Nonsteroidal anti-inflammatory drugs (NSAIDs) are after antibiotics, the second most frequently suspected agents causing drug hypersensitivity. The prevalence of acetyl salicylic acid (ASA, aspirin) hypersensitivity ranges from 0.6% to 2.5% in the general population, from 4.3% to 11% in asthmatic patients [
Hypersensitivity may occur shortly, within 15 minutes or longer, up to 24 hours after NSAID intake. In general it develops within 1 to 4 hours [
In most patients the adverse reaction is nonallergic. Those with eicosanoid metabolism dysfunction or other alterations are prone to hypersensitivity when NSAIDs inhibit the enzyme cylooxygenase-1 (Cox-1) [
Comparison of NSAIDs and acetaminophen concentrations incubated with leukocytes to serum concentrations at usual therapeutic dosage and to 50% inhibitory concentrations (IC50) of cyclooxygenase-1 and -2.
Drug | Concentrations | ||||
Tested* | In serum** | IC50*** | |||
mg/mL | COX-1 | COX-2 | |||
ASA | 0.01–0.1–1 | 31–310–3100 | 111 | 4.45 | 13.88 |
Diclofenac | 0.0013–0.013–0.13 | 4.2–42.2–422 | 6.1 | 0.26 | 0.01 |
Ketoprofen | 0.025–0.25–2.5 | 98.3–983–9830 | 9.4 | 0.11 | 0.88 |
Celecoxib | 0.005–0.05–0.5 | 13.1–131–1310 | 1.6 | 82 | 6.8 |
Acetaminophen | 0.01–0.1–1 | 66.1–661–6610 | 117 | 113.7 | 25.8 |
*Each NSAID and and acetaminophen (APAP) were tested at three ten-fold serial dilutions.
**Serum concentrations at usual therapeutic dosage
***Concentration of drug that inhibited 50% of COX-1 in platelets or COX-2 in monocytes [
The diagnosis of NSAID hypersensitivity is based on clinical histories and provocation challenges with aspirin or NSAIDs [
In ASA-induced urticaria or asthma, in addition to basophils, neutrophils or other leukocytes are also activated [
Sixty-five patients referred by the patients’ physician or by an emergency unit to the Dermatology and Allergy Center of Hôpital Tenon, Paris, between 2006 and 2009 for evaluation of a history of NSAID and/or APAP (acetaminophen, paracetamol) hypersensitivity were included in the study. Among them, 5 patients had APAP hypersensitivity alone. Half of patients according to the physician report or the patient’s declaration had a history of hypersensitivity to one of these drugs during the last year, for the other half it was older.
NSAID-induced or exacerbated urticaria or angiooedema and associated symptoms were clinically classified [
Patients suspected of immediate or delayed allergy to an NSAID were discarded. Those with asthma or Widal syndrome were not included because they were followed up in the department of respiratory system diseases.
The study protocol was in accordance with the local ethical committee guidelines, and all subjects gave their consent before being included. Tests were done only for a diagnostic purpose.
Leukocytes from 12 normal subjects, members of the hospital’s staff, and 8 patients with allergic reactions induced by drugs other than NSAIDs were used as controls for BAT, MAT, and NAT. All the controls had never experienced NSAID hypersensitivity and had taken at least once 1 g aspirin within the last 12 months.
Skin tests (STs), prick ST and intradermal ST, were done as previously described [
In this study no oral challenge test with NSAIDs was done for a diagnostic purpose. In patients needing an analgesic antipyretic drug, double blind, placebo-controlled, orally given APAP tests were carried out when the STs were negative. Protocol for patients needing NSAID therapy was similar, they received selective NSAIDs celecoxib or nimesulide. All the oral challenges were done by the same practitioner.
