Flow cytometric investigation of peri-anaesthetic anaphylaxis using CD63 and CD203c
The investigation of anaphylactic reactions in the peri-operative period is difficult. Elevation of serum tryptase levels is a good indicator of an anaphylactic event but the ability of subsequent investigations to identify the drug(s) responsible for the reaction is still potentially unreliable. The aim of this study was to examine basophil activation as an investigative tool. We performed flow cytometric analysis of the expression on the cell surface of the basophil activation markers CD63 and CD203c and measured histamine release in 21 patients who were referred with possible peri-operative anaphylaxis. The sensitivity of CD63, CD203c, basophil histamine release and skin prick for the muscle relaxants was found to be 79%, 36%, 36% and 64%, respectively; the specificity was found to be 100%. These results demonstrate the difficulty in investigating the cause of an unexpected clinical event following drug administration, but the higher sensitivity of neo-expression on the cell surface of CD63 suggests that flow cytometric analysis of its neo-expression on basophils in vitro may be a diagnostic aid.
Investigations commonly used in the diagnosis of allergy and anaphylactic reactions include demonstrating an almost immediate rise in serum tryptase, measuring the specific IgE to the allergen and skin prick tests. Mast cell tryptase is released during an anaphylactic reaction and the serum concentration peaks in the first 6 h . Skin prick tests are by far the most common tests performed to investigate allergy in the outpatient clinic. Raised serum tryptase and positive skin prick tests demonstrate mast cell reactivity. However, the mast cell is a differentiated tissue cell type that is not as accessible for detailed study as its circulating precursor, the basophil.
Basophil activation tests have been used in the diagnosis of allergy to various allergens. Basophil activation studies include those of mediator release (histamine and sulphidoleukotrienes) and flow cytometric analysis of changes in the molecules expressed at the cell surface. The most common basophil activation marker used in current research is the neo-expression of CD63. CD63, an activation marker on basophils, has been used as a reporter molecule of the state of activation produced by a large number of allergens [2, 3] and neuromuscular blocking drugs [4, 5]. In both the latter studies, the specificity has been shown to be high (93% and 100%) and the sensitivity low (63% and 54%). More recently, the expression of CD203c, found primarily on basophils, mast cells and CD34 progenitors , has been reported to be a more sensitive marker of basophil activation  but there is some disagreement about this . There are no studies known to us evaluating CD203c expression in patients with anaesthetic drug allergy.
We studied the effect of the incriminated drug or drugs on basophils from 21 patients known to have had a peri-anaesthetic anaphylactic reaction. We measured the histamine release and the changes in the expression of the cell surface activation markers CD63 and CD203c when whole blood was challenged with the drug/drugs in question in in vitro experiments.
Twenty-one patients were studied following a clinical event suggestive of an adverse drug reaction at induction of anaesthesia (Table 1). The patients were recruited after informed written consent and the study had the approval of Bro Taf Local Research Ethics Committee. We also recruited 10 normal volunteers as controls. A normal volunteer was defined as one with no remarkable medical history, no history of allergy and negative skin prick tests to a panel of common allergens. Blood from one of the 10 normal volunteers was run as a negative control alongside every patient sample in all tests performed. In all cases of neuromuscular block allergy we used a second neuromuscular blocking drug belonging to a different pharmacological group to assess cross-reactivity.
Table 1. Drug that patients are allergic to and the clinical symptoms demonstrated when exposed to the drug.
| 1||F||Lignocaine||Bronchospasm, syncope||Negative||Positive||Negative||Negative|
| 3||F||Prilocaine||Syncope, tachycardia||Negative||Negative||Negative||Negative|
| 5||F||Thiopentone||Rash, profound hypotension||Not Done||Positive||Negative||Negative|
| 6||M||Alfentanil||Bronchospasm, hypotension||Positive||Negative||Negative||Negative|
| 7||F||Morphine||Profound hypotension||Positive||Negative||Negative||Negative|
| 8||F||Alcuronium||Hypotension, bronchospasm||Positive||Positive||Negative||Positive|
| 9||M||Atracurium||Hypotension, bronchospasm||Positive||Positive||Positive||Positive|
The following monoclonal antibodies were used: phycoerythrin (PE)-conjugated anti-human CD63 (Pharmingen, Becton Dickenson, Oxford, UK); 20 μl for every 100 μl whole blood; PE-conjugated anti-human CD203c (Immunotech, Marseille, France); 20 μl for every 100 μl whole blood; phycoerythrin-cyanin conjugated anti-human CD45 (Dako, Ely, UK); 10 μl for every 100 μl whole blood, and polyclonal fluorescein isothiocyanate FITC-conjugated antihuman IgE (Dako; 10 μl for every 100 μl whole blood).
Analysis was carried out on heparinised peripheral venous blood within 3–4 h of sampling. For each sample, we prepared both negative (phosphate-buffered saline, PBS) and positive (10−6m formyl-methionyl-leucyl-phenylalanine, fMLP; Sigma, Dorset, UK) controls. Aliquots (100 μl) of each blood sample were incubated with three different dilutions in PBS of the suspected drug. The following dilutions were used: 1/10 of stock solution, 1/100, 1/1000. All the drugs used were commercially available and used routinely in clinical practice. After mixing blood with the drugs in appropriate dilution, PBS or fMLP, the samples were incubated in a water bath at 37 °C for 15 min followed by a 15-min staining performed at +4 °C in a dark chamber. Finally, samples were lysed using Uti-LyseTM (Dako).
Samples were analysed on a Dako-Partec Galaxy flow cytometer. Similar protocols were used for CD63 and CD203c, both protocols having been previously described by Boumiza et al.  in the diagnosis of latex allergy. When cells were stained with CD45-CY5, IgE-FITC and CD63-PE (or CD203c) monoclonal antibodies with gating on the anti-CD45 positive population and the IgE-positive population results were considered positive if greater than 10% of this subpopulation stained with anti-CD63 or anti-CD203c (with < 5% staining in the control group). Results were considered positive only if the two dilutions of the drug demonstrated greater than 10% positivity, the criterion used by Monneret et al. .
Histamine release was measured in the supernatant using an immunoassay (Immunotech, Marseille, France). Aliquots (100 μl) of diluted blood were simultaneously challenged with 10 μl of three different concentrations of the drug under investigation and a negative (PBS) and positive control (fMLP).
The results were evaluated as percentage release as compared to the total histamine content. The total histamine content was measured by freeze-thawing a given amount of diluted blood (as per the manufacturer's protocol). A result was considered positive if the release was greater than 10% of the total histamine release.
The clinical sensitivity of the assay was calculated as the ratio of the number of drug allergic patients with a positive test to the total number of such patients. Specificity was calculated as the ratio of the number of control subjects with negative test results to the total number of control subjects tested. The odds ratio for a test was calculated as the ratio of the number of allergic patients with a positive test relative to the number of allergic patients with a negative test result divided by the ratio of the number of control subjects with a positive result relative to the number of control subjects with a negative test result.
We recruited 21 cases, of whom 17 were females. Fourteen were referred for investigation of reactions to neuromuscular blocking drugs (11 females), three with reactions following local anaesthetic administration (all females), two female patients with reactions to intravenous agents (midazolam and thiopental) and two with reactions to opiate anaphylaxis (one male, one female) (alfentanil, morphine) (Table 1).
Eleven of the 14 cases who reacted to neuromuscular blocking drugs showed neo-expression of CD63 and five showed increased expression of CD203c on stimulation with the appropriate drug. Six patients showed cross-reactivity between the different groups of muscle relaxants (Table 2). The flow cytometry analysis was negative in three cases of muscle relaxant reactivity (rocuronium, succinylcholine and vecuronium, patients Nos. 14, 16 and 21, respectively). There were three cases of muscle relaxant anaphylaxis in whom the RAST test for specific IgE was done at the time of the incident and in only one of these three, was the test positive for succinylcholine allergy. The RAST test is commonly available only for succinylcholine. In seven cases the mast cell degranulation marker, serum tryptase, was assayed. The serum tryptase levels were raised in three of the seven cases.
Table 2. Muscle relaxants and cross reactivity.
The three cases of local anaesthetic reaction (patients Nos. 1, 2 and 3) tested negative to skin prick tests, flow cytometry and histamine release.
Two patients (patients Nos. 6 and 7) who reacted to opiates had strongly positive skin prick tests but both failed to express CD63 and CD203c. They also did not release histamine. The patients (Nos. 4 and 5) with midazolam and thiopental sensitivity were both positive for CD63. While the patient who reacted to midazolam demonstrated histamine release and positive CD203c, the one who reacted to thiopental was negative for both CD203c and histamine release.
The sensitivity of neo-expression of CD63 and CD203c, basophil histamine release, and positive skin prick tests for the respective muscle relaxants that patients had reacted to was found to be 79%, 36%, 36% and 64%, respectively; the specificity was found to be 100%(Table 3). None of the normal volunteers demonstrated any positive reaction to the drugs tested. Statistical analysis was not performed for the other group of drugs, as the numbers of patients were too small.
Table 3. Sensitivity and specificity of the tests performed for the muscle relaxant group.
|Skin prick tests||64||100|
Changes in the expression of activation markers CD63 and CD203c on the basophil surface were identifiable in 77% of patients who had adverse reactions to muscle relaxants and induction agents. The sensitivity of CD63 neo-expression for reactivity to muscle relaxants was better than in the two previously reported studies [4, 5]. The sensitivity for CD203c was only 36% but specificity was 100%. However, changes in basophil CD63 and CD203c expression were not observed in vitro in patients who were thought to have reacted to opiates or to local anaesthetics.
Activation of basophils induces the fusion of cytoplasmic granules with the plasma membrane, resulting in the release of inflammatory mediators and cytokines. CD63 is a transmembrane protein which is anchored in the basophil granule membrane. As basophils degranulate, it is exposed to the exterior . CD63 expression on basophils has been shown to be increased by a number of different basophil activators including IgE, fMLP and calcium ionomycin. Flow cytometric analysis of in vitro activated basophils using CD63 neo-expression has proved quite reliable for the diagnosis of different IgE-mediated allergies including allergies to inhalant allergens (e.g. pollen allergy) , food  and venom . It has also been used with some success in evaluating reactions to drugs, including muscle relaxants. However, CD63 is also present in other cell types (platelets and macrophages) in the blood. CD203c, on the other hand, is reported to be exclusively expressed on basophils, mast cells and their progenitors and to be up-regulated in response to allergen-specific and anti-IgE-induced IgE cross-linking as for CD63. CD203c (neural cell surface differentiation antigen E-NPP3) belongs to the ectonucleotide pyrophosphate/phosphodiesterases, which also include E-NPP2 (human tumour motility stimulating protein PD-1α (also called autotoxin) . CD203c has recently been used in the investigation of latex allergy . Discovery of the restricted expression of CD203c on basophils and mast cells generated excitement, as early studies suggested an increased sensitivity compared with CD63. However, the superiority of CD203c as a marker of basophil activation has since been questioned . In our study, 11 of the 14 patients with suspected reactivity to muscle relaxants expressed CD63, but only five of the 14 patients expressed positive with CD203c. One reported advantage of a cell surface marker's expression being restricted to a type of circulating cell, as CD203c is to basophils, is the ability easily to isolate and study these cells on a flow cytometer with little interference from other circulating cells . Using 3-colour flow cytometric staining (similar to that used to analyse CD63 expression) to define the basophil subpopulation in peripheral blood, the variable level of expression within that subpopulation of the marker CD203c can be analysed.
CD63 expression and CD203c expression were both normal in the two patients who reacted to opiates. Both patients had strongly positive skin prick tests and reacted unequivocally clinically. In both cases, serum tryptase was not assayed. Why CD63 and CD203c were not positive in the presence of such a clear clinical event is not known, though it suggests either that opiates act on mast cells without affecting basophils or alternatively that any basophil action induced by opiates does not entail CD63 or CD203c neo-expression. A larger population needs to be studied.
