Drug reintroduction testing and beta-lactam hypersensitivity in children: Protocols and results, a 12-year experience

Abstract

Introduction: Oral provocation test (OPT) to beta-lactam antibiotics (BL) is a gold standard in allergology investigation. We aimed to demonstrate the contribution of OPT in BL hypersensitivity (HS) indicated as a first step in diagnosis. Methods: We conducted a retrospective study from 2007 to 2019, in a single Tunisian tertiary care academic center. It concerned children with presumed non-severe allergic manifestations to BL, with a reaction that has occurred at least 6 months before the OPT. Results: We identified 35 children for inclusion. After the first OPT, a second OPT with a different BL was performed in case of a positive result of the first one. In 12 cases (34.2%), the OPT elicited a reaction. In eight cases the allergy was to penicillin and in two cases to cephalosporins (cefixim). Cross-reactivity was noted in two cases. Conclusion: An OPT to BL indicated in the first instance in non-severe reaction in children will allow a rapid diagnosis in case of suspicion of HS to BL.

Introduction

Beta-lactams (BL) are the central pillar of the current antibiotic arsenal. The cornerstone of the current antibiotic arsenal includes BL. BL is the most prescribed therapeutic class, with 65% of prescriptions in outpatient medicine and 59% in hospitals due to their efficacy, a broad spectrum of activity, good tolerance, and low cost (Unger et al., 2013).

Although the incidence of true allergy is low, the widespread utilization of BL antibiotics has resulted in a rise in antibiotic allergy labeling (Ratzon et al., 2016).

Being labeled with a BL allergy can carry a significant burden as it can lead to various risks, including the potential for adverse effects from alternative treatments, an increased risk of developing drug-resistant infections, and higher healthcare costs. Indeed, the cost of a course of antibiotics for penicillin-allergic patients is 63% higher than for non-penicillin-allergic patients (Sade et al., 2003). Therefore, a rigorous investigation is essential, either to confirm an allergy and thus avoid more severe reactions or to eliminate a differential diagnosis. This investigation includes an anamnesis first to determine the possible substances responsible for the allergic manifestations, the nature of the reactions, and the chronology of the appearance of the clinical signs skin tests are part of this investigation. They are easy to perform in a consultation, but the certainty of diagnosis is becoming more and more complex, especially in children given the frequency of benign skin lesions.

Therefore, the oral provocation test (OPT) to BL performed after the first instance of a reaction has its place in allergology exploration in children. To this end, we proposed to conduct a retrospective and analytical study with the following objectives: To describe the procedures of the reintroduction tests to BL in case of suspicion of a hypersensitivity (HS) reaction and to demonstrate the contribution of the BL OPT as a first-line test in the diagnosis of BL allergy.

Methods

This is a retrospective descriptive study conducted in the pneumo-allergology department (C) of Abderrahmane Mami Hospital in Ariana during a period from January 2007 to May 2020.

We included children over 1 year of age with suspected non-severe allergic reactions to BL for at least 6 months before the OPT. Children whose guardians did not agree to sign a written consent for the oral challenge test were excluded from the study. All possible children and their guardians were informed about the different steps of the OPT procedure, and the side effects of the test, and the guardian signed the informed consent and was signed on the day of the OPT. We reviewed the medical records of all infants. The indication for BL has been noted as well as the clinical characteristics of the initial allergic reaction.

All infants have had a pharmacovigilance consultation. Drug imputability was established according to the method of Bégaud et al. (Bégaud et al., 1985). It is composed of intrinsic imputability relative to the observation data (probable, plausible, and doubtful) and extrinsic imputability based on bibliographic data.

OPT was performed after verification of contraindications according to European Network on Drug Allergy (ENDA) recommendations (Blanca et al., 2007):

• - Proven or suspected history of type II (hemolytic anemia), type III (serum sickness, systemic vasculitis, hepatitis, interstitial nephritis), or type IV (Stevens-Johnson syndrome, toxic epidermal necrolysis, DRESS syndrome, acute generalized exanthematous pustulosis) delayed allergic reaction.

• - Some contraindications are relative: poorly controlled asthma.

All patients had a first-line opt-out without skin testing. All infants had an OPT in a hospital setting near and an emergency kit containing epinephrine, injectable corticosteroids, antihistamines, short-acting bronchodilators, filling solutions, oxygen catheter, and aerosol mask was nearby. The medical resuscitation service was notified for possible management of a severe allergic reaction. Medications that could interfere with the result of OPT were interrupted: antihistamines, systemic and local corticosteroids, and β-blockers. A 24-h hospitalization for monitoring was indicated if the patient experiences a severe allergic reaction during the OPT.

The different OPT protocols

The OPT was performed as the first step in all cases. Three molecules were used: amoxicillin, amoxicillin-clavulanic acid, and cefixime. The therapeutic dose calculated before starting OPT was 50–100 mg/Kg for amino-penicillin (amoxicillin, amoxicillin-clavulanic acid) and 8 mg/Kg for cefixime. Children included in the study during the period from 2007 to 2014 had received five doses respectively 5 mg, 10 mg, 50 mg, 100 mg, 250 mg, and the rest of the therapeutic dose for OPT amoxicillin or amoxicillin-clavulanic acid (Table 1), and respectively 2 mg, 4 mg, 10 mg, 20 mg, 40 mg, and the rest of the therapeutic dose for OPT cefixime (Table 2).

Table 1.

Penicillin OPT protocol (2007 to 2014).

T (mn)	Dose delivered
0	5 mg
30	10 mg
60	50 mg
90	100 mg
150	250 mg
210	Remaining therapeutic dose

Table 2.

Cefixime OPT protocol (2007–2014).

T (mn)	Dose delivered
0	2 mg
30	4 mg
60	10 mg
90	20 mg
150	40 mg
210	Remaining therapeutic dose

Children included in the study during the period 2015–2019 had received four doses of both penicillin and cefixime (Table 3).

Table 3.

BL OPT protocol as of 2015.

T (mn)	Dose delivered
0	1% of the therapeutic dose
30	10%
60	20%
90	70%
150	Remaining therapeutic dose

The patients were monitored for a reaction by the medical team with each dose constantly over the OPT.

In case of a positive OPT to a molecule, a second-line OPT to another molecule of the therapeutic subclass was indicated to search for cross-reactivity or a therapeutic alternative.

Interpretation of OPT

The OPT is positive if, during the test, the patient experienced allergic events requiring immediate discontinuation of OPT and administration of symptomatic treatment, or if allergic reactions occur 24–48 h after the end of OPT (documented allergic reaction). Supervision was provided by the doctor on duty. The cumulative dose administered at the time of onset of allergic events was noted as well as the nature of the allergic manifestations, the time of appearance, the treatment administered, and the evolution were noted. If there is no allergic reaction, the OPT is negative.

Results

Among 65 children referred for suspected BL HS, 56 drug-induced OPT were indicated. Nine children were not included in the study because the diagnosis of HS to BL was ruled out by the accidental reintroduction of a BL. Seven children were lost to follow-up after the first consultation and five children who consulted for suspected HS had an incomplete allergology workup.

Thirty-five OPTs with at least one BL were performed in the hospital setting after the consent of the guardian (Figure 1).

Figure 1.

Flow chart of patients.

We included 14 girls and 21 boys with a sex ratio of 1.5. The mean age was 7 years with extremes ranging from 4 to 12 years. Nine patients had personal atopy. Four cases had atopic dermatitis, two involved asthma, and three involved allergic rhinitis. In 10 cases, familial atopy was noted.

All patients in our series had non-immediate initial reactions with clinical manifestations developing in more than 1 h, without any signs of severity.

The average delay between the onset of allergic manifestations and the consultation in allergology was 3 years (6 months–8 years). The main clinical manifestations were dominated by pruritic maculopapular eruptions reported in 21 cases and delayed urticaria in nine cases. This urticaria was associated with angioedema in three cases.

BL has been prescribed for pediatric infectious diseases: acute otitis media in 12 cases, community-acquired pneumonia infections in nine cases, and acute bacterial bronchitis in 14 cases. BL was the only treatment prescribed in 21 cases, combined with anti-inflammatories and analgesics in 14 cases. These medications were discarded following spontaneous reintroduction by the patient without incident.

According to the Ring and Messmer classification, all reported clinical manifestations were grade Ⅰ and Ⅱ. No cases of grade III or IV anaphylaxis were noted in our study (Ring and Messmer, 1977).

A pharmacovigilance consultation had been held before the OPT. The study of the imputability of the occurrence of the allergic reaction was established according to the recommendations and the imputability score (Bégaud et al., 1985) and revealed a score of I2B3 in six cases (plausible) and a score of I1B3 (doubtful) in 29 cases (Table 4). Thus, in most cases (97%), allergology investigations were suggested. The oral challenge test was performed as the first-line test in all cases. The mean time from OPT to the HS reaction was 4 years [9 months–8 years]. Alternative BL OPT was indicated as a second-line test after the first positive BL OPT. In 23 (66%) cases, OPT was negative (Table 5).

Table 4.

Distribution of the imputability score according to the incriminated antibiotics.

	Plausible (I2B3) N = 6	Doubtful (I1B3) N = 29	Total
Amoxicillin	3 cas	11 cas	13 cas
Amoxiclav	3 cas	7 cas	10 cas
Cefixime	0 cas	6 cas	6 cas
BL not specified	0 cas	5 cas	5 cas

Table 5.

