Antibiotics Shuaib Nasser Cambridge University Hospitals NHS Foundation Trust NAP6 Steering Committee member

Key Findings Antibiotics most common cause of perioperative anaphylaxis in UK 35% of all cases (NMBAs ≈25% of cases) Antibiotic anaphylaxis in 5.15 per 100,000 patients /yr receiving at least one antibiotic perioperatively 3.96 per 100,000 antibiotic administrations /yr develop antibiotic anaphylaxis Co-amoxiclav caused17.3% and teicoplanin 13.5% of the 35% total ie 89% of all cases Perioperative anaphylaxis incidence: Teicoplanin -16.39 per 100,000 administrations /yr, ie16-fold that of cefuroxime Co-amoxiclav -8.64 per 100,000 administrations /yr ie 9-fold that of cefuroxime vancomycin was 3rd and piperacillin-tazobactam 4th place - distant Most common first clinical feature was hypotension - 42% of all antibiotic cases presented within 5 min of exposure in 76% of cases Two fifths of the patients with a pre-operative history of penicillin allergy reacted to teicoplanin Allergy clinics did not identify the antibiotic culprits in 25% cases due to incomplete investigations ie skin tests and drug provocation challenges In two thirds of cases inappropriate advice on future avoidance was given by allergy clinics Allergy clinics may be underdiagnosing antibiotic allergy, potentially placing patients at risk NMBAs, second leading cause, ≈25% of cases The onset of anaphylaxis was within 5 minutes in 74% of cases, within 10 minutes in 92% and in all cases within 30 minutes Previous history of penicillin allergy appears to be a risk factor for anaphyalxis to Teicoplanin and therefore: This evidences the need for robust programmes to investigate and de-label, where appropriate, patients with reported history of penicillin allergy.

Antibiotic allergy reported pre-operatively Allergy to Penicillin Cephalosporin Other antibiotic 45 3 4 2 10 total 52 7 16 73 patients had pre-operative label of antibiotic allergy: 52 to penicillins (49 penicillin, 2 amoxicillin, 1 piperacillin-tazobactam), 3 3 also had a label of cephalosporin allergy. 7 had a label of cephalosporin allergy 16 - allergy to a variety of antibiotics including trimethoprim, co-trimoxazole, erythromycin, metronidazole, doxycycline or tetracycline. Four of these also had a label of penicillin allergy. One patient had a label of multiple antibiotic allergy

Demographics –antibiotic allergy Only 12%<35yrs of age

Antibiotic culprits

Time interval from IV injection to first clinical feature The onset of anaphylaxis was within 5 minutes in 74% of cases, within 10 minutes in 92% and in all cases within 30 minutes

First Clinical Feature

Clinical features These are discussed in Chapter 10 Clinical features. The most common first clinical feature (42%) was hypotension followed by bronchospasm/high airway pressure (15%) and tachycardia (13%). During teicoplanin anaphylaxis hypotension was a dominant presenting feature with bronchospasm uncommon (Figure 5).   Considering clinical features present at any time during the episode, hypotension was universal and blood pressure was unrecordably low in a quarter of cases. Flushing/non-urticarial rash, bronchospasm/high airway pressure and tachycardia were the next most common features (67%, 53% and 50%, respectively). Bradycardia was present in 11% of cases. Table 6.

Grade of anaphylaxis for all antibiotics Total Co-amoxiclav 24 21 1 46 Teicoplanin 18 16 2 36 Cefuroxime   4 Gentamicin 3 Flucloxacillin Piperacillin- tazobactam Metronidazole Vancomycin 48 43 94 *Two cases where teicoplanin and gentamicin were joint probable causes therefore 92 patients but 94 antibiotics Severity There were 46 (50%) grade 3 and 43 (47%) grade 4 reactions. Three (3%) cases were fatal of which two were due to teicoplanin and one co-amoxiclav.

Allergy clinic diagnosis of 92 cases of antibiotic anaphylaxis 92 cases of antibiotic-induced anaphylaxis were identified by the review panel Allergy clinics considered 69 cases (50 definite) to have been caused by an antibiotic. But in some cases a single culprit was not confirmed and >2agents were recommended for avoidance In 12 cases the review panel considered that an oral challenge had been omitted by the allergy clinic In another 11 cases the review panel suggested IV drug provocation should have been considered In 61 /92 cases(66%) of antibiotic allergy identified by the review panel appropriate advice on future avoidance was not provided by the allergy clinic either not given, inappropriate due to incomplete investigation of all culprits and/or no culprit identified, no safe alternatives clearly stated, excessive avoidance advice (e.g. multiple antibiotics) based on incomplete investigations Ie. allergy clinics may be underdiagnosing antibiotic allergy potentially placing patients at risk Culprit established by the review panel Allergy clinic diagnosis of antibiotic allergy with high certainty Allergy clinic diagnosis of antibiotic allergy with intermediate certainty or certainty not stated Not investigated by Allergy clinic or culprit not identified Antibiotic   number % of cases established by the review panel Co-amoxiclav 46 24 52.2% 8 17.4% 14a 30.4% Flucloxacillin 2 100% 0% Piperacillin-tazobactam 1 Cefuroxime 4 50.0% Teicoplanin 36 19 52.8% 22.2% 9 25.0% Vancomycin Gentamicin 3 33.3% 66.7% Metronidazole Total 94 50 53.2% 20.2% 25.5% In two cases both tecicoplanin and gentamicin were judged equally probable as culprits so there were 94 definite or probable antibiotic culprits

Antibiotic use & relative incidence of anaphylaxis Exposure from activity survey Use in NAP6 Cases suspected by anaesthetist Allergic reactions in NAP6* Incidence   Relative incidence Estimated annual caseload Total no % of all cases cases per 100,000 administrations per year Cefuroxime = 1 Co-amoxiclav 532,580 71 26.7% 40 46 17.3% 8.64 9.16  Flucloxacillin 211,973 12 4.5% 6 2 0.8% 0.94 1.0 Piperacillin-tazobactam 28,237 4 1.5% 3 1 0.4% 3.54 3.76 Cefuroxime 424,143 16 6.0% Teicoplanin 219,621 62 23.3% 33 36 13.5% 16.39 17.38 Vancomycin 17,648 5.67 6.01 Gentamicin 616,899 56 21.1% 1.1% 0.49 0.52 Metronidazole 272,173 17 6.4% 0.37 0.39 Amoxicillin Data not collected 0% - Ceftriaxone Ciprofloxacin Clindamycin Meropenem Amikacin Total 2,323,274 256 92 (94)# 34.6% (35.3%) 3.96 (4.05) 4.20 (4.29) Patients receiving at least one antibiotic 1,787,360 5.15  5.46 The incidence of perioperative anaphylaxis to teicoplanin is considerably high at 16.39 per 100,000 administrations per year, which 16-fold that of cefuroxime. Co-amoxiclav was the antibiotic with the second highest risