Clinical evaluation of Pharmacia CAP System RAST FEIA amoxicilloyl and benzylpenicilloyl in patients with penicillin allergy

Authors


Miguel Blanca
Invesigation Laboratory
Hospital Civil
29009 Malaga
Spain

Abstract

Background: The diagnosis of IgE-mediated immediate reactions to penicillins can be supported by in vivo or in vitro tests using classical benzylpenicillin determinants. The wide variety of β-lactams and the description of new specificities requires a re-evaluation of the different tests available. The objective was to evaluate the diagnostic capacity of Pharmacia CAP System RAST FEIA amoxicilloyl c6 (AXO) and benzylpenicilloyl c1 (BPO) in patients with a documented IgE-mediated penicillin allergy.

Methods: We studied 129 patients in five groups. Groups 1, 2, and 3 had developed an immediate reaction after penicillin treatment. Group 1 (n=19) were skin test positive to amoxicillin (AX) and/or BPO and/or minor determinant mixture (MDM); group 2 (n=29) were skin test positive to AX but negative to BPO and MDM; and group 3 (n=26) were skin test negative to all determinants, the diagnosis being confirmed by a previous repetitive history or controlled administration. Two control groups, one with nonimmediate reactions – group 4 (n=25) – and one with good tolerance to penicillin – group 5 (n=30) – were included. All samples were analyzed in vitro for AXO and BPO, and the results compared to the in vivo diagnosis.

Results: AX was the drug most often involved. In group 1, 53% were in vitro positive for AXO and 68% for BPO, but 74% had at least one positive test result. In group 2, only 10% had a positive in vitro test to BPO compared to 41% to AXO. In group 3, 42% had positive BPO and/or AXO in vitro tests. In the control groups 4 and 5, the negative in vitro results for AXO were 96% and 100%, and for BPO 100% and 97%, respectively. A positive correlation between specific IgE levels and the time interval from the reaction to the evaluation was found only for group 3.

Conclusions: This in vitro assay is beneficial for evaluating subjects allergic to β-lactams. It is necessary to test for specific IgE to AXO in addition to BPO in patients with immediate allergic reactions after AX. The combination of in vivo and in vitro tests for estimating IgE antibodies to penicillins is important because of the existence of patients with a positive history but negative skin test.

Allergic reactions to penicillin are the most common cause of immunologic drug reactions. Both IgE-mediated and T-cell-dependent responses have been reported after the administration of β-lactams (1), the IgE-mediated reactions being the most frequent and best characterized (2). Subjects develop IgE antibodies which recognize the structure formed by a combination of the penicillin molecule plus a protein carrier (3). In studies in both man and animals, penicilloyl has long been recognized as the major antigenic determinant (4). Soon after the introduction of benzylpenicillin (BP), a number of other β-lactams were developed, resulting in the presence of a wide number of chemical structures that are quite different from those initially used (5). This means that populations are exposed to various haptens that can be recognized by the immunologic system (2, 3). Furthermore, the influence of these chemical structures varies over time because of the different patterns of consumption. BP, the drug that generates the major determinant benzylpenicilloyl (BPO), is now no longer the most often used β-lactam. These changes have influenced not only the specificity of antibody (3, 6) but also the pattern of allergic responses (7).

The diagnostic methods for evaluating subjects with an immediate allergic reaction include an adequate clinical history plus the performance of in vivo and/or in vitro tests (2). Soon after the description of allergic reactions to penicillin, the development of a penicillin-polylysine conjugate helped contribute to the diagnosis of these cases (9). Later, the addition of the so-called minor determinants (MDM) improved the skin testing (2, 8, 9). However, there is limited access to MDM, since they have not been licensed in all countries and they can only be used for research purposes (8–10); therefore, the pattern of responses in different populations cannot always be determined. Recently, several groups have reported an increase in cases of anaphylaxis over urticaria, so that skin testing in these patients may involve additional risk (7, 11–13).

Another approach in evaluating allergic reactions to penicillins is in vitro quantitation of specific IgE antibodies to the culprit drug (14, 15). These specific IgE antibodies were detected in a population of allergic patients taking penicillin V in 1979 (16). Additional reports of selective reactions to other penicillins and cephalosporins have since been published (17–21). The most relevant and best-studied selective response and further evidence of blood IgE-antibody-positive results but negative skin tests has been observed in Spain with amoxycillin (AX) (11, 22, 23).

The objective of the study was to evaluate the diagnostic advantages of in vitro testing for specific IgE antibodies with the Pharmacia CAP System, in a group of patients allergic to penicillins according to clinical history and in vivo tests. Results indicate that in vitro testing is beneficial, especially in patients positive in clinical history but skin test negative.

Material and methods

Patient and control selection

Subjects who developed an immediate reaction after the administration of a penicillin derivative were included in the study. Two clinical entities were defined: anaphylaxis and urticaria. The criterion for considering a reaction as anaphylaxis was the presence of several of the following symptoms: systemic pruritus, generalized erythema, angioedema in one or more parts of the body, dysphonia, difficulty in breathing, abdominal cramps, and hypotension. The criterion for urticaria was whether the immediate manifestations were limited exclusively to the skin and consisted of maculopapulae at various sites of the body, whether or not associated with angioedema.

Three groups of patients were considered, based on the skin test reactivity to the penicillin determinants:

Group 1

This group comprised subjects with an immediate reaction to BP and/or AX. The skin test had to be positive to BPO and AX independently of positivity to the other determinants MDM and ampicillin (AMP). This group comprised 19 subjects, six women (32%) and 13 men (68%). The mean age was 47.5 years, range: 19–69 (Tables 1 and 2).

Table 1.  Clinical characteristics of subjects who responded to penicillin and AX (groups 1 and 2)
 Patient
no.
Age
(years)
SexDiagnosisDrug involvedInterval
(days)
Group 1132MAnaphylaxisAX550
 255MAnaphylaxisAX128
 334FUrticariaAX73
 449MAnaphylaxisAX15
 543FUrticariaAX92
 637MAnaphylaxisAX80
 737MAnaphylaxisAX324
 859MAnaphylaxisAX69
 961MAnaphylaxisPG746
 1052FAnaphylaxisAX716
 1154MAnaphylaxisAX3
 1247FAnaphylaxisAX5
 1364MAnaphylaxisAX156
 1450MAnaphylaxisAX52
 1565MAnaphylaxisAX40
 1669MAnaphylaxisAX5
 1736MUrticariaAX102
 1819FAnaphylaxisAX57
 1940FAnaphylaxisAX7
Group 22032MAnaphylaxisAX788
 2154FAnaphylaxisAX11
 2221FUrticariaAX134
 2331FUrticariaAX22
 2436FAnaphylaxisAX plus clavulanic acid20
 2531FUrticariaAX348
 2621MUrticariaAX454
 2737FUrticariaAX212
 2829FAnaphylaxisAX58
 2940FAnaphylaxisAX15
 3039FUrticariaAX plus BP (benzathine)99
 3140FUrticariaAX73
 3245MUrticariaAX533
 3351FAnaphylaxisAX2
 3433FAnaphylaxisAX27
 3533FAnaphylaxisAX57
 3630MAnaphylaxisAX399
 3733MAnaphylaxisAX52
 3827FAnaphylaxisAX681
 3946MAnaphylaxisAX128
 4032MAnaphylaxisAX189
 4144MAnaphylaxisAX8
 4235FAnaphylaxisAX4
 4335FUrticariaAX25
 4431FAnaphylaxisAX28
 4524MAnaphylaxisAX10
 4637MAnaphylaxisAX16
 4730MAnaphylaxisAX98
 4842MUrticariaAX155
Table 2.  Skin test and serum IgE antibody results in patients from groups 1 and 2
 Patient
no.
Skin testSerum IgE antibodies
(kUA/l)
BPOMDMAXAMPBPOAXO
Group 11++++  <0.35  <0.35
 2++++  <0.35  <0.35
 3+++  <0.35  <0.35
 4+++  <0.35  <0.35
 5++++  <0.353.6
 6++++  <0.35  <0.35
 7++++0.5  <0.35
 8++++0.5  <0.35
 9++++1.4  <0.35
 10+++1.61.1
 11++++2.63.1
 12++2.83.7
 13++++9.17.8
 14++9.911.1
 15++++16.8  <0.35
 16+++17.43.4
 17+++21.733.7
 18++++36.423.2
 19++++>100>100
Group 220++  <0.35  <0.35
 21+  <0.35  <0.35
 22+  <0.35  <0.35
 23+  <0.35  <0.35
 24++  <0.35  <0.35
 25++  <0.35  <0.35
 26++  <0.35  <0.35
 27+  <0.35  <0.35
 28+  <0.35  <0.35
 29++  <0.35  <0.35
 30+  <0.35  <0.35
 31++  <0.35  <0.35
 32++  <0.35  <0.35
 33++  <0.35  <0.35
 34+  <0.35  <0.35
 35++  <0.35  <0.35
 36++  <0.35  <0.35
 37+  <0.350.8
 38++  <0.350.8
 39+  <0.350.9
 40++  <0.351.1
 41+0.61.8
 42+  <0.352.0
 43++  <0.352.2
 44++  <0.352.5
 45+  <0.352.7
 46++  <0.352.9
 47+2.12.9
 48++0.45.8

Group 2

This group comprised subjects who, after having an immediate reaction to an AX derivative, were skin test positive to AX determinants and negative to BPO and MDM, and had good tolerance to BP. Details of the characterization of this group are reported elsewhere (11, 12). There were 29 subjects, 17 (59%) women and 12 (41%) men. The mean age was 35.1 years, range: 21–54 (Tables 1 and 2).

Group 3

This group comprised subjects with an immediate reaction to penicillin or AX who were skin test negative to all penicillin derivatives used in the study. These included:

1)patients who had experienced at least two or more episodes of immediate response to a penicillin derivative

2)patients who developed systemic symptoms compatible with anaphylaxis after skin testing in spite of being skin test negative

3)patients who had a positive challenge after penicillin administration, and who were skin test negative.

A total of 26 patients, 14 women (53.8%) and 12 men (46.2%), were included in this group. The mean age was 43.8 years, range: 23–64 (Table 3).

Table 3.  Clinical characteristics and drug involved in patients with negative skin test to penicillin determinants and results of in vitro assay to serum IgE antibodies to BPO and AXO in patients from group 3
Patient
no.
Age
(years)
SexDiagnosisDrug
involved
Interval
(days)
Serum IgE antibodies (kUA/l)
BPOAXO
4946FAnaphylaxisAX65<0.350.5
5047MAnaphylaxisAX143<0.35<0.35
5163FUrticariaAX545<0.35<0.35
5232FUrticariaBP587<0.350.5
5341MUrticariaAX95<0.35<0.35
5443MAnaphylaxisAX55<0.35<0.35
5550FAnaphylaxisCLX55<0.35<0.35
5644MUrticariaAX92<0.35<0.35
5749MAnaphylaxisBP53<0.350.4
5838MAnaphylaxisBP5<0.35<0.35
5935FAnaphylaxisAX42<0.35<0.35
6056FAnaphylaxisAX95<0.35<0.35
6130FAnaphylaxisAX32<0.35<0.35
6244FUrticariaAX67<0.35<0.35
6359FAnaphylaxisBP113<0.35<0.35
6425MUrticariaAX22<0.35<0.35
6543MUrticariaAX62<0.35<0.35
6632FAnaphylaxisAX116<0.35<0.35
6764MAnaphylaxisAX290.5<0.35
6823FUrticariaAMP6190.70.4
6934FUrticariaAX2061.11.1
7038FUrticariaAX5591.12.7
7163FUrticariaAMP5441.20.7
7253MAnaphylaxisAX56181.41.8
7338MUrticariaAX9402.10.9
7450MAnaphylaxisBP69113.32.9

Two control groups were included:

Group 4

This group comprised patients with a clinically documented non-IgE-mediated reaction to penicillin. Subjects who developed maculopapular or exanthemic reactions with an interval greater than 6 h and usually within 24–48 h after taking the drug were included in this group. In all these cases, confirmation of penicillin allergy was established by controlled administration. Immediate skin tests to BPO, MDM, AX, and AMP had to be negative. The group included 22 (88%) women and three (12%) men. The mean age was 40.0 years, range: 17–71.

Group 5

This group comprised 30 subjects with no history of allergic reaction to β-lactams; a negative skin test to BPO, MDM, AX, and AMP; and good tolerance to BP and AX. There were 18 (60%) women and 12 (40%) men. The mean age was 39.7 years, range: 14–65.

Skin tests

These were carried out as previously reported (11, 12), with 0.02 ml of solution. The reagents were benzylpenicilloyl polylysine (BPO) (Allergopharma Merck, Reinbek, Germany) (5×10−5 M), MDM (Allergopharma) (2×10−2 M), AX (SB SmithKline Beecham, Madrid, Spain) (20 mg/ml), and AMP (Antibiotic, León, Spain) (20 mg/ml). The concentrations shown were the maximum used. Skin tests to BPO and MDM were made on one day and to AX and AMP 1 week later. In order to determine whether systemic symptoms were induced by each particular hapten, we initiated the skin test the first day with BPO beginning with prick testing at the concentration of 5×10−5 M and then intradermal testing following the procedure described elsewhere (11, 12). In all cases, the concentration was a 10-fold lower dilution; if the test was negative, we proceeded to the next higher concentration. After 1 h, we proceeded to test the MDM in the same way. One week later, we proceeded to the skin testing with AX and AMP. We began with AMP at 20 mg/ml by prick, followed by intradermal administration at 2 mg/ml and then 20 mg/ml. We then performed the AX tests. The sequence was scheduled in this way because of the high incidence expected of selective responses to AX, thereby placing the most common hapten last for testing in order to complete the whole study. Responses were classified as positive or negative, as previously described (24). In the skin prick tests, a wheal larger than 2 mm with a negative response to the control saline was considered positive. In the intradermal tests, the wheal area was marked initially and 20 min after testing, and an increase in diameter greater than 3 mm was considered positive.

Sera collection

Samples were obtained from the patients at the moment they were evaluated in the clinical outpatient department before the skin test procedure, and were frozen at −20°C until use. All samples were run in parallel for the in vitro test.

Controlled administration

In order to establish the involvement of the penicillin in the induction of the reaction in skin-test-negative subjects, a controlled challenge was done in those in whom only one episode occurred. Those patients who developed systemic symptoms after skin testing were also considered positive and hence not given the drug. Basically, this was carried out as previously described (11, 12) with some modifications. In the first week, after completion of skin testing with BPO and MDM, we began with the parenteral administration of 1ml of BP (104IU/ml), followed by 1ml at 105IU/ml, and finishing with 1ml of 106IU/ml if good tolerance was established at the previous doses. Similarly, if skin tests with AMP and AX the next week were negative, AX was given by the oral route at the following doses: 5, 50, 100 and 500 mg with a 1-h interval between each.

Specific IgE antibody determination: Pharmacia CAP System

The Pharmacia CAP System FEIA is an in vitro test system, based on ImmunoCAP technology, for determination of circulating specific IgE antibodies. The drug of interest, covalently coupled to ImmunoCAP, reacts with the specific IgE in the patient's serum specimen. After nonspecific IgE is washed away, enzyme-labeled antibodies against IgE are added to form a complex. After incubation, unbound enzyme anti-IgE is washed away, and the bound complex is then incubated with a developing agent. When the reaction is stopped, the fluorescence of the elute is measured by FluoroCount 96. For classification of test results, the fluorescence of patient samples is compared directly by fluorescence for standards run in parallel. The pharmacia CAP System specific IgE measuring range is 0.35–100 kUA/l, with a cutoff value of ≥0.35 kUA/l for positive test results and <0.35 kUA/l for negative test results.

Statistical studies

For each patient group, a descriptive analysis of demographic characteristics was made, and the arithmetic mean, age range, and sex distribution were calculated. The frequency of positive and negative results for Pharmacia CAP System RAST FEIA c1 (BPO) and c6 (AXO) was also calculated.

The time interval means for groups 1, 2, and 3 were compared by the Levene test. The relationship of the time interval between the occurrence of the reaction and the sera collection and the specific IgE values was studied by Spearman rank correlation. These variables were transformed into categoric variables for chi-square analysis. Time interval and blood IgE antibody values were classified into two categories

1)for time interval, less or greater than 60 days

2)for the Pharmacia CAP System RAST FEIA c1 (BPO) and c6 (AXO) as positive (≥0.35 kUA/l) or negative (<0.35 kUA/l).

The sensitivity of the CAP test was calculated by clinical history, skin tests, and/or controlled administration as the reference methods in groups 1, 2, and 3, and the specificity was calculated in the negative controls (groups 4 and 5).

Results

Clinical characteristics of subjects

This was a study based on clinical and immunologic data of patients from a population that formed part of a large cohort of allergic patients in Spain. A total of 129 patients were included in the study, of whom 74 (groups 1, 2, and 3) developed an immediate reaction after penicillin treatment. Patients in groups 1 and 2 were skin test positive to at least one penicillin determinant, as described in Material and methods. Group 3 comprised 26 patients with a positive history but negative skin tests. Patients in group 4 had developed a nonimmediate reaction after the penicillin intake, and subjects in group 5 had shown good tolerance to penicillin and no symptoms after the drug intake. The most common drug involved in the adverse reactions in all five groups was AX (87.8%), followed by BP (8.1%), AMP (2.7%), and cloxacillin (1.3%).

The demographic and clinical characteristics of the patients were as follows.

Group 1

In 18 (95%) patients, AX was the drug involved, and in one (5%) it was BP. Anaphylaxis occurred in 16 (84%) patients and urticaria in three (16%). The mean interval between the occurrence of the reaction and sera collection was 136±44 days (Table 1).

Group 2

AX was involved in 27 cases (93%), AX plus clavulanic acid in one (3%), and AX-BP (benzathine) in one (3%). Anaphylaxis occurred in 19 (66%) patients and urticaria in 10 (34%). The mean interval between the occurrence of the reaction and sera collection for the study was 160±41 days (Table 1).

Group 3

AX was the drug involved in 18 patients (69%), BP in five (19%), AMP in two (8%), and cloxacillin in one (4%). Of these patients, 14 (54%) had anaphylaxis and 12 (46%) urticaria. The mean interval between the occurrence of the reaction and sera collection in this group was 440±214 days (Table 3).

Group 4

The drug most frequently involved was AX in 84% of patients, followed by BP in three (12%), and AMP in one (4%).

Group 5

This group comprised subjects who tolerated AX. The intake of β-lactams was AX (85%), BP (11%), and AMP (1%).

In vivo and in vitro studies

Patients in group 1 had positive skin test results to at least BPO and AX (Table 2). Twelve patients (63%) had a positive skin test to all penicillin determinants tested, and the remainder were positive to BPO and AX; of the latter, 17 were positive to AMP (89%) and 12 to MDM (63%). The immunoassay to BPO was positive in 13 (68%) patients while that to AXO was positive in 10 (53%) patients. The immunoassay was positive to AXO and negative to BPO in only one patient (5%), and four patients (21%) were positive only to BPO. In nine patients (47%), the immunoassay was positive to both haptens.

The skin test results of group 2 are shown in Table 2. Skin tests were positive to AX and negative to BPO and MDM; 16 (55%) of the 29 patients were also positive to AMP. With AX, eight (27%) patients were positive by prick test, and the remainder were positive by intradermal test. Two of the 16 patients positive to AMP were positive by prick test and the remainder by intradermal test. In all cases, skin tests were negative to BPO and MDM, and the skin response to AX was stronger than to AMP. The results of specific serum IgE antibodies are also presented in Table 2, showing 12 (41%) patients positive to AXO. Three (10%) patients were also positive to BPO.

Patients in group 3 had negative skin test results. The details of controlled administration and systemic symptoms after skin testing are presented in Table 4. Nine patients developed a clinical reaction on two or more occasions and therefore were considered positive (nos. 49, 50, 51, 61, 67, 71, 72, and 74 of Table 3). Seven of the 26 patients (27%) developed systemic symptoms after skin testing. Except for patients 58 and 73, who responded at the maximum dose of determinants used, in the remaining patients the concentration that induced the reaction was lower than the maximum. Interestingly, in patient 73, the responsible hapten was the major determinant BPO.

Table 4.  Clinical data in patients from group 3
Patient no.Clinical characteristics
Patients who developed systemic reaction after skin testing
52Systemic reaction with skin test, pruritus, and facial angioedema with 2 mg i.d. of AX
58Maculopapulae after skin test with MDM (2×10−2 M)
59Systemic pruritus and dysphonia after skin testing with AX (2 mg/ml)
62Systemic pruritus after i.d. testing with AX (20 mg/ml)
65Systemic pruritus and eyelid angioedema after i.d. testing with AX (2 mg/ml)
70Regional urticaria after skin testing with AX (2 mg/ml)
73Systemic pruritus and hypotension after skin testing with BPO (5×10−5 M)
Patients with negative skin test and positive challenge
53Maculopapulae in abdomen and legs with 5 mg of AX
54Hypotension, systemic pruritus, and generalized erythema with 5 mg of AX
55Maculopapulae in abdomen and chest after administration of 100 mg of PV
56Pruritus in hands, legs, and eye, and angioedema with 15 mg of AMP
57Maculopapulae in abdomen, legs, and chest with 100 IU of BP
60Systemic pruritus, erythema, abdominal cramps, and hypotension with 5 mg of AX
63Systemic pruritus and facial angioedema with 1000 IU of BP
64Systemic pruritus and facial erythema with 5 mg of AX
66Pruritus in palms and soles plus hypotension after 5 mg of oral AX
68Systemic pruritus, facial angioedema, and maculopapulae in abdomen with 5 mg of AX             

The second part of Table 4 shows the 10 (38%) patients with negative test results to penicillin determinants who developed a positive response after controlled administration of the drug. The symptoms were variable and hypotension was present only in patients 54 and 66. In the remaining patients, the symptoms were of minor intensity. The dose that induced the response was variable, but in no instance was the full dose required to induce the symptoms.

The results of the Pharmacia CAP System for group 3 are presented in Table 3. Positive results to BPO were found in eight patients (31%) and to AXO in 10 (38%). Combined positivity to either one or the other was exhibited by 11 (42%). With one exception, all patients that were positive to BPO were also positive to AXO.

Statistical studies

Although the time interval means in groups 1 and 2 were shorter than those found in group 3, the Levene test showed that these differences were not statistically significant. The relationship between specific IgE measurements and the time interval between occurrence of the reaction and the sera collection was studied with Spearman's rank correlation. The correlations obtained were low and nonsignificant. In group 1, for BPO, it was r=0.30 (P=0.22) and, for AXO, r=0.16 (P=0.52); in group 2, for BPO, it was r=0.06 (P=0.77) and, for AXO, r=0.22 (P=0.26). Plots of these data are presented in Figs. 1A and 1B for group 1, and Fig. 2 (AXO only).

Figure 1.

A) Serum specific IgE antibodies to BPO in group 1 in relation to time interval between serum collection and occurrence of reaction. B) Serum specific IgE antibodies to AXO in group 1 in relation to time interval between serum collection and occurrence of reaction.

Figure 2.

Serum specific IgE antibodies to AXO in group 2 in relation to time interval between serum collection and occurrence of reaction.

Moreover, when these data were transformed into categorical variables, chi-square analysis showed no significant association between a time interval of less than or greater than 60 days with a specific IgE level lower than or greater than 0.35kUA/l. In group 3, however, the study of the relationship between time interval and specific IgE levels showed significant correlations for both BPO (r=0.3; P=0.047) and AXO (r=0.4; P=0.003).

The numbers of positive and negative in vitro test results are presented in Tables 5 and 6. The proportion of positive results in group 1 was 68% for BPO and 53% for AXO. However, for positivity to either AXO or BPO, the test results rose to 74%. In group 2, which reacted specifically to AXO, only 10% had a positive in vitro test result to BPO compared to 41% with a positive in vitro test result to AXO. In group 3, which had a negative skin test but positive penicillin allergy diagnosis, 42% showed positive BPO and/or AXO test results. The proportions of negative test results for AXO in the control groups 4 and 5 were 96% and 100%, respectively, compared to 100% and 97% for BPO. Table 6 shows the sensitivity and specificity of the in vitro test considering the contribution of a single hapten (BPO or AXO) or the combination of both.

Table 5.  Number of positive and negative in vitro test results to BPO and AXO in five groups studied
GroupSerum specific IgE to BPOSerum specific IgE to AXO
PositiveNegativePositiveNegative
113 (68%)6 (32%)10 (53%)9 (47%)
23 (10%)26 (90%)12 (41%)17 (59%)
38 (31%)18 (69%)10 (38%)16 (62%)
40 (0%)25 (100%)1 (4%)24 (96%)
51 (3%)29 (97%)0 (0%)30 (100%)
Table 6.  Sensitivity in positive group (on left) and specificity in negative group (on right). Positive group comprised 1, 2, and 3 groups according to clinical history, skin tests, and/or controlled administration of drug
HaptenPositiveFalse negativeSensitivityTotalNegativeFalse positiveSpecificityTotal
BPO245032%7454198%55
AXO324243%7454198%55
BPO+AXO373750%7453296%55