They were purchased in solution for intravenous or intramuscular use. Acetyl salicylate lysine (Aspegic, Sanofi-Aventis), acetaminophen (APAP, paracetamol, perfalgan, Bristol-Myers Squibb), ketoprofen (profenid, Sanofi-Aventis), diclofenac (voltaren, Novartis Pharma) and celecoxib (Celebrex, Pfizer Inc.) were diluted in the dilution buffer (2 mM MgCl2, 1.2 mM CaCl2, and 2 g/L bovine serum albumin in phosphate-buffered saline, PBS). The dilution buffer was used, instead of a NSAID, as the negative control in two tubes, and calcium ionophore 2.5
R-phycoerythrin-conjugated anti-CD33 monoclonal antibody (MAB), fluorescein-isothiocyanate- (FITC-) conjugated anti-CD63 MAB, R-phycoerythrin-conjugated anti-CD203c MAB, tandem dye R-phycoerythrin-cyanin-5.1-conjugated anti-CD45 MAB, and ECD- conjugated (tandem dye phycoerythrin-Texas red) streptavidin were purchased from Beckman Coulter-Immunotech, Marseille, France. Biotinylated goat anti-human IgE polyclonal antibody was purchased from Vector Laboratories, Burlingame.
The BASIC (basophils isolated from blood and analysed by flow cytometry) assay was done as previously described [
100
The leukocytes were then quadruple labelled by adding 10
Analysis of leukocytes surface markers was performed on an Epics XL flow cytometer (Beckman Coulter, Marseille, France). On the histogram defined by forward scatter and side scatter, the first gating was done by a bit map around lymphocytes and monocytes, basophils were found in this gate. Another gate was done around polymorphonuclear leukocytes, for neutrophils. The next gatings were done for basophils around IgE+ CD45+ CD33dim cells (Figure
Detection by flow cytometry of CD63 upregulation on leukocytes activated by NSAIDs. (a) and (b) Targeting of leukocytes: (a) shows at the top CD33 bright cells, the monocytes, CD33 dim cells, polymorphonuclears, and CD33-negative cells, lymphocytes. (b) Shows among anti-IgE-labelled cells, CD33dim cells, the basophils, and C33 bright, the monocytes. The percentage of monocytes among IgE-labelled cells varies from 0 to 50%, depending on the patient. (c) and (d) Activation of basophils: (c) basophils incubated with buffer. (d) Basophils incubated with 1 mg/mL of ASA and upregulation of CD63 on 30% of basophils. (e) and (f) Activation of monocytes and neutrophils. (e) Monocytes at the top, and neutrophils under, incubated with buffer. (f) After incubation with 1 mg/mL of ASA, upregulation of CD63 on 25% of monocytes and 5% of neutrophils.
In each assay, upregulation of CD63 was measured on at least 400 cells for basophils, monocytes, or neutrophils.
The tube-to-tube reproducibility was determined by labelling leukocytes of 10 patients and counting 12 tubes per patient. A cut-off value for positive results was chosen at 2 standard deviations (6%) exceeding the value of the nonstimulated control tube.
The best cut-off value for activated basophils and monocytes was established by analysing the receiver operating characteristic (ROC) curves of results observed in a selected group of patients with severe hypersensitivity versus controls (Figure
ROC curves for basophils and monocytes incubated with 1 mg/mL of ASA. (a) shows the sensitivity (Se) and the specificity (Sp) of BAT. The cut-off value was determined at 6% of activated basophils (vertical dashed line). (b) shows ROC curves for the activation of basophils and monocytes. The diagnostic performance of the activation of basophils was better (area under the curve for activated basophils 0.855) than that of the activation of monocytes (area under the curve 0.718),
Results of BAT and MAT were blind-analysed, the operator was not informed of the patients’ diagnosis. Conversely, the diagnosis of allergy was done before the results of BAT and MAT were available. Statistical analysis was done with
Among 60 patients with NSAID hypersensitivity, 30 had no underlying disease and had recovered when they abstained from taking NSAIDs. These patients were diagnosed as having NSAIDs-induced urticaria/angioedema according to [
The remaining 30 patients had chronic and/or episodic (intermittent) idiopathic urticaria and/or angiooedema (C/E IU). NSAID-induced angiooedema was as frequent as urticaria. These patients were diagnosed as having NSAIDs-exacerbated urticaria/angioedema according to [
Description of patients suffering from NSAID hypersensitivity.