Results of first-line OPT.

OPT (+): HS to BL confirmed	12 cases (34%)
- Amoxicillin	5 cases
- Amoxiclav	5 cases
- Cefixime	2 cases

First-line OPT

In 12 cases (34.2%), the OPT confirmed a BL allergy. It was an allergy to penicillin in eight cases: OPT positive to amoxicillin (n = 5) and OPT positive to amoxicillin-clavulanic acid (n = 5) and allergy to cephalosporins in two cases (Table 6).

Table 6.

Results of second-line OPT.

MOLECULE	Number of tests	TPO (+)
- Amoxicillin	2	1
- Cefixime	10	1

The clinical manifestations were in the majority of cases of the cutaneous type made of erythema (n = 3), cutaneous rash (n = 2), urticaria (n = 4), and maculopapular lesions (n = 1). In three cases, fever, dyspnea, and digestive manifestations were noted (Table 7).

Table 7.

Clinical manifestations and time of their appearance according to the tested molecule.

Case/drug	Reaction time/last dose (mn)	Cumulative discontinuation dose (mg)	Initial reaction	Type of reaction
12 years amoxicillin	20	165	Primary maculopapular eruptions	Erythema + + fever
5 years amoxicillin	10	415	Generalized urticaria	Angioedema
8 years amoxiclav	120	1370	Urticaria + fever	Urticaria
7 years amoxiclav	210	960	Urticaria and eyelid edema	Edema of the face + urticaria
5 years amoxiclav	20	310	Skin rash	Urticaria + conjunctival hyperemia
6 years amoxiclav	25	710	Pruritic skin rash	Erythema + maculopapular rash
4 cefixime	480	212	Generalized erythema and facial edema	Urticaria
6 years amoxicillin	360	1500	Skin rash, fever, and vomiting	Skin rash and fever
10 years Amoxiclav	180	1500	Urticaire généralisé	Generalized hives + vomiting
5 years cefixime	60	200	Diffuse maculo-pauve lesions	Pruritic maculo-papular eruption over the whole body
10 years Amixicillin + cefixime	15 (amoxicillin) 75	150 60	Generalized erythema and facial edema	Urticaria (amoxicillin) urticaria (cefixime)
11 years Amixicillin + cefixime	30 (amoxi) 60 (cefixime)	1000 400	Urticaria and dyspnea	Erythema and dyspnea (amoxicillin) Skin rash (cefixime)

Second-line OPT

OPT was performed to look for a therapeutic alternative and penicillin-cephalosporin cross-reactivity in 12 children. Penicillin/cephalosporin cross-reactivity was noted in two cases (Table 6).

The average time of onset of the allergic manifestations about the last administered dose was 119 min.

Discussion

Almost 10% of parents report a suspected allergic reaction to antibiotics in their children (Erkoçoğlu et al., 2013). However, only a few can be confirmed after an allergology workup. Thus, it is difficult to determine with certainty the epidemiology of drug HS.

In children, the majority of suspected allergic reactions are the result of infectious and/or inflammatory diseases. These are often benign skin rashes. The diagnostic process to confirm or rule out this allergy is time-consuming.

Questioning is the first step of the allergology investigation. It includes a precise description of the symptomatology, the patient’s history (previous allergic incidents), the onset of symptoms (previous contacts, time of appearance after the last intake), other drugs taken (at the time of the reaction), and drugs of the same class. Skin tests (prick tests, intradermal tests, and patch tests) are usually the next step in the allergic investigation, but their reliability varies according to the classes of antibiotics (Blanca et al., 2009). Skin tests with classical penicillin reagents which are the major and minor determinants of benzylpenicillin, lose their sensitivity especially since they are not available and marketed everywhere.

In many cases the skin tests are negative despite a very suggestive clinical history, this may be explained by a selective immunological response to a side chain epitope requiring the selection of an alternative BL (Richter et al., 2013).

OPT has its place as a sensitive diagnostic tool for allergic reactions with immediate or delayed HS without severe systemic reaction (Romano et al., 2004).

In several studies, OPT to BL was positive despite negative skin tests, thus proving HS to the drug (Romano et al., 1995; Romano et al., 2008).

Romano et al. reported that only three (2.5%) of 117 subjects reacted to OPT with the suspect penicillin. Whereas in the study by Trcka et al., three (3.8%) of 78 subjects who underwent OPT with BL reacted with exanthematic rashes (Romano et al., 2002; Trcka et al., 2007). However, these studies were conducted on adults. OPT is the key to diagnosing cross-allergy. Indeed, many individuals allergic to cephalosporins develop a reaction to aminopenicillins but have a good tolerance to benzylpenicillin; reagents used in skin tests, responsible for false negatives during the allergology investigation (Blanca et al., 2009). OPT can expose this tolerance. In our study, we confirmed a cross-allergy to penicillin-cephalosporin in two cases (5.7%). Over the past decade, several investigators have performed first-line OPT without skin testing in subjects, especially children, with suspected BL allergy (García Rodríguez et al., 2019; Johansson et al., 2001; Labrosse et al., 2018; Mill et al., 2016; Moral et al., 2011). If the child’s clinical history suggests a low risk of allergy, and the initial suspicion is weak, some experts may consider skipping skin testing and proceeding directly to oral provocation. However, this should be done with caution and expertise.

This approach has also been validated since 2015 by the EAACI DAIG pediatric working group (Gomes et al., 2016) and the British Society for Allergy and Clinical Immunology for use in children (Mirakian et al., 2015).

In our study, all infants had a first-line OPT without a prior skin test. OPT was performed in a day hospital by a trained team in a specialized department. No patient reported a severe HS reaction. The oral challenge test is the gold standard for the diagnosis of BL allergy and the best diagnostic tool for benign skin rashes in children (Anonymous).

However, OPT protocols vary considerably in terms of dose escalation, time intervals between incremental doses, and days of administration, as well as the diagnostic criteria for judging a positive result. Several factors contribute to these differences: the timing of the reaction (immediate vs non-immediate reactions), the severity of the reaction (anaphylaxis vs urticaria/maculopapular exanthema), the population involved (children vs adults), the experience of the operators, and the resources available.

There is no standardization of dosages and the number of different doses. Several approaches exist.

Some use a 3-step OPT: for example, 1%-10%-100% (27, 28) or 25%-25%-50% (26) to 2 tiers: 10%-100% (Romano et al., 2012) or to a single dose 100% (Macy and Ngor, 2013) of the maximum unit dose.

Some authors propose protocols that extend over several days. Vezir et al. proposed that the offending drug be administered in divided doses every 30 min, for a total of five doses on the first day, and continuous full-dose home dosing for 5 days. The results showed that a total of four patients (3.4%) developed an urticarial rash after OPT. Three patients developed reactions on the first day, and one patient developed a reaction on the fourth day of DTP (Vezir et al., 2016).

Mill et al. propose a two-phase test. Authors challenged amoxicillin with a two-stage therapeutic dose: 10% of the therapeutic dose, then 20 min later, 90% of the therapeutic dose. Koosakulchai et al. opted for a 2-dose OPT: 10% followed by 30 min of observation and then 90% residual until the daily therapeutic dose was reached (Koosakulchai et al., 2021). Two-dose protocols have the advantage of being less expensive, less time-consuming, and more applicable to children. Opinions also differ on dose titration. Fransonn S et al. proposed titrated or not titrated OPT according to the estimated risk of inducing a severe reaction (severity of the index reaction and the patient’s comorbidities) (Fransson et al., 2017).

Titrated OPT was performed in three steps (1%, 10%, and 100% of the maximum unit dose). No titrated OPT was performed by directly administering the maximum unit dose.

Another study identified trigger levels based on survival analysis and suggested the following steps: 5%, 15%, 30%, and 50% of the therapeutic dose at 30-min intervals. For index reactions of anaphylaxis, OPT was performed with additional lower doses of 0.01%, 0.1%, and 1% of the therapeutic dose (Chiriac et al., 2017). We opted for a 1-day OPT titration. The number of doses administered to reach the therapeutic dose varied between five and six doses. This was proposed according to the severity and age of the reaction and the tolerance of the administered doses.

To our knowledge, this is the first Tunisian study that has looked at BL OPT indicated in the first instance in children, carried out over 12 years in an allergology reference center. In our study, the percentage of children with a confirmed HS to BL is higher than that of the literature. This can be explained by a selection bias, as our center is the most renowned center in the field of allergology exploration. Generally, it carries patients from all over the country.

This leads us to talk about the main limitation of our study, which is the monocentric character. A calculation of a sample size was not carried. We cannot generalize our results to the Tunisian population. Another limitation is that over 12 years, we collected only 35 patients, which can be explained by the strict exclusion criteria, all severe reactions were excluded but also it can be explained by the lack of awareness of primary care physicians about the heavy label of an antibiotic allergy issued without diagnostic confirmation.

Conclusion

In children, we can immediately propose an OPT with the offending molecule if it was known or use a molecule that would be the most prescribed. This attitude could shorten the time needed for the allergology exploration proposed in advance while respecting the safety conditions (OPT performed in a hospital setting and after the consent of the guardian). Finally, it is necessary to sensitize primary care physicians and pediatricians to refer these children to a specialized consultation as soon as possible to avoid the abusive eviction of BL.

ORCID iD

Chirine Moussa https://orcid.org/0000-0002-8123-9843