• aminopenicillins;
  • children;
  • drug-provocation tests;
  • nonimmediate reactions;
  • skin testing


  1. Top of page
  2. Abstract
  3. Material and methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Sources of funding
  8. References

Background:  Nonimmediate allergic reactions (NIR) to aminopenicillin include several entities, the most common of which are urticaria-like and maculopapular exanthemas.

Aims of the study:  To evaluate a group of children who developed one or more episodes of skin reactions suggestive of NIR after aminopenicillin administration.

Methods:  The inclusion criteria required negative immediate skin tests and absence of specific IgE antibodies to different penicillins. Intradermal and patch tests were carried out with delayed readings and, if negative, a drug-provocation test including a full therapeutic course of the drug was given. Two different groups were compared: A) children with positive skin testing or a positive drug-provocation test and B) children with negative skin testing and good tolerance after a drug-provocation test.

Results:  Group A was composed of 20 patients. Positive intradermal/patch tests were found in one patient and in the remaining 19, a positive response to a drug-provocation test confirmed the diagnosis. Group B (the control group) consisted of 19 patients with similar symptoms after aminopenicillin intake but good tolerance. No differences in age, dose or number of previous treatments were observed between the groups. The clinical entities were also similar in both groups.

Conclusions:  Reproducible nonimmediate skin reactions to aminopenicillins may occur in children in spite of negative skin testing. The value of this diagnostic procedure seems to be limited in this type of reaction, with drug-provocation tests (DPT) being a reasonable and safe alternative if the diagnosis has to be confirmed.


nonimmediate reactions


immediate reactions


benzylpenicilloyl determinant


minor determinant mixture


drug-provocation tests




European Network for Drug Allergy

Nonimmediate (NIR) reactions to betalactams occur more than 1 h after drug administration, and most are T-cell mediated (1). Although considerable interest has been evinced on immediate reactions (IR), the same cannot be stated for NIR (2). The entities involved are greater than that for IR and includes maculopapular exanthema, urticaria-like exanthema, serum-sickness-like syndrome, the Steven Johnson-Toxic epidermal necrolysis syndrome complex, acute exanthematic pustulosis and drug hypersensitivity syndrome, in addition to organ-specific reactions (3). Exanthematic reactions were initially reported with ampicillin therapy and usually associated with viral infections, with the subsequent replacement by other aminopenicillins, such as amoxicillin and amoxicillin-clavulanic acid, which are the most frequently involved (4). Although a high proportion of subjects with exanthematic reactions to aminopenicillins show good tolerance (5–7), a few develop further reactions after re-exposure, and are therefore considered as allergic (4, 7). Although reports suggest that the incidence of IR in children appears to be lower than in adults (1, 8, 9) no information is available for NIR.

We studied a group of children aged 1.5–12 years referred to the allergy unit of Gregorio Marañon children’s hospital to rule out allergy to aminopenicillins. After the allergological study, we found that intradermal/patch testing to different penicillin determinants seemed to be of limited value to diagnose children with NIR, and a drug-provocation tests (DPT) was necessary.

Material and methods

  1. Top of page
  2. Abstract
  3. Material and methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Sources of funding
  8. References

Patient selection

Children referred for evaluation of a NIR to aminopenicillin were included. Reactions were maculopapular or urticaria-like exanthema, whether or not accompanied by osteoarticular symptoms (Serum sickness-like syndrome). Symptoms occurred at least several hours after first drug intake and usually within an interval of 24–48 h (2).

The allergollogical work-up included skin testing and, if negative, a DPT. Children who developed a positive response, with skin testing or DPT, were diagnosed as having a NIR (Group A). We compared this group with children with similar symptoms after aminopenicillin intake but having a negative skin test and DPT (Group B).

The institutional review boards approved the study, and informed consent for the diagnostic procedures was obtained from the parents of all the patients.

Skin testing

Intradermal testing was carried out as described (2), using penicilloyl-polylisine (PPL) at 5 × 10−5 mmol/l, minor determinant mixture (MDM) at 2 × 10−2 mmol/l (Diater, Madrid, Spain), benzylpenicillin (Normon, Madrid, Spain) at 10 000 IU/ml and amoxicillin (GSK, Madrid, Spain) at 20 mg/ml. Readings were made at 15–20 min, to exclude an immediate reaction, and at 24, 48 and 72 h. A wheal area with a diameter >5 mm was considered a positive reaction.

Patch testing was carried out with amoxicillin in petrolatum at 5% w/w, as described (2). The occlusion time was 48 h and readings were made at 24, 48 and 96 h. At 24 h, the occlusion was opened for the evaluation and then reapplied.

Specific IgE determination

In order to exclude an IgE-dependent reaction, specific IgE to benzylpenicillin and amoxicillin was measured by Phadia CAP system (Uppsala, Sweden), as reported (10).

Drug-provocation test

In the event of a negative skin test, a DPT was made as described (11), but with slight modifications to adapt the protocols to the ages and body weights of the children. Briefly, DPT was done with oral amoxicillin, one quarter of the dose followed by three quarters 1 h later, with patients kept under observation for 6 h, after which they continued taking the drug at home for 5 days twice a day under ambulatory surveillance.

Statistical studies

Comparisons of age, time interval and number of episodes occurring in both groups were made by t-test and the gender and type of reaction were compared by Chi square.


  1. Top of page
  2. Abstract
  3. Material and methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Sources of funding
  8. References

A total of 20 children (7 girls and 13 boys) were finally diagnosed as having a NIR (Group A) (Table 1). The mean age was 4.9 (1.5–13) years, and the total number of episodes was 37 (six children had one episode, 11 had two and three children had three episodes). Amoxicillin was involved in 18 and amoxicillin-clavulanic acid in 19 episodes. Urticaria-like exanthema appeared in nine cases of patients, maculopapular exanthema in four and in seven, osteoarticular symptoms occurred in addition to skin involvement. The time between the occurrence of the previous reaction and the allergological evaluation was 13.8 months (1–28). The mean time between first drug intake and the appearance of symptoms was 5.8 days (range: 2 to 10 days). In five cases of patients, both amoxicillin and amoxicillin-clavulanic acid were involved.

Table 1.   Clinical characteristics in children who developed a positive allergological work-up (Group A)
CaseGenderAge (years)ReactionDrugInterval (drug-reaction) (days)Interval (reaction-study) (months)Episodes
  1. AX, amoxicillin; AX-CLAV, amoxicillin-clavulanic.

 1Male3Serum sickness-like syndromeAX7/4112
 2Male7Urticaria-like exanthemaAX-CLAV7241
 3Female3Urticaria-like exanthemaAX-CLAV7/7/7123
 4Male5Serum sickness-like syndromeAX/AX-CLAV7181
 5Female5Serum sickness-like syndromeAX-CLAV7241
 6Male2Serum sickness-like syndromeAX10/1062
 7Female9Maculopapular exanthemaAX7121
 8Male2Serum sickness-like syndromeAX-CLAV/AX8/8122
 9Female5.5Urticaria-like exanthemaAX5/5182
10Male3.5Urticaria-like exanthemaAX-CLAV3/3182
11Male9Urticaria-like exanthemaAX-CLAV711
12Male8Urticaria-like exanthemaAX-CLAV7/7282
13Female6Serum sickness-like syndromeAX7/7182
14Female4Urticaria-like exanthemaAX6121
15Male3Urticaria-like exanthemaAX (2)/AX-CLAV8/7/793
16Female3Serum sickness-like syndromeAX-CLAV/AX7/4172
17Male4Maculopapular exanthemaAX/AX-CLAV2/5122
18Male13Urticaria-like exanthemaAX-CLAV7121
19Male3Maculopapular exanthemaAX-CLAV5/662
20Male1.5Maculopapular exanthemaAX2/262

Table 2 shows the results of the allergological work-up in Group A. Both patch and intradermal testing were positive in just one patient (case 20), and a DPT was required to establish the diagnosis in the others. Amoxicillin was the drug inducing the response in all DPT, including those where the association amoxicillin-clavulanic acid was the culprit drug. The mean time interval between administration and reaction was 5.78 (1–8) days. In most instances, except cases 3, 10, 13, 16 and 19, the reactions occurred after the first dose of the day. Figure 1 shows the typical findings in three children who had a positive response.