Discussion

The use of in vivo and in vitro tests in the diagnosis of patients allergic to β-lactams needs to be adapted to the present situation in which the pattern of consumption of these antibiotics has changed, and BP is no longer the drug most frequently prescribed (6, 25), having been replaced mainly by AX (6, 22, 23, 25–27) and, to a lesser extent, by the cephalosporins (28, 29). We therefore used a well-established in vitro method to evaluate the presence of IgE antibodies in a defined allergic population in which AX was the drug most commonly inducing the immediate reaction. The advantage of using the Pharmacia CAP System is that it is a nonisotopic immunoassay system, well validated with classical allergens. Since the relevant drug that induced the reactions in the groups studied was AX, the Pharmacia CAP System assay employed used both AXO and BPO for the validation.

The purpose of this work was to study the IgE response to BP and AX in two groups defined by clinical history and positive skin tests, and a third, skin-test-negative group with positive history confirmed by either controlled administration or repeated reactions to the penicillin derivatives. The reason for including this third group in the Pharmacia CAP System validation was the evidence of allergic but skin-test-negative subjects (11, 12, 30, 31). In all these groups, the most relevant drug involved in the reaction was AX. For the assays, the positive and negative study groups had to be defined. The definition of a positive control for these studies is complex because controlled administration cannot be used as a reference standard due to the risk of inducing a systemic response. Therefore, the criteria considered for defining the positive study groups was a history of a clearly anaphylactic or urticaria reaction and skin test positivity to AX and at least one other penicillin determinant in group 1, and positivity to only AX in group 2. In group 3, subjects had to have had several episodes of an immediate reaction to AX or BP or confirmation by controlled administration. In order to assess the clinical specificity of the test, we chose two further groups, one with a non-IgE-mediated allergic response (group 4) and another with good tolerance to AX and other β-lactams (group 5). The proportion of positive in vitro test results in group 1 suggested that in this group cross-reactivity existed between the penicillin determinants. These results also demonstrated the good agreement between the in vitro and skin tests. The results of group 2, with specific reactions to AX, confirmed that this was a selective group in whom AXO was the relevant determinant.

Although we chose very well-defined groups for the study, the sensitivity of IgE in vitro testing was not as high as with classical allergens, where values over 90% have been described (34). However, the specificity in our groups was high and similar to that of other protein allergens (34, 35). If we consider the combination of both haptens, the sensitivity increased to 74% in group 1, but there was no change (41%) in group 2. This indicates that reactions in 26% of the patients in group 1 and 59% of those in group 2 were not confirmed by in vitro testing. The explanation for this could be the interval between the occurrence of the reaction and the serum collection, defined as time interval, or that an inadequate hapten carrier conjugate was used. Since there was no relation between the time interval and in vitro assay values, it is difficult to establish the cause of the false negatives, but the existence of patients, with a negative in vitro assay with a short time interval indicates that other causes need to be considered. However, in a previous study, we found an association between IgE antibody response in the in vivo test and the time interval, a finding which seems surprising and contradictory (36) – although similar data have been found by others (31) – but which can possibly be explained by nontherapeutic exposure to penicillins (31, 37). If we consider together the data of the three positive groups, and combine the results of the in vivo test to BPO and the in vitro test to AXO, the number of positive patients rose from 19 (25.6%) to 41 (55.4%) out of a total of 74 patients studied.

In conclusion, the results suggest that the Pharmacia CAP System RAST FEIA in vitro assay may be used to evaluate subjects allergic to β-lactams. It demonstrates the necessity of testing for specific IgE to AXO, in addition to BPO, in patients who have had an immediate allergic reaction after treatment with AX. Since there are some indications that anaphylaxis predominates over urticaria in allergic reactions to β-lactams (7), the ability to use an in vitro test is a good alternative means to detect positive patients. Furthermore, because of the possibility of finding patients with a positive history but negative skin test, the combination of in vivo and in vitro tests to estimate IgE antibodies to penicillins is a useful approach.

Acknowledgments

We thank Ian Johnstone for his help with the English language version. This study was supported by Pharmacia & Upjohn CAP96/03, FIS 00/0838, and FEDER 1FD97–0516.

Ancillary