Group of patients* | Age | Sex ratio | Atopy** | C/E IU*** | HS ≥ 2**** NSAIDs | |
---|---|---|---|---|---|---|
Grade I | 44 | 29F/9M | 9 (24%) | 18 (47%) | 25 (66%) | |
Grade II | 43 | 16F/6M | 7 (32%) | 12 (55%) | 17 (77%) | |
Total | 44 | 45F/15M | 16 (27%) | 30 (50%) | 42 (70%) |
*Patients with a history of urticaria or angiooedema and no visceral disorder were classified as having had a hypersensitivity of grade I and those with at least one visceral disorder were classified as grade II according to published works [
**Patients had history of atopic dermatitis, allergic rhinitis, or asthma, but they had healed at the time of NSAID hypersensitivity.
***The patients with “C/E IU” had chronic or episodic idiopathic urticaria or angiooedema; NSAID hypersensitivity was diagnosed when symptoms worsen relapsed, or were unusual and recovered when they stopped NSAID intake.
****Patients who had histories of hypersensitivity induced by at least two chemically distinct NSAIDs.
Skin symptoms and time to onset of the NSAID-induced hypersensitivity.
Group of patients | Time to onset* | NSAID-induced skin symptoms | ||||
Median | <6 H | AO | AO + Urticaria | Urticaria | ||
Grade I | 8 H | 14 (37%) | 18 (47%) | 9 (24%) | 11 (29%) | |
Grade II | 1 H | 21 (95%) | 12 (55%) | 7 (32%) | 3 (14%) | |
Total | 4.5 H | 35 (58%) | 30 (50%) | 16 (27%) | 14 (23%) |
*The time to onset of symptoms after NSAID intake was shorter in patients with grade II hypersensitivity than in patients with grade I (
Forty-two among 60 patients (70%) had histories of hypersensitivity induced by at least two chemically distinct NSAIDs (Table
Twenty-two patients with a history of urticaria or angiooedema and with at least one visceral disorder hypotension, laryngeal oedema, dyspnoea, abdominal pain, vomiting or diarrhoea after NSAID intake, were classified as having had a hypersensitivity of grade II (Table
Description of severe reactions (grade II) observed in patients with NSAID hypersensitivity.
Grade II reactions/time to onset* | Visceral disorders (VD) | |||||
Laryngeal oedema | Dyspnoea** | Hypotension | G-intestinal disorders | At least one VD*** | ||
<6 H | 9 (43%) | 8 (38%) | 6 (29%) | 3 (14%) | 21 (100%) | |
8 H | 1 | 0 | 0 | 0 | 1 | |
Total | 10 (45%) | 8 (37%) | 6 (27%) | 3 (14%) | 22 (100%) |
*Time to onset of symptoms after NSAID intake. **Dyspnoea observed in patients with no laryngeal oedema. ***Patients with a history of urticaria or angiooedema and with at least one visceral disorder were classified as having had a hypersensitivity of grade II according to published works [
Among 18 patients with only one known NSAID hypersensitivity, 7 had at least one visceral disorder, frequency (7/18) was similar to patients with hypersensitivity induced by at least two chemically distinct NSAIDs (15/42). No clinical or biological data could discriminate one group from the other.
Targeting of leukocytes and detection of CD63 upregulation on the membrane of activated cells were done as indicated in Figure
Twenty-two out of 60 patients with a history of NSAID hypersensitivity had a positive BAT to ASA at 1 mg/mL; therefore the sensitivity was 37%. Two controls among 20 (10%) had positive BAT to ASA, the specificity of the test could be estimated close to 90%.
In patients suffering from NSAID hypersensitivity restricted to cutaneous reaction (grade I), positive BAT to ASA was not more frequent than in control group (Figure
Activation of basophils (BAT), monocytes (MAT), and neutrophils (NAT) induced in vitro by ASA (1 mg/mL) or diclofenac (0.125 mg/mL).