Table 2.   Results of skin testing and drug-provocation tests in children who developed a positive allergological work-up (Group A)
CaseSkin testsDPTDrugD/CD (mg/g)Interval (drug-reaction) (days)Symptoms
  1. n.d., not done; AX, amoxicillin; D, dose given b.i.d. in mg and CD cumulative dose indicating the total amount of drug given before the appearance of the reaction.

 1(−)(+)AX250 (3.75)8Generalized exanthema and itching
 2(−)(+)AX250 (3.25)7Itching on the scalp followed by maculopapular exanthema
 3(−)(+)AX125 (1)4Erythema and wheals on shoulders, chest and abdomen.
 4(−)(+)AX250 (2.25)5Generalized pruritus and wheals on chest and back.
 5(−)(+)AX250 (3.25)7Pruritus on the neck and retroauricular regions followed by exanthema in the same areas plus lower abdomen.
 6(−)(+)AX250 (3.25)7Generalized confluent erythematous wheals with pruritus
 7(−)(+)AX500 (6.5)7Generalized exanthema and itching
 8(−)(+)AX250 (2.25)5Generalized confluent erythematous wheals with pruritus plus fever and swelling of elbows and knees.
 9(−)(+)AX250 (2.75)6Erythema on neck, back and lower abdomen plus swelling and redness of knees and elbows
10(−)(+)AX125 (0.75)3Pruritus and wheals on the face, lower extremities and waist plus redness on elbows and knees.
11(−)(+)AX500 (6.5)7Pruriginous confluent wheals on lower extremities.
12(−)(+)AX250 (3.25)7Wheals on the back plus erythema on neck and retroauricular area.
13(−)(+)AX250 (3.5)7Eyelid edema plus generalized exanthema followed by joint swelling and fever.
14(−)(+)AX250 (3.5)6Generalized wheals and itching.
15(−)(+)AX125 (2.75)7Generalized exanthema and itching.
16(−)(+)AX250 (2)4Eyelid edema plus generalized wheals followed by joint swelling and fever.
17(−)(+)AX250 (2.25)5Maculopapular exanthema on lower extremities followed by joint swelling and edema
18(−)(+)AX500 (6.5)7Generalized itching and wheals on abdomen
19(−)(+)AX125 (0.25)1Erythema and pruritus on abdomen and armpits

Figure 1.  Examples of three children (cases 4, 13 and 17 from Group A) who developed a positive response after the drug provocation test.

Download figure to PowerPoint

Group B consisted of 19 patients, 13 girls and six boys (Table 3). Their mean age was 5.63 (1.5–12) years. These patients experienced a total of 30 episodes: 11 children had one episode each, six children had two episodes each, one had three and yet another child had four episodes. Amoxicillin was involved in 16 cases and amoxicillin-clavulanic acid in 14. Urticaria-like exanthema appeared in five cases, maculopapular exanthema in seven and in another seven episodes, osteoarticular symptoms occurred in addition to skin involvement. The time between the occurrence of the last reaction and the allergological evaluation was 13.21 (2–36) months. The mean time between first drug intake and the appearance of symptoms was 5.6 days, range: 1–10 days. In two children who had two different episodes, different drugs were involved (amoxicillin and amoxicillin-clavulanic acid). All patients in Group B had a negative allergological work-up. No significant differences were found in age, time interval between drug intake and appearance of symptoms, or the number of episodes. Nor were the distribution by gender and the clinical entities in each group significantly different.

Table 3.   Clinical characteristics in children with a negative allergollogical work-up (Group B)
CaseGenderAge (years)ReactionDrugInterval (drug-reaction) (days)Interval (reaction-study) (months)Episodes
  1. AX, amoxicillin; AX-CLAV, amoxicillin-clavulanic.

