Cross‐Reactive NSAID Hypersensitivity: Clinical Findings From Aspirin Provocation and Alternative Drug Challenge Testing

ABSTRACT

Nonsteroidal anti‐inflammatory drug (NSAID) hypersensitivity reactions are commonly reported but often overestimated due to reliance on clinical history alone. Accurate diagnosis and identification of safe alternative medications are essential for appropriate management. This retrospective study aimed to evaluate the clinical manifestations of NSAID hypersensitivity, assess the diagnostic value and safety of aspirin oral provocation testing, and investigate the tolerability of alternative medications, including acetaminophen, meloxicam, and celecoxib. Patients diagnosed with NSAID hypersensitivity through aspirin provocation testing at a single tertiary hospital were included. Demographics, reaction phenotypes, time intervals from index reactions, provocation outcomes, and tolerability to alternative agents were analyzed. Based on the aspirin provocation results, 36, 24, 192, and 58 patients were classified as having NSAID‐exacerbated cutaneous disease, NSAID‐exacerbated respiratory disease, NSAID‐induced urticaria/angioedema, and NSAID‐induced blended reactions (NIBR), respectively. Among patients suspected of NIBR based on clinical history, the confirmation rate was 74.1%. Overall, 832 oral provocation tests with alternative drugs were performed in 310 patients: 309 with acetaminophen, 307 with celecoxib, and 216 with meloxicam. The tolerability rates for acetaminophen 1300 mg, meloxicam 15 mg, and celecoxib 200 mg were 90%, 96%, and 98%, respectively. Epinephrine was administered in 19.4% of patients overall, with the highest rate observed in the NIBR group (44.8%); however, no patients required hospitalization or experienced fatal outcomes following the provocation test. Aspirin provocation testing and sequential evaluation of alternative medications can be safely performed in outpatient settings when conducted by experienced allergists. Clinical history alone is insufficient for diagnosing NSAID hypersensitivity, and confirmatory testing is crucial to avoid unnecessary drug avoidance and support accurate delabeling.

Study Highlights.

• What is the current knowledge on the topic?

  • ⚬
    COX‐1‐mediated NSAID hypersensitivity typically results in intolerance to multiple agents. Current practice emphasizes history‐based suspicion and recommends diagnostic testing to identify safe alternative medications.
• What question did this study address?

  • ⚬
    This study addressed whether current guidelines that recommend diagnosing NSAID hypersensitivity and selecting alternative medications based solely on clinical history, without provocation testing, are appropriate.
• What does this study add to our knowledge?

  • ⚬
    Up to 25% of patients show reactions inconsistent with their clinical history, challenging history‐based diagnosis. Aspirin provocation at 30‐min intervals is safely manageable by allergists. In cross‐reactive NSAID hypersensitivity, oral challenge confirmed cross‐reactivity in 4% with meloxicam and 2% with celecoxib.
• How might this change clinical pharmacology or translational science?

  • ⚬
    Accurate diagnosis is best achieved through allergist‐supervised testing rather than history alone. Acetaminophen, celecoxib, and meloxicam are effective alternative agents for patients with cross‐reactive NSAID hypersensitivity; therefore, active evaluation of tolerability to these drugs is strongly recommended in this population. This approach supports individualized patient care and reduces unnecessary drug avoidance.

1. Introduction

Nonsteroidal anti‐inflammatory drugs (NSAIDs) are among the most commonly used medications for various inflammatory and pain‐related conditions [1, 2]. In addition, they are one of the leading causes of drug hypersensitivity reactions worldwide [1, 3]. The phenotypes of NSAID hypersensitivity can be classified as either cross‐reactive (CR) or selective reactive (SR), depending on the presence or absence of cross‐reactivity [4].

Although NSAIDs have chemically diverse structures, most exert their effects via COX‐1 or COX‐2 inhibition [4, 5]. Hypersensitivity to multiple chemically unrelated NSAIDs is largely attributed to COX‐1 inhibition, while SR is mediated by drug‐specific immunologic mechanisms involving immunoglobulin E or T cells [4]. Patients with SR usually react only to NSAIDs within the same chemical class, whereas those with CR hypersensitivity exhibit immediate‐type symptoms following COX‐1 inhibitor intake and are intolerant to most NSAIDs [4, 6]. While CR is more prevalent than SR, one study reported that approximately 75% of NSAID hypersensitivity cases could be categorized as CR [3, 7].

However, in clinical practice, NSAID hypersensitivity reactions are often assumed based solely on patient history, leading to mislabeling [8]. Consequently, NSAIDs are often unnecessarily avoided, leading to increased reliance on opioids [9, 10, 11]. One study demonstrated that unconfirmed NSAID‐related adverse drug reactions (ADRs) are associated with increased opioid prescriptions, particularly in postpartum pain management [9]. Similarly, patients who reported NSAID‐related ADRs more frequently received cumulative opioid doses at discharge following musculoskeletal illness [8, 10]. Exposure to opioids for medical illness is not only associated with immediate side effects, such as nausea, but also an increased risk of serious opioid‐related events with prolonged use [9].

Thus, accurate diagnostic testing and identification of safe alternative medications are essential for patients with CR NSAID hypersensitivity; evidence suggests that some may tolerate selective COX‐2 inhibitors or weak COX‐1 inhibitors [12, 13, 14, 15]. However, adverse reactions to these alternatives have been observed, particularly at higher doses or in patients with underlying conditions [16, 17]. A Korean study reported CR rates of 24.8% for acetaminophen and 10.3% for celecoxib, indicating that even widely considered “safe” alternatives may not be universally tolerated [18].

Therefore, this study aimed to clarify the diagnostic utility of aspirin provocation testing in patients with CR NSAID hypersensitivity and to evaluate their clinical characteristics, phenotype distribution, and tolerability to commonly used alternative medications within a Korean population.

2. Materials and Methods

2.1. Patients

We reviewed the medical records of patients who were issued a drug allergy alert card for CR to NSAIDs at a single tertiary hospital in Korea between March 2017 and December 2024 [19]. Only patients diagnosed with CR to NSAID hypersensitivity, confirmed by a positive oral aspirin provocation test performed in the outpatient clinic, were included in the analysis.

Patients who were positive to oral aspirin challenge were further evaluated with additional oral provocation tests to identify tolerable alternative medications. Acetaminophen, celecoxib, and meloxicam were assessed as potential substitutes. The choice of alternative drugs for testing was based on clinical judgment and patient history. Importantly, no delayed‐onset reactions were observed among participants beyond the immediate post‐challenge observation period. Clinical data, including age, sex, comorbid allergic diseases, time interval between the suspected NSAID‐induced hypersensitivity reaction and the oral aspirin provocation test, type of hypersensitivity reaction reported at the initial visit, and results of oral provocation tests with aspirin, acetaminophen, celecoxib, and meloxicam, were collected from electronic medical records.

2.2. Aspirin Oral Provocation Test

The oral aspirin provocation test was performed based on the EAACI/GA²LEN guidelines, with modifications tailored to our outpatient clinic setting [20]. The key differences from the original protocol were as follows: (1) an open‐label design was used instead of a single‐blind design; (2) six‐step dose escalation (5, 15, 50, 150, 280, and 500 mg) was administered at 30‐min intervals, rather than the original five‐step protocol at 90‐min intervals; and (3) the entire challenge was completed in a single day without a preceding placebo day. Detailed information is provided in Table S1.

The test, an open challenge of patients without any acute medical condition, was performed under the direct supervision of a physician and technicians experienced in performing aspirin provocation tests. Patients with a baseline forced expiratory volume in 1 s (FEV₁) of < 70% of the predicted value were excluded from the test. Additionally, pregnant women or those receiving beta‐blocker therapy did not undergo testing. Medications that could potentially affect the outcome of the provocation test were also discontinued before the oral challenge. These included short‐acting bronchodilators, inhaled corticosteroids, long‐acting bronchodilators, long‐acting theophylline, short‐acting antihistamines, systemic corticosteroids, and leukotriene antagonists.

The challenge was completed within a single day, with the procedure outlined as follows. FEV₁ was measured, and the baseline value was chosen as the best of three consecutive efforts. Patients with an FEV₁ of ≥ 70% of the predicted value were eligible to proceed to the next stage of the test. To monitor bronchial response during cumulative dosing, FEV₁ was measured before each aspirin dose and at 30‐min intervals thereafter. Details regarding the timing of FEV₁ measurements and other relevant procedures are provided in Table S1.

Aspirin was administered orally in six stepwise increasing doses of 5, 15, 50, 150, 280, and 500 mg at 30‐min intervals, up to a cumulative dose of 1000 mg. In patients who did not exhibit any reaction following the 280 mg dose (cumulative dose of 500 mg), an additional 500 mg dose was administered 30 min later, resulting in a total cumulative dose of 1000 mg. After the final 500 mg dose (cumulative dose of 1000 mg), patients were observed for an additional 2 h. Clinical symptoms were also closely monitored throughout the test. Subjective symptoms (e.g., throat discomfort, nasal stuffiness, mild chest tightness) were also recorded and carefully monitored before administering the next dose. The aspirin provocation test was considered positive if any of the following occurred: cutaneous symptoms (e.g., urticaria, angioedema), respiratory symptoms (e.g., rhinorrhea, nasal congestion, dyspnea, wheezing), a ≥ 20% drop in FEV₁, or systemic reactions consistent with anaphylaxis per EAACI/GA²LEN criteria [20]. Multiple symptoms involving distinct organ systems in a single patient were counted as separate events for analysis purposes.

If the patient exhibits significant changes in symptoms and signs, the challenge was stopped, and medications, such as antihistamines, systemic steroids, short‐acting beta‐2 agonist inhalation, and intramuscular epinephrine, were administered to treat and relieve the symptoms. The feasibility of conducting the aspirin provocation test in an outpatient clinic setting was assessed, and its diagnostic performance was evaluated in terms of reaction onset timing and safety, including epinephrine usage rate and the absence of severe delayed reactions.

2.3. Classification of CR to NSAID Hypersensitivity

CR to NSAID hypersensitivity reactions were divided into four types according to the classification proposed recently by a group of experts representing the European Network Drug Allergy and European Network on Hypersensitivity to Aspirin and Nonsteroidal Anti‐Inflammatory Drugs as follows: (1) NSAIDs exacerbated respiratory disease (NERD): aspirin and NSAID‐induced respiratory reactions, such as rhinorrhea, nasal congestion, bronchospasm, wheezing, and dyspnea in patients with underlying asthma and/or chronic rhinosinusitis and nasal polyps; (2) NSAID exacerbated cutaneous disease (NECD): aspirin and NSAID‐exacerbated urticaria and/or angioedema in patients with chronic urticaria; (3) NSAID‐induced urticaria/angioedema (NIUA): urticaria and/or angioedema induced by various NSAIDs, including aspirin, in individuals without a history of underlying chronic skin and/or respiratory disease; and (4) NSAID‐induced blended reaction (NIBR): mixed or blended reactions induced by various NSAIDs but are not categorized into any of the proposed categories, such as anaphylactic reactions induced by multiple NSAIDs or blended reactions in patients with both asthma/chronic rhinosinusitis and chronic urticaria [21, 22].

2.4. Oral Provocation Test for Alternative NSAIDs

To evaluate cross‐reactivity and identify tolerable analgesics for the patients, acetaminophen and/or celecoxib and/or meloxicam oral provocation tests were performed similarly for aspirin‐intolerant patients on separate days. If the first test was positive, the next was performed after a minimum interval of 7 days. However, testing of specific drugs was omitted if a patient's clinical history indicated possible hypersensitivity to acetaminophen, celecoxib, or meloxicam. The dose‐escalating protocol for the drugs was as follows: for acetaminophen, an initial dose of 650 mg was administered, followed by an additional 650 mg after 30 min if no adverse reactions were observed; for celecoxib, an initial dose of 100 mg was administered, followed by an additional 100 mg after 30 min if no adverse reactions were observed; and for meloxicam, an initial dose of 7.5 mg was administered, followed by an additional 7.5 mg after 30 min if no adverse reactions were observed. For alternative NSAIDs, oral provocation tests were performed at 30‐min intervals. FEV₁ and blood pressure were monitored every 30 min, and if no reaction occurred 30 min after the second dose of a given drug, the next alternative agent was administered. The patient's symptoms and signs, as well as the criteria for a positive test, were identical to those used in the aspirin oral provocation test.

