Allergic reactions occurring during anaesthesia

Abstract

Introduction Although our understanding of allergic reactions during anaesthesia has substantially increased over the past 30 yr, they remain a major cause of concern to anaesthetists. Anaphylaxis is an acute allergic reaction resulting primarily from the rapid antigen-induced, usually IgE-dependent release of potent, pharmacologically active mediators from tissue mast cells and peripheral blood basophils. Anaphylactic reactions may be exacerbated in severity or prolonged in duration by mediators derived from these cells or from other secondary recruited inflammatory cells. These reactions differ from 'anaphylactoid reactions' that are clinically indistinguishable from anaphylaxis but are triggered independently of IgE antibody. Since the initial report describing an anaphylactic reaction to succinylcholine [1], followed by clinical observations reported by Fisher [2], Sigiel and colleagues [3] and Vervloet and colleagues [4], increasing interest has focused on immune-mediated reactions occurring during anaesthesia. Moreover, in light of the increasing number of anaesthetic drugs, the need for confirmation and quantitative risk assessment of suspected rare serious adverse reactions requiring precise epidemiological studies [5] has continually been reinforced. In addition, awareness of the constant changes in anaesthetic practice has led to elaboration of a growing number of practice guidelines concerning the diagnosis, high-risk group identification and management of anaphylaxis [6-11]. Mechanism The term 'anaphylaxis' was introduced in 1902 by Charles Richet and Paul Portier to describe the hypersensitivity reactions they observed in dogs after repeated injections of sea anemone toxin [12]. In immunological terms, anaphylaxis is an example of an immediate Type 1 hypersensitivity reaction. Most cases are IgE or rarely IgG mediated. The typical sequence of events in immediate hypersensitivity begins with the production of IgE by B-cells in response to initial exposure to an antigen. These antibodies bind to specific Fc receptors on the surface of effector cells such as mast cells and basophils. The interaction of reintroduced antigen with the bound IgE leads to activation of these cells and the release of various mediators [13]. The mediators are usually classified as preformed or newly synthesized. They include histamine from mast cells and basophils, and tryptase from mast cells [14-16]. The substances are responsible for the clinical manifestations of immediate hypersensitivity. Release is produced by cellular activation and transduction triggered by the bridging of IgE-receptor complexes with allergens. The magnitude of degranulation is influenced by the affinity of a given drug for cell-bound IgE antibodies as well as by their number on the cell surface. It is estimated that the human basophil contains 40 000-100 000 IgE antibody receptors. Basophils and mast cells bind IgE via a high-affinity receptor (FcεRI), whereas lymphocytes, eosinophils and platelets bind them via a low-affinity receptors (FcεRII) triggering the release of further mediators including kinins, prostaglandins, leukotrienes, serotonin or eosinophil cationic protein [17]. Most anaesthetic drugs such as muscle relaxants or opioid derivatives are low molecular weight molecules and are considered as haptens. As such, they are incapable of inducing the production of drug-specific antibodies by themselves. Consequently, previous conjugation of the native drug or one of its degradation products with some protein carrier and processing by professional antigen-presenting cells such as dendritic cells before presentation of peptides on the cell surface in close association with a Class I or II histocompatibility molecule, is usually regarded as the initial step of sensitization [18-20]. Direct interaction with proteins present on the surface of the dendritic cell, as reported for sensitizing metal and some antibiotics, and Class II histocompatibility molecular interactions, should also be given consideration [21-23]. Although antigen presentation by dendritic cells is regarded as an essential step for induction of an immune response in modern immunology, the elicitation of an immediate IgE-mediated reaction to the same drug is influenced by different immunochemical requirements. In anaphylaxis, mast cells and basophils are activated by the cross-linking of FcεRI molecules. This is thought to occur by the binding of multivalent antigens to the attached molecules. This could be a possible explanation for the increased number of anaphylactic reactions to neuromuscular-blocking agents, in comparison with other drugs used during anaesthesia. With respect to muscle relaxants, the main antigenic determinants involved in the generation of specific IgE antibodies are substituted ammonium ions. This was initially demonstrated by Baldo and Fisher [24]. As a result, it has been hypothesized that most neuromuscular blocking agents that bear two similar quaternary ammonium ions per molecule are capable of bridging IgE antibodies and eliciting anaphylaxis. In this regard, the flexibility of the chain between the ammonium ions as well as the distance between the quaternary ammonium ions might be of importance during the elicitation phase of anaphylaxis [25,26]. Flexible molecules such as succinylcholine can stimulate sensitized cells more strongly than rigid molecules (e.g. pancuronium). The relative affinities of the various muscle relaxants to their corresponding IgEs may also play a role. This also explains the cross-reactivity between the different muscle relaxants observed with IgE antibodies of most patients, and initially evidenced by skin testing [27,28]. This cross-reactivity, however, was only observed in 70% of patients presenting with anaphylaxis to muscle relaxants [29]. In addition to the above-mentioned steric considerations, which could explain why two muscle relaxants do not necessarily behave similarly, further hypotheses have been proposed. In some cases, the antigenic determinant may either correspond to the quaternary ammonium epitope or extend to an adjacent part of the molecule. Another rare possibility might be that IgE antibodies could be complementary to structures other than the ammonium group [30]. The main antigenic determinants involved in the generation of specific IgE antibody towards other anaesthetic drugs has also been defined. Two antigenic determinants have been identified in the thiopental molecule: the secondary pentyl and ethyl groups attached in position 5 of the pyrimidine ring nucleus and the thiol region on the opposite side [30,31]. The antigenic determinant on morphine comprises the methyl-substitute attached to the N-atom and the cyclohexenyl ring with a hydroxyl group at carbon 6. Cross-reactivity between morphine, codeine and other narcotics has been reported [32]. Finally, 240 potentially allergenic proteins have been identified in processed latex products. Seven sensitizing proteins have been identified or cloned and assigned allergen designation Hev b1-b7 [33-36]. A 14-kD protein (rubber elongation factor) is one of the major allergens responsible for allergic reactions among healthcare workers, and a 27-kD protein has been implicated among several vulnerable patient populations [37,38]. Hevein [39], pro-hevein [40], latex lysosome and rubber elongation factor [41] are other potent implicated allergens. In some instances, IgE-mediated anaphylactic reactions have been reported at the first known contact with an incriminated drug. This suggests a possible cross-reaction with IgE antibodies generated by previous contact with apparently unrelated chemicals. This is a particularly attractive hypothesis in cases where patients react to relatively small and ubiquitous epitopes such as a quaternary ammonium group [21]. The latter is the case of neuromuscular blocking agents [24]. Such a hypothesis has been proposed for adverse reactions to neuromuscular blockers and IgE antibodies generated towards acetylcholine, different membrane phospholipids or food lecithins from soy or egg [42]. Similar observations have been made concerning anaphylactic reactions to latex in patients with a history of food allergy to different fruits (avocado, kiwi, banana, fig, chestnut, hazelnut, sweet pepper, melon, pineapple, papaya). They contain several proteins similar or identical to those found in latex [43-58]. Epidemiology The observation of possible non-specific histamine release triggered by compounds having a simple chemical structure [59] has led to intense debate about the real nature of adverse reactions associated with anaesthesia during symposia organized in France, the UK and Germany [60-66]. The concept of shock related to non-specific histamine release by various anaesthetic substances was supported by many studies carried out by Lorenz and colleagues [67,68]. However, since the first report of an anaphylactic reaction to succinylcholine [1], followed by clinical observations reported by Fisher [2], Sigiel and colleagues [3] and Vervloet and colleagues [4], increasing interest has been focused on immune-mediated reactions occurring during anaesthesia. The development of skin tests [69-71] and the identification by Baldo and Fisher [24] of the particular role played by quaternary ammonium ions on specific IgE production led to the recognition of the frequent implication of muscle relaxants in anaphylaxis during anaesthesia [72]. Most reports on the incidence of anaphylaxis originate in France [29,73-76], Australia [2,77-79], the UK [80-84] and New Zealand [69,85]. They reflect an active policy of systematic clinical and/or laboratory investigation of anaphylactoid reactions suspected to be mediated by an immune mechanism. Anaphylactic reactions have also been reported for smaller series in the USA [86,87]. Nevertheless, the true incidence of anaphylactic reactions and their associated morbidity/mortality remain poorly defined. This is due to uncertainties over reporting accuracy and exhaustiveness. This is illustrated by the differences observed between reports from various developed countries. First, the clinical reaction must be recognized and it should be emphasized that patients experiencing anaphylaxis during anaesthesia present with a variety of signs and that these may not appear all at once. In addition, some debate remains about the interpretation and significance of skin and laboratory tests usually required to distinguish between anaphylactic and anaphylactoid reactions. Finally, there are often difficulties in obtaining valid data on the number of patients exposed to the risk in the population from which reactions are reported. With these limits in mind, the estimated incidence of anaphylaxis was 1:10 000-20 000 in Australia in 1993 [78] and 1:13 000 in France in 1996 [29]. Although rare, these may lead to death, even when appropriately treated [29,88], with a mortality rate ranging from 3.5% [89] to 4.7% [90]. The prevalence of the risk of anaesthetic anaphylaxis in the overall population, which corresponds to the proportion of patients who would respond, if exposed to antigenic anaesthetic substances remains poorly defined. Skin tests and/or specific IgE assays performed on the general population could estimate it. However, their clinical significance is questionable. If positive, they reflect an IgE-dependent sensitivity but do not necessarily indicate that an allergic reaction will occur. Nevertheless, the prevalence of allergy to anaesthetic drugs is probably higher than the incidence of reactions, because patients undergoing anaesthesia do not necessarily receive the drug to which they are allergic. The prevalence of muscle relaxant sensitivity, based on skin test positivity and/or detection of specific IgE to quaternary ammonium ions, shows wide variation. A positive reaction in 9.3% of the general population has been reported [91], with extremes ranging from 1.6% in patients with no history of atopy and/or drug allergy to 16% in patients presenting these risk factors [92,93]. This contrasts with the incidence of anaphylaxis to neuromuscular blocking agents, which has been estimated as 1:6500 for anaesthesia where a muscle relaxant is included in the anaesthetic protocol [29]. The prevalence of latex sensitization, a major cause of anaphylaxis during anaesthesia, varies depending on the population studied. As with all allergy-causing substances, the greater the exposure in a given population, the greater the number of sensitized individuals. Nevertheless, significant differences have been reported. The prevalence of latex sensitization in the literature has been reported to vary from approximately 1 to 6.6% [94,95], reaching 15.8% in anaesthesiology staff [96]. However, as previously mentioned, this IgE-dependent sensitivity does not necessarily indicate that an allergic reaction will occur in case of exposure to latex. Causative agents Drugs Muscle relaxants. Among the drugs and other agents involved in anaphylaxis, muscle relaxants are most frequently involved, with a range of 50-70% [29,84,86,87,97-99](Fig. 1). In France the incidence of anaphylaxis to muscle relaxants was estimated at 1:6500 anaesthetic procedures involving a muscle relaxant in 1996 [29]. Anaphylactic reactions have been reported for all neuromuscular blocking agents, even with recently commercialized substances [29,99-111]. In most series, succinylcholine appears to be more frequently involved [74,76,112], with some differences reflecting variations in anaesthesiological practices from one country to another [13,99]. In the French survey conducted between 1 January 1997 and 31 December 1998, the percentage of anaphylactic reactions observed with each drug was compared with the estimated number of patients who effectively received these drugs over the study period [99]. The results derived from such estimates should, however, be carefully contemplated. However, they indicate that succinylcholine and rocuronium seemed to be more frequently involved. Vecuronium and pancuronium followed them whereas atracurium was the least frequently involved (Fig. 2). It should also be noted that in this series, anaphylaxis to a neuromuscular blocking agent was observed in 48 patients (14.7%) with no history of anaesthesia, and, as a consequence, no previous administration of any muscle relaxant.Figure 1: Agents involved (%) in anaphylaxis during anaesthesia in France (n = 477) from January 1997 to December 1998.Figure 2: Neuromuscular-blocking agents (%) responsible for anaphylaxis in France (n = 336) from January 1997 to December 1998.Hypnotics. The estimated incidence of anaphylactoid reactions with thiopental was estimated as 1:30 000 [80]. It was suggested that most of the generalized reactions were related to its ability to elicit direct leukocyte histamine release [113]. However, there is evidence for IgE-mediated anaphylactic reactions based on skin tests and a specific IgE assay [114-117]. Although the radioimmunoassay developed for the detection of antibodies that react with thiopental is a valuable aid in confirming the diagnosis of Type I allergy to this drug, its use requires some specific consideration. IgE antibody formed in patients who react to a neuromuscular blocking agent could cross-react in vitro with the thiopental solid phase. In this case, however, thiopental does not inhibit the binding of these antibodies to the thiopental solid phase but inhibition is observed with the neuromuscular blocking agent. This allows one to distinguish between sensitization to thiopental and neuromuscular blocking drugs [30,31,118,119]. Ever since Cremophor EL (used as a solvent for some non-barbiturate hypnotics) has been avoided, many previously reported anaphylactoid reactions have disappeared. Although less frequent, anaphylaxis to all induction agents has been observed [13,120]. In the last French survey, five cases to thiopental, 10 to propofol and three to midazolam were recorded [99](Fig. 1), whereas no anaphylactic reaction to etomidate and ketamine was observed. Opioids. Reactions to morphine, codeine phosphate, meperidine, fentanyl and its derivatives are uncommon [32]. Because of their direct histamine-releasing properties, distinction between anaphylaxis and non-immune-mediated histamine release is not always easy [121,122]. Hapten inhibition studies performed in the serum of a subject who experienced an anaphylactic reaction following the administration of papaveretum [123] has led to the identification of the allergic determinant (cyclohexenyl ring with a hydroxyl group at C-6 and methyl substituent attached to the N atom) involved in IgE binding [30]. Only seven cases were recorded in the last 2-yr epidemiologic survey in France (morphine = 1, fentanyl = 4, sufentanil = 2) [99](Fig. 1.). Local anaesthetics. Allergic reactions to local anaesthetics are rare despite their frequent use. It is estimated that <1% of all reactions to local anaesthetics have an allergic mechanism [76,124,125]. Inadvertent intravascular injection leading to excessive blood concentrations of the local anaesthetic, or systemic absorption of epinephrine that was combined with the local anaesthetic, are by far the most common causes of adverse reactions produced by these drugs. In a series by Fisher and Bowey [125] that reports the results of an investigation conducted in 208 patients with a history of allergy to local anaesthetics over 20 yr, four patients were reported to have had an immediate allergy, and four patients had delayed allergic reactions. Although severe anaphylactic reactions have been reported with both types of local anaesthetics [126-129], ester local anaesthetics, having a benzoic acid ring in their structure and the capability of producing metabolites related to para-aminobenzoic acid [130-132], are more likely than amide local anaesthetics to provoke an allergic reaction. Allergy to local anaesthetics may also be due to methyl-paraben [130-133], paraben [134] or metabisulphites [135,136] used as preservatives in commercial preparations. Non anaesthetic drugs. Antibiotics are commonly administered perioperatively and can cause allergic reactions. A discussion of allergic reactions to antibiotics is beyond the scope of this review. However, their frequency has increased over the last 20 yr. They account for between 2 and 8% of reported anaphylactic reactions [99,112], cephalosporin being most commonly incriminated in Australia, whereas penicillins remain most frequently involved in France. Vancomycin, which is increasingly used for prophylaxis, has also been incriminated in some instances [137]. However, in most cases, the adverse reactions observed are related to the chemically mediated red-man syndrome associated with rapid vancomycin administration [138-140]. Protamine, whose use to reverse heparin anticoagulation has increased over the last two decades, has also been incriminated [76,103,141]. Reactions may involve a number of mechanisms including IgE, IgG and complement [142-145]. Aprotinin, a naturally occurring serine protease inhibitor, has found widespread applications either by the i.v. route or as a component of biological sealants, because of its ability to decrease blood loss and, as a consequence, transfusion requirements. Anaphylactic reactions are mediated by IgG and IgE antibodies [146]. The risk of anaphylactic reactions has been estimated as between 0.5 and 5.8%. Patients previously treated with this drug present an increased risk [147-149]. Perioperative exposures to agent other than to drugs Latex. Allergy to natural rubber latex, which contains a complex mixture of water-soluble plant proteins, has become a major source of concern in clinical practice. It is the second most common cause of anaphylaxis during anaesthesia in the general population. However, in children subjected to numerous operations, particularly