were classified as NIUAA and two as SNIUAA. These findings underscore the importance of structured outpatient follow-up and comprehensive evaluation to identify the exact causative drug, which not only ensures the safe and appropriate use of medications in affected patients but also prevents unnecessary avoidance of the entire NSAID class, as illustrated by the two cases of NIUAA. Although cross-reactivity is primarily attributed to nonselective cyclooxygenase-1 inhibition, cross-intolerance has also been reported in a minority of patients exposed to selective cyclooxygenase-2 inhibitors, such as celecoxib. In a Korean cohort, the overall celecoxib cross-reaction rate was 10.3%, with particularly high rates of aspirin-exacerbated chronic urticaria (17.4%) and aspirin-exacerbated chronic urticaria (8.7%) [22]. The two celecoxib-reactive cases in our study likely fell within this small but clinically relevant exception category. Sixth, because the radiocontrast media in this study were predominantly ICM, with only one GBCA case, the distribution largely reflects the specific agents used at our center and may not be generalizable to other institutions or populations. Finally, owing to the retrospective observational design, causal relationships cannot be firmly established, and potential confounding factors cannot be ruled out.

In conclusion, this study is the first to provide a comprehensive analysis of patients presenting to the ED with DHRs, including the outcomes of culprit drug identification during outpatient follow-up. Both immediate- and delayed-type reactions were frequent, and radiocontrast media emerged as the most common culprit. The high confirmation rate among patients who attended follow-up emphasizes the importance of early referral to allergy specialists in the ED. Our findings highlight several priorities: preemptive risk assessment for high-risk drugs, standardized diagnostic procedures for patients with multiple drug exposures, greater recognition of delayed reactions, and improved adherence to epinephrine use for anaphylaxis. Future multicenter prospective studies are needed to validate these strategies and to establish a structured ED-based drug allergy management system.