Patients with positive tests* | |||||||
Group of patients | BAT+ | MAT+ | NAT+ | ||||
ASA | Diclofenac | ASA | Diclofenac | ASA | Diclofenac | ||
NSAID HS | 22 (37%) | 20 (33%) | 14 (23%) | 15 (25%) | 8 (13%) | 16 (27%) | |
Grade I** | 8 (21%) | 10 (26%) | 4 (11%) | 9 (24%) | 6 (16%) | 12 (31%) | |
Grade II | 14 (64%) | 10 (46%) | 10 (46%) | 6 (27%) | 2 (9%) | 4 (19%) | |
Controls | 2 (10%) | 5 (25%) | 5 (25%) | 3 (15%) | 0 (0%) | 2 (10%) |
*The optimal cut-off point for a positive test deduced from ROC curves was 6% of activated cells (see Section
Activation of patients’ and controls’ basophils and monocytes incubated with 1 mg/mL of ASA. (a) Basophils, (b) monocytes. Grading of reactions in the NSAID hypersensitivity was done according to severity of clinical symptoms [
In contrast, in patients who had had a grade II hypersensitivity, basophils were more strongly activated and BAT was more frequently positive (14/22 = 64%) than in patients with grade I (8/38 = 21%) or controls (Figure
A positive BAT to ASA correlates with the precocity of the hypersensitivity. The patients with a positive BAT had reported an interval of 2.5 hours (median value) between NSAID intake and symptoms. On the other hand among patients with negative test the interval was 8 hours (
Though less sensitive (46%), results of MAT correlate quite well with BAT (Spearman’s rank correlation,
Activation of neutrophils with ASA was rather insignificant (Table
Activation by diclofenac of basophils and monocytes was of the same magnitude of ASA and results correlated rather well (Spearman rank correlation,
Each NSAID and APAP were tested at three ten-fold serial dilutions (Table
In 23 patients and 8 controls, basophils were double labelled with the two activation markers. Activation of patients’ basophils with ASA, diclofenac, or ketoprofen was detected 32 times with CD63, 6 times with CD203c, and 4 times with both CD63 and CD203c (results not shown). CD63 was at least 5 times more sensitive than CD203c (
The sensitivity (37%) and specificity (90%) of BAT for the diagnosis of NSAID induced hypersensitivity were low. These values rather agree with published results. The sensitivity of BAT was assessed between 33 and 77% with ASA and between 17 and 52% with diclofenac [
Our study further shows that, though less frequent than on the basophils, ASA upregulates CD63 on the monocytes of some patients with NSAID hypersensitivity and of some control subjects who tolerate ASA well. Because BAT and MAT have a low sensitivity and specificity in the diagnosis of NSAID hypersensitivity, this raises the question of the clinical significance of a positive test.
Our results suggest a linkage between positive BAT or MAT with ASA and a history of a severe NSAID hypersensitivity. Indeed positive tests were more frequent in patients with severe hypersensitivity than in patients with only cutaneous symptoms or controls (
The patients included in this study had been suspected of a nonallergic hypersensitivity to NSAIDs. For most of them this was confirmed because 70% had histories of hypersensitivity with at least two chemically distinct NSAIDs; moreover, in addition to basophils, monocytes were activated in vitro by NSAIDs. Because activation of monocytes is not IgE dependent, a positive MAT might contribute to identify a nonallergic hypersensitivity. Among 18 patients with only one known NSAID hypersensitivity, 7 patients were MAT positive with ASA or diclofenac.
Results observed in BAT and MAT agree with clinical studies about harmlessness of celecoxib, a selective COX-2 inhibitor [
Twenty out of 60 (33%) patients reported in addition to NSAID APAP hypersensitivity, but, when tested with APAP, skin tests, BAT, and MAT were negative. The mechanism by which APAP affects fever and pain is still debated. It remains a weak COX inhibitor, but it is more potent in inhibiting COX-2 than COX-1, like selective COX-2 inhibitors [
One team published contradictory results about the sensitivity of the BAT, it reported that diclofenac induces basophil degranulation without increasing CD63 expression in sensitive patients [
CD203c upregulation, compared to CD63, poorly detected basophil activation by NSAIDs, but there is a controversy about CD203c upregulation by NSAIDs [
In summary, in patients suffering from NSAID hypersensitivity restricted to cutaneous reaction, in vitro activation of basophils or monocytes by NSAIDs was similar to that of control subjects. In contrast, a group of patients who had had early and quite severe reactions (grade II) had with ASA significantly stronger activation of basophils and monocytes.
The authors state no conflict of interests.
The authors thank for critical discussions and review of the paper Roland Abuaf, M.S., and Gregoris Gerotziafas, M.D., Ph.D. This work was supported by Hôpital Tenon (AP-HP) and Université Pierre et Marie Curie, Paris.