those suffering from spina bifida, it is the primary cause of anaphylaxis [150,151]. Most patients are sensitized to proteins originating from rubber tree sap (Hevea brasiliensis) present in products made from latex, such as gloves, catheters and various medical or non-medical products containing natural rubber. Latex exposure can occur as a result of contact with the skin or mucous membranes, with inhalation, ingestion and parenteral injection or with wound inoculation. The incidence of allergy to latex has rapidly increased, rising from 0.5% before 1980 to 19% in 1994 in France [76]. However, in the last French survey, latex sensitization was held responsible for 12.1% of recorded cases [99](Fig. 1). These results appear to be somewhat encouraging. They seem to indicate that increasing awareness of the risk of latex sensitization in children with spina bifida [152,153] or healthcare workers [154,155], combined with the efficacy of surgery in a latex-safe environment [156,157], could be responsible for the decrease of anaphylaxis to latex we observed. Colloids. All synthetic colloids have been shown to produce clinical anaphylaxis. The overall incidence of reactions has been estimated to range between 0.033% [158] and 0.22% [75]. Although direct release of histamine has been reported with urealinked gelatin [159], evidence for IgE-mediated adverse reactions to gelatin has been reported [75]. In addition, adverse reactions to urea-linked gelatin (0.852%) seem to be more frequent than with modified fluid gelatin (0.338%) [75], whereas IgG-mediated adverse reactions to hydroxyethyl starch are uncommon [160-163]. Adverse reactions to dextrans were estimated as 0.275%, when it was 0.099% for albumin and 0.058% for hydroxyethyl starch solutions [75]. Eleven anaphylactic reactions to gelatin and two reactions to hydroxyethyl starch solutions were reported in the last French survey of anaphylaxis during anaesthesia [99](Fig. 1). Clinical features The intensities of allergic reactions show striking variation from one patient to another. Manifestations may range from mild non-life-threatening anaphylaxis to severe anaphylactic shock and death [13,78,99,164]. The onset and severity of the reaction are related to the mediator's specific end organ effects. Consequently, the difference between anaphylactoid and true anaphylactic reactions cannot be made on clinical grounds alone. IgE-dependent anaphylaxis was evident in 53% of cases [29] in a recent study involving patients investigated for an anaphylactoid reaction during anaesthesia. Clinical symptoms reported in patients with a true anaphylactic reaction and in those presenting with non-IgE-mediated anaphylactoid reactions were similar. However, when a classification based on symptom severity was applied (Table 1)[158], clinical manifestations were more severe in patients with documented anaphylaxis. Nevertheless, some cases corresponding to true IgE-mediated anaphylactic reactions were classified as Grade I or II. As a result, any suspected anaphylactoid reaction occurring during anaesthesia should be thoroughly investigated to establish a precise diagnosis and appropriate recommendations.Table 1: Grade of severity for quantification of the anaphylactoid reaction.Anaphylaxis may occur at any time during anaesthesia and may progress slowly or rapidly. Ninety per cent of reactions appear within minutes after the i.v. injection of anaesthetic products or antibiotics. Alertness is essential because reactions may be well established before they are noticed. The most commonly reported initial features are pulselessness, a difficulty in lung inflation and desaturation [103]. In our experience, a decreased end-tidal CO2 expiration is also of valuable diagnostic interest. If the signs appear later during the maintenance of anaesthesia, they suggest an allergy to latex or volume expanders [75,165,166]. Latex allergy should also be considered when gynaecological procedures are performed. Particles from obstetricians' gloves, which accumulate in the uterus during obstetrical manoeuvres, could suddenly be released into the systemic blood flow following oxytocin injection [29,167]. Anaphylactic reactions to antibiotics have also been reported following removal of a tourniquet during orthopaedic surgery [168,169]. Factors that influence allergic reaction symptom severity in the sensitized individual include the distribution and reactivity of sensitized mast cells and basophils, individual organ susceptibility, released mediators and the endogenous response they elicit [170]. Anaphylaxis commonly involves the skin, cardiovascular and respiratory systems, as well as virtually any system, including the gastrointestinal, central nervous and genitourinary systems. The most recent French epidemiological survey, conducted between January 1997 and December 1998, involved 477 patients having experienced a true anaphylactic reaction during anaesthesia and most adverse reactions were of Grades