2.5. Ethics Statement

The study protocol was approved by the Institutional Review Board of Chonnam National University Hospital (IRB number CNUH‐2025‐033). The need for informed consent was waived.

2.6. Statistical Analysis

Continuous variables were reported as means ± standard deviations for parametric data and medians (range) for nonparametric data, while categorical variables were summarized as frequencies and percentages. Subgroup comparisons were conducted using the Mann–Whitney U and Kruskal–Wallis tests, followed by Dunn's post hoc test for continuous variables and Pearson's chi‐squared test or Fisher's exact test for categorical variables. Bonferroni correction was applied for multiple comparisons. All statistical analyses were performed using SPSS software, version 28.0 (SPSS Inc., Chicago, IL, USA), with a p‐value < 0.05 considered statistically significant.

3. Results

3.1. Demographic and Clinical Characteristics of the Study Participants

Overall, 310 patients underwent oral aspirin provocation tests with 832 alternative drug tests. The number of cases for each alternative drug was as follows: acetaminophen (309 tests), celecoxib (307 tests), and meloxicam (216 tests). The median time between the index reaction and diagnostic testing was 56 days (range: 14–9125 days). The median interval between the last hypersensitivity reaction and diagnostic testing varied depending on the type of hypersensitivity (Figure 1). In cases of NERD, 4.2% of patients demonstrated reactions even after a decade had elapsed, suggesting that this phenotype may persist for several years following the last reported reaction. Based on the clinical history before testing, 22 patients (7.1%), 42 (13.5%), 188 (60.6%), and 58 (18.7%) were classified into the NERD, NECD, NIUA, and NIBR groups, respectively. Table 1 summarizes the baseline demographic characteristics and comorbid allergic diseases, which were subsequently analyzed according to NSAID hypersensitivity phenotypes. These findings highlight the clinical heterogeneity and variability in diagnostic timing across NSAID hypersensitivity phenotypes.

FIGURE 1.

Median time from the index hypersensitivity reaction to diagnostic testing, categorized by types of NSAID hypersensitivity according to the results of the aspirin oral provocation test. NSAID, nonsteroidal anti‐inflammatory drug.

TABLE 1.

Demographic and clinical characteristics of the study participants.

	N	%
Age, median (range) (years)	43 (18–79)
Sex
Female	186	60.0
Male	124	40.0
Allergic disease
Allergic rhinitis	63	20.3
Chronic spontaneous urticaria	52	16.8
Asthma	37	11.9
Others	14	4.5

3.2. Oral Aspirin Provocation Test

Table 2 presents the clinical manifestations observed during the aspirin provocation test. Symptom events were categorized by affected organ systems, such that simultaneous reactions involving multiple systems (e.g., skin and respiratory) were recorded as separate events. Each patient exhibited a median of one symptom (range: 1–5), and 99 patients (31.9%) experienced two or more symptoms simultaneously. The most common clinical finding was skin involvement (65.6%), with angioedema being the most frequent skin manifestation (37.2%). Based on the oral aspirin provocation test results, 36 (11.6%), 24 (7.7%), 192 (61.9%), and 58 (18.7%) patients were classified into the NECD, NERD, NIUA, and NIBR groups, respectively (Table 3). A statistically significant difference in age was observed among the four groups, with patients in the NERD group being older than those in the NECD and NIUA groups (p = 0.009). Regarding underlying diseases, allergic rhinitis was significantly more prevalent in the NIBR group than in the NIUA group, and asthma was significantly more common in the NERD group than in the NIUA and NECD groups (Table 3).

TABLE 2.

Results of aspirin oral provocation test.

Reaction of aspirin provocation test	N (events) a	%
Skin
Angioedema	178	37.2
Urticaria	107	22.3
Facial flushing	29	6.1
Respiratory
Dyspnea	70	14.6
Cough	16	3.3
Nasal obstruction	26	5.4
Rhinorrhea	32	6.7
Laryngeal edema	17	3.5
Gastrointestinal	1	0.2
Cardiovascular	3	0.6

^a Each symptom was counted as a separate event. One patient could present with multiple symptoms.

TABLE 3.

Clinical characteristics of patients with nonsteroidal anti‐inflammatory drug hypersensitivity, categorized by type based on aspirin provocation test results.

	NECD (n = 36)	NERD (n = 24)	NIUA (n = 192)	NIBR (n = 58)	p
Age, median (range) (years)	36 (20–66)	52 (21–73)	43 (18–79)	43 (20–73)	0.009
Female, n (%)	21 (58.3%)	15 (62.5%)	108 (56.3%)	42 (72.4%)	0.174
Allergic disease, n (%)
Allergic rhinitis	7 (19.4%)	7 (29.2%)	30 (15.6%)	19 (32.8%)	0.025
Chronic spontaneous urticaria	36 (100.0%)	1 (4.2%)	0 (0.0%)	14 (24.1%)	0.000
Asthma	2 (5.6%)	9 (37.5%)	14 (7.3%)	12 (20.7%)	0.000
Others	4 (11.1%)	1 (4.2%)	9 (4.7%)	4 (6.9%)	0.469
Reaction to test interval, median (range)	56 (14~2190)	56 (14~3650)	56 (14~9125)	56 (14~7300)	0.888
Aspirin provocation test
Positive cumulative dose of aspirin, median (range)	495 (15~995)	495 (15~995)	495 (15~995)	495 (15~995)	0.056
Time to positive reaction, median (range) (min)	30 (5~60)	30 (10~140)	30 (5~300)	30 (5~50)	0.508
Patient history‐test result discrepancy, n (%)	2 (5.6%)	4 (16.7%)	11 (5.7%)	15 (25.9%)	0.000
Intramuscular epinephrine for symptom relief, n (%)	3 (8.3%)	11 (45.8%)	20 (10.4%)	26 (44.8%)	0.000
Cross‐reactivity to alternative drugs
Acetaminophen, n (%)	4 (11.1%)	3 (12.5%)	17 (8.9%)	7 (12.3%)	0.844
Celecoxib, n (%)	1 (2.8%)	2 (8.7%)	3 (1.6%)	0 (0.0%)	0.076
Meloxicam, n (%)	3 (11.5%)	0 (0.0%)	8 (6.3%)	2 (4.1%)	0.451

The cumulative eliciting doses that resulted in positive reactions during the provocation tests among patients diagnosed with NSAID hypersensitivity were as follows: 7 patients (2.3%) at 20 mg, 9 patients (2.9%) at 70 mg, 64 patients (20.6%) at 220 mg, 129 patients (41.6%) at 500 mg, and 101 patients (32.6%) at 1000 mg. Although slight differences exist depending on the type of NSAID hypersensitivity, hypersensitivity reactions were identified at a cumulative aspirin dose of approximately 220 mg in 26% of patients. In contrast, 74% of patients exhibited hypersensitivity reactions at cumulative aspirin doses of ≥ 500 mg.

Among all patients, the median time to positive reaction in the aspirin provocation test was 30 min (range: 5–300 min) (Table 3). According to the protocol, patients were observed in the clinic for 2 h after the final dose. However, six patients experienced symptoms beyond this observation period (Figure 2): one patient in the NERD group (at 140 min post‐final dose) and five patients in the NIUA group (at 150, 180, 190, and 300 min in one, two, one, and one patient, respectively). None of these six patients required intramuscular epinephrine.

FIGURE 2.