Amide local anaesthetic agents are not known to cause anaphylaxis and yet from time to time patients do present with such a clinical history. A number of possible explanations for such reactions have been attempted. Some of them include the presence of preservatives, inadvertent intravascular injections and possibly vasovagal reactions. The three cases of local anaesthetic reactions we studied had syncope. One of them had tachycardia, and the patient with lidocaine anaphylaxis in addition to tachycardia had bronchospasm. All three failed to show increased CD63 or CD203c neo-expression after incubation with the respective drug.
Poor reporting systems, confusing terminology and the lack of centralised registries mean that there has been very unsatisfactory reporting and thus analysis of the extent of anaphylaxis in the peri-operative period. The European Academy of Allergology and Clinical Immunology recently revised the nomenclature for allergy . They defined hypersensitivity as ‘causing objectively reproducible symptoms or signs, initiated by exposure to a defined stimulus at a dose tolerated by normal subjects’ and anaphylaxis as a ‘severe, life threatening, generalised systemic hypersensitivity reaction’. They stipulated that all reactions should be termed anaphylaxis until further investigation into the mechanism of action has been performed. They also suggested that words like histaminoid and anaphylactoid should no longer be used, to remove confusion.
The mechanisms responsible for these drug reactions remain unclear. The finding of specific IgE against the implicated drug in some of these cases might suggest IgE-mediated anaphylaxis. However, in an equally large number of these cases, specific IgE is not detected. This suggests an alternate mechanism or poor sensitivity of the assay to specific IgE antibodies, available for only very few drugs .
Skin prick tests may be useful in demonstrating mast cell reactivity to the drug being investigated. Three methods described are prick testing, intradermal testing and patch testing. Tests used at different centres vary with respect to these and the concentration of drugs used. Case reports of anaphylactic reactions to the test dose have been reported . The sensitivity and specificity of skin testing can be variable, and it is not reliable for certain drug groups. With a positive history very suggestive of clinical anaphylaxis and a negative skin prick test, the picture is potentially very difficult to interpret. There is laboratory evidence that mast cells in different tissues react differently to muscle relaxants , suggesting that even if the skin mast cells are negative, mast cells from elsewhere might be reactive, making evaluation complex. Basophil activation tests are not directly comparable to mast cell reactivity using skin prick tests, even though both these cell types have FcεR1-type receptors on their cell surface which cross-link via IgE. Tests of basophil and mast cell function may contribute to the diagnosis and help in further research but they test cells in different stages of differentiation.
In the absence of established reliable diagnostic algorithms, a thorough clinical history and assessment by a senior anaesthetist and an immunologist or allergist is currently the optimal evaluation. The Association of Anaesthetists of Great Britain and Ireland guidelines suggest how an acute anaphylactic reaction should be managed and also puts the onus on the anaesthetist to arrange for a relevant referral to a specialist and ensure subsequent follow-up . The guidelines also advise collecting samples for testing serum tryptase levels at the time of the acute event, following immediate resuscitation and at later time points. Despite such strong recommendations, tryptase levels are not taken in many cases or may be taken but, for whatever reason, not assayed. It is essential for individual anaesthetic departments to know how to take and handle samples for tryptase assay and where they should be sent. Tryptase is a stable molecule in serum and the assay for it is reliable. A raised serum tryptase concentration significantly above the upper limit of normal followed by a decrease indicates mast cell activation.
Anaphylaxis in the peri-operative period has the potential to develop into a catastrophic event; besides the immediate threat to health and life there are serious implications for future management.
The lack of understanding of the mechanisms, poor record keeping and confusion in the terminology have continued to hamper progress in the diagnosis and identification of the responsible drug or drugs. Thorough documentation of events and case analysis by specialists with relevant insight help in identifying the causative drugs. Elevation of serum tryptase levels confirms that mast cell degranulation has occurred. Further evaluation of which drug is responsible is more difficult. In these studies, the sensitivity of skin prick testing to implicate a suspected drug was 64%, whereas tests of basophil activation using CD63 neo-expression had a higher sensitivity of 79%.
We are very grateful to The Association of Anaesthetists of Great Britain and Ireland for a project grant which made this study possible.