II (22.9%) or III (62.6%), whereas only 10.1% of Grade I and 4.4% of Grade IV cases were recorded. Interestingly, in this series, reactions to neuromuscular blocking agents were more severe than those to latex. Cutaneous symptoms were present in 69.6% of cases (n = 332), angio-oedema in 11.7% (n = 56), bronchospasm in 44.2% (n = 211), arterial hypotension in 17.8% (n = 85), cardiovascular collapse in 53.7% (n = 256), bradycardia in 2.1% (n = 10) and cardiac arrest in 4% (n = 19) [99]. No difference in the severity of clinical symptoms was observed with regards to gender, history of atopy, asthma and food or drug intolerance. However, a significant association between the onset of clinical bronchospasm and a history of atopy or asthma was observed. Clinical features may occur as an isolated condition [29,78,99,171]. Therefore, an anaphylactic reaction restricted to a single clinical symptom (e.g. bronchospasm, tachycardia with hypotension) can easily be misdiagnosed because many other pathological conditions may present identical clinical manifestations [13]. In mild cases restricted to a single symptom, spontaneous recovery may be observed even in the absence of any specific treatment. However, it should be kept in mind that under such circumstances, the lack of a proper diagnosis and appropriate allergologic assessment could lead to fatal re-exposure [172]. In our last survey, cardiovascular symptoms (hypotension or cardiovascular collapse) were the sole features in 10.5% (n = 50), bronchospasm in 3.2% (n = 15) and cutaneous symptoms in 7.8% (n = 37) of cases [99]. In most cases, after adequate treatment, clinical signs regress within 1 h without sequelae. However, in some cases, bronchospasm can be particularly severe and resistant to treatment, with a risk of cerebral anoxia. Prolonged inotropic support might also be required in some patients. Moreover, previous treatment by β-adrenoreceptor blocking agents is a potential risk factor explaining a lack of tachycardia, as well as resistance of arterial hypotension to adrenaline [164]. Risk factors The potential severity of anaphylaxis during anaesthesia underscores the interest of developing a rational approach to reduce its incidence by identifying potential risk factors before surgery. With respect to drug allergy, different items such as gender, previous general anaesthesia, atopy and other drug allergies should be taken into account. Special attention should also be paid in case of allergy to latex. Gender A female predominance has been demonstrated in perioperative anaphylactic reactions, particularly those concerning allergic reactions to muscle relaxants with a female: male ratio ranging from 8:1 to 4:1 in some series [78,173]. In the last French epidemiological survey, it was 2.7:1 [99]. This difference was not related to an increase in female exposure to neuromuscular blocking agents. It persists even if the gender ratio (1.1 female:1.0 male) of anaesthetized patients as established by the French survey of anaesthesia is taken into account [174]. Similarly, a female predominance was observed with respects to latex sensitization. This does not, however, imply any need for systematic allergy investigation in females before anaesthesia. Age Children who have undergone many operations, in particular those suffering from spina bifida, are considered at an increased risk for latex sensitization [6,10,150-152,175]. This increased risk has not been confirmed in adult patients repeatedly exposed to latex [176,177]. In addition, in the last French epidemiological survey of anaphylaxis during anaesthesia, a significant difference was observed regarding the distribution of anaphylaxis to latex according to age ranges. It was significantly different from those observed with neuromuscular blocking agents, with a higher incidence in the younger age ranges [99]. On the contrary, although the peak incidence of anaphylaxis to neuromuscular blocking agents was observed in the fourth decade in females and in the fifth decade in males, anaphylaxis was reported both in young and in elderly patients. Atopy Atopy is a hereditary predisposition in which subjects synthesize IgE antibodies to various allergens introduced into the body via natural routes. It has long been considered a risk factor for sensitization to muscle relaxants, in light of the high number of atopic patients found in early studies of anaphylactic shock during anaesthesia, when atopy was defined on clinical grounds alone. However, when confirmed by specific immunological tests, atopy does not appear to be a significant risk factor for muscle relaxant sensitivity [91,93,178]. A history of atopy and/or asthma has a very low specificity and sensitivity is a predictor of anaphylactic reactions. Moreover, it is considered to have an unacceptably high false-alarm rate [103,179]. Nevertheless, one should bear in mind the fact that basophils of atopic patients release histamine more readily [180,181]. As a consequence