Timing of positive reactions during aspirin provocation by hypersensitivity phenotype. Each dot represents a patient, colored by reaction time: black (0–30 min), dark gray (31–60 min), gray (61–90 min), light gray (91–120 min), very light gray (121–150 min), and white (151–300 min).

In 32 patients (10.3%), the type of hypersensitivity inferred from clinical history differed from that confirmed through the aspirin provocation test (Table 3). A statistically significant difference was observed depending on the type of NSAID hypersensitivity (p < 0.001). When the clinical history suggested NIUA, the diagnosis was confirmed by testing in 94.3% of cases. In contrast, when NIBR was suspected based on history, the confirmation rate was 74.1%.

Medications were administered to alleviate hypersensitivity reactions based on their severity, using a combination of antihistamines, systemic corticosteroids, and intramuscular epinephrine (Table 3). Intramuscular epinephrine was administered in 60 patients (19.4%), and its use varied significantly according to the type of NSAID hypersensitivity (p < 0.001). Specifically, 44.8% of patients with provoked symptoms received intramuscular epinephrine, consistent with the rate observed for NIBR, and 45.8% of those with NERD symptoms. Meanwhile, only 8.3% and 10.4% of patients with NECD and NIUA symptoms, respectively, required its use. No cases required additional hospitalization for symptom management, and no fatal events were observed. The oral aspirin provocation test demonstrated distinct symptom patterns, reaction thresholds, and treatment requirements among different hypersensitivity subtypes. While intramuscular epinephrine was required in some patients, all reactions were effectively managed without hospitalization or serious complications.

3.3. CR Hypersensitivity to Alternative Drugs

Acetaminophen provocation tests were performed in 309 patients with NSAID hypersensitivity (24 patients with NERD, 36 with NECD, 192 with NIUA, and 57 with NIBR). Among patients who underwent acetaminophen challenge, 10.0% (n = 31) showed CR hypersensitivity. Among patients with intolerance to acetaminophen, two also showed intolerance to meloxicam. CR hypersensitivity was highest in the NERD group (12.5%), followed by the NIBR (12.1%), NECD (11.1%), and NIUA (8.9%) groups. No significant difference was observed in the cross‐reactivity to acetaminophen according to the type of NSAID hypersensitivity.

Celecoxib provocation tests were performed in 307 patients with NSAID hypersensitivity (23 patients with NERD, 36 with NECD, 190 with NIUA, and 58 with NIBR). Cross‐reactivity to celecoxib 200 mg was observed in six patients with NSAID hypersensitivity, accounting for 2.0% of those who underwent the celecoxib challenge. Cross‐reactivity was the highest in the NERD group (8.7%), followed by the NECD (2.8%) and NIUA (1.6%) groups. No significant difference was observed in cross‐reactivity to celecoxib based on the type of NSAID hypersensitivity.

Meloxicam provocation tests were performed in 216 patients with NSAID hypersensitivity (14 patients with NERD, 26 with NECD, 127 with NIUA, and 49 with NIBR). Cross‐reactivity to meloxicam 15 mg was observed in 13 patients with NSAID hypersensitivity, accounting for 6.0% of those who underwent an oral meloxicam challenge. Cross‐reactivity was the highest in the NECD group (11.5%), followed by the NIUA (6.3%) and NIBR (4.1%) groups. No significant difference was noted in cross‐reactivity to meloxicam based on the type of NSAID hypersensitivity.

Among patients with intolerance to acetaminophen, two exhibited intolerance to meloxicam, and no patient showed intolerance to celecoxib. Among patients with intolerance to meloxicam, none were intolerant to celecoxib. No patients with aspirin intolerance showed intolerance to all three drugs simultaneously. Among patients who exhibited intolerance to at least one of the three drugs (n = 48), 13 (27.1%) had a history of allergic disease as an underlying condition. However, no significant difference was observed in the prevalence of allergic disease between patients with and without intolerance to any of the three drugs. Cross‐reactivity to alternative agents was low across all subtypes, supporting their potential as viable alternatives in clinical practice. The absence of intolerance to all three alternatives and the lack of association with allergic comorbidities underscore the importance of individualized drug testing.