 1Female2Serum sickness-like syndromeAX-CLAV7111
 2Female3Serum sickness-like syndromeAX7/7122
 3Female5Serum sickness-like syndromeAX-CLAV/AX10/10112
 4Female1.5Maculopapular exanthemaAX641
 5Female12Maculopapular exanthemaAX-CLAV6121
 6Female12Delayed urticariaAX-CLAV7121
 7Male3.5Serum sickness-like syndromeAX-CLAV4122
 8Male6Delayed urticariaAX6/6/6/6364
 9Male4Maculopapular exanthemaAX-CLAV /AX1/1362
10Female2.5Maculopapular exanthemaAX-CLAV6101
11Female11Serum sickness-like syndromeAX-CLAV7241
12Female4Delayed urticariaAX5/591
13Female7Delayed urticariaAX-CLAV721
14Female3Serum sickness-like syndromeAX-CLAV7/5222
15Male10Maculopapular exanthemaAX731
16Female5Maculopapular exanthemaAX-CLAV841
17Male1.5Maculopapular exanthemaAX-CLAV2/172
18Female11Delayed urticariaAX5121
19Male3Serum sickness-like syndromeAX5/5/5123


  1. Top of page
  2. Abstract
  3. Material and methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Sources of funding
  8. References

Exanthematic reactions during the course of aminopenicillin treatment are a common phenomenon, with maculopapular exanthema reported in 68% of patients and urticaria in 32% (5). Although these reactions are usually considered nonallergic in origin, a number of studies have shown that they can be allergic and induced by T-cells (10).

Terrados et al. reported that 65% of adults with NIR to penicillins had a positive skin test (12). Evidence for allergic reactions is less common in children and data are mainly derived from IR. Mendelson et al. studied 240 children thought to be allergic to beta-lactams, 1.4% of children with negative skin tests experienced maculopapular rash or urticaria, with an interval of up to 10 days after drug administration (13). Romano found that delayed skin testing was positive in 4.9% of cases (6).

In a prospective study in NIR, Padial et al. found that only 9% of subjects had a positive intradermal test (11). Our results in children support these findings, as only one of the 20 children had a positive patch and intradermal test and the rest required a DPT to establish the diagnosis.

Exanthematic reactions in children may be IgE positive (6, 13), in our study, no case with an immediate reaction was reported.

Our study does not provide the incidence of NIR in children with a suggestive history as we did not estimate the total number of patients evaluated at our clinic and only used a well-defined positive group and a control group for comparative purposes.

As a general rule, the negative predictive value for skin testing, including subjects with skin rashes, is greater than 99% (14). However, a very high proportion of the children with NIR, in our study, developed a positive response after a DPT in the absence of positive skin tests. To confirm these observations, further studies need to be carried out calculating the total number of patients with suggestive symptoms in comparison with the final number of cases confirmed as positive.

The association amoxicillin-clavulanic acid was involved in over 50% of the subjects included in our study. Similar results have been reported by others (15, 16). In all cases in our study, when a subject developed a positive response, this was in response to the administration of amoxicillin alone. The capacity of amoxicillin-clavulanic acid to induce NIR seems to be limited and most instances concern isolated case reports (16, 17).

The precise history of the nature of the process leading to drug administration was not available in either the group with confirmed allergy or in the group with a suggestive history and a negative allergological work-up. Furthermore, eight cases in the negative group (Table 3) had two or more episodes. Whether subsequent exposure can induce the reactions is something that we are unable to answer at this stage, as the same and requires further studies.

In summary, maculopapular or urticaria-like exanthematic reactions to aminopenicillins may occur in children; the utility of skin testing being limited, DPT is required to establish the diagnosis. This procedure seems to be safe providing that symptoms fall within the criteria specified above.


  1. Top of page
  2. Abstract
  3. Material and methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Sources of funding
  8. References

We thank Ian Johnstone for help with the English language version of the manuscript.


  1. Top of page
  2. Abstract
  3. Material and methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Sources of funding
  8. References