4. Discussion

This study aimed to evaluate the clinical performance of aspirin provocation testing and the tolerability of patients with NSAID hypersensitivity reactions to alternative medications, such as acetaminophen, meloxicam, and celecoxib. The findings support the feasibility, safety, and diagnostic value of a structured provocation protocol in this population, particularly when conducted by trained allergists in an outpatient setting.

Our findings reinforce prior concerns that NSAID hypersensitivity is frequently misclassified when based solely on patient history [8]. In this study, approximately 25% of patients exhibited reactions during provocation testing that were inconsistent with their reported clinical histories, underscoring the limitations of relying exclusively on patient‐reported information. Such misclassification may lead to the unnecessary avoidance of NSAIDs and an increased reliance on opioids, as previously documented [8, 9, 10]. One study found that unconfirmed NSAID‐related ADRs were linked to higher opioid prescriptions, particularly in postpartum pain management [9]. Similarly, patients who reported NSAID‐related ADRs more frequently received cumulative opioid doses at discharge following musculoskeletal illness [8, 10]. Exposure to opioids for medical illness is not only associated with immediate side effects, such as nausea, but also a heightened risk of serious complications with prolonged use [9].

In patients with CR hypersensitivity, aspirin provocation test was used to confirm NSAID hypersensitivity, assess cross‐reactivities, and identify safe alternative medications [11]. Although the EAACI‐recommended protocol has been the most widely adopted, its inclusion of a placebo and the 90‐min interval between doses may limit its practicality in outpatient settings [11, 20]. In contrast to previous studies [11], this study consistently implemented a 30‐min interval protocol over a 7‐year period. Among patients with positive results on the aspirin provocation test, 98.1% developed symptoms within 2 h after the final dose, while 1.9% experienced symptoms beyond this period—presumably delayed reactions—but none required epinephrine administration. These findings suggest that, when conducted under appropriate supervision, the risk associated with delayed reactions can be adequately managed.

Currently, the American Academy of Allergy, Asthma, and Immunology guidelines recommend against performing an aspirin provocation test when the diagnosis of NSAID hypersensitivity, including aspirin‐exacerbated respiratory disease (AERD), is clinically evident based on history [6]. The ENDA/EAACI further suggests that testing may be considered in patients with at least two episodes of urticaria and reactions to three or more structurally distinct NSAIDs [11]. According to Gorgulu Akin et al., among 60 patients with a clinical history suggestive of NSAID‐induced hypersensitivity, only 15% were confirmed through drug provocation testing [23]. These findings challenge the conventional approach of diagnosing NSAID hypersensitivity and selecting alternative medications based solely on clinical history.

In addition, as demonstrated in this study, approximately 25% of patients may exhibit symptoms during the provocation test that do not match their previous clinical history. This finding underscores the potential for misclassification based on history alone and reinforces the need for confirmatory testing. In addition, patients should be clearly informed prior to testing that symptoms may differ from previous experiences, including the possibility of new or unexpected reactions.

Long‐term persistence patterns of CR to NSAID hypersensitivity phenotypes remain insufficiently understood. Although previous studies have suggested that up to 70% of patients with NIUA may lose their hypersensitivity within 6 years, AERD and NERD are generally recognized as non‐resolving condition [5]. Consistently, our findings showed that 4.2% of patients with NERD had a positive aspirin provocation test although their last hypersensitivity event had occurred more than a decade earlier—the highest rate among all phenotypes examined.

Considering the known risk of severe bronchospasm in patients with NERD, graded aspirin challenges should be approached with particular caution in this group [11]. In our study, the frequency of epinephrine administration during testing was significantly higher in patients with respiratory‐involving phenotypes, such as NERD and NIBR, compared with those presenting primarily with cutaneous symptoms (e.g., NECD or NIUA). These findings suggest that clinicians should exercise heightened caution when performing oral aspirin provocation tests in patients whose clinical history indicates a potential for respiratory involvement.

Acetaminophen and celecoxib are widely used in adults with NSAID hypersensitivities [5, 6]. Acetaminophen, which is a weak selective COX‐1 inhibitor, may elicit adverse reactions at a high dose (≥ 1000 mg) in some patients with NSAID hypersensitivity, especially AERD [6, 16]. However, in this study, we found that acetaminophen has a markedly lower intolerance rate of 10.0% in our cohort [18]. In Korea and the United States, single‐ingredient oral formulations of acetaminophen for adult use are available in tablet forms of 325, 500, and 650 mg [24]. In adults, acetaminophen is typically administered at doses of 325–1300 mg every 4–6 h, with a total daily limit of 4 g [24]. In our study, the tolerable dose of acetaminophen observed was 1300 mg, which corresponds to the maximum allowable single dose for adults. This finding represents the first identification of a higher tolerable dose than the previously suggested 1000 mg, which had been considered the standard tolerable dose in patients with NSAID hypersensitivity [16]. According to previous studies, intolerance to highly selective COX‐2 inhibitors is rare in patients with NERD but occurs in approximately 33% of patients with NECD or NIUA [25, 26]. In this study, intolerance to celecoxib was observed in only 2% of patients with NSAID hypersensitivity, reinforcing its infrequent occurrence.

This study demonstrated that meloxicam, a preferential COX‐2 inhibitor, can be used as a safe alternative drug in patients with CR NSAID hypersensitivity. Although meloxicam at a dose of 15 mg may inhibit COX‐1 to some extent, previous studies have reported that approximately 80% of patients with cross‐reactivities tolerate meloxicam [27, 28]. Recent studies have reported that cross‐reactivity at a meloxicam dose of 15 mg occurs in approximately 5% of patients, regardless of age group [13, 29]. However, in our study, meloxicam was tolerated in 96% of patients with CR NSAID hypersensitivity. These findings suggest that meloxicam may be considered a safe alternative in most adult patients with NSAID hypersensitivity.

The alternative medications evaluated in this study—acetaminophen, celecoxib, and meloxicam—are approved and widely used across major healthcare systems, including those in the United States, Europe, Canada, Japan, and China. Accordingly, we believe that our findings on the tolerability of these agents in patients with CR NSAID hypersensitivity are likely to be generalizable and applicable beyond the Korean healthcare setting.

This study has some limitations. The retrospective design may have introduced selection bias, and the sample size for certain subgroups was relatively small. However, considering the true prevalence of NSAID hypersensitivity, the sample size is not insignificant. In addition, not all patients underwent standardized testing for each alternative medication, which may have affected the estimation of tolerability. Evaluation of known risk factors for cross‐reactivity, such as sensitization to aeroallergens or relevant laboratory findings, was limited, as such data were available for only one‐third of the cohort. Future prospective studies are warranted to validate these findings and further elucidate risk factors associated with cross‐reactivity in NSAID hypersensitivity.

In conclusion, this study supports the role of aspirin provocation and structured alternative drug testing as valuable tools in the evaluation and management of NSAID hypersensitivity. When guided by clinical judgment and performed by trained allergists, these procedures can be safely incorporated into routine outpatient practice. The development and dissemination of standardized protocols may enhance diagnostic accuracy, reduce unnecessary drug avoidance, and align with current trends in allergy delabeling and personalized medicine.

Author Contributions

D.W.S. and Y.‐I.K. wrote the manuscript; D.W.S., J.E.Y., and Y.‐I.K. designed the research; D.W.S., J.E.Y., and Y.‐I.K. performed the research; D.W.S. analyzed the data; D.W.S. contributed new reagents/analytical tools.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Table S1: cts70335‐sup‐0001‐TableS1.docx.

Koh Y.‐I., Yu J. E., and Sim D. W., “Cross‐Reactive NSAID Hypersensitivity: Clinical Findings From Aspirin Provocation and Alternative Drug Challenge Testing,” Clinical and Translational Science 18, no. 9 (2025): e70335, 10.1111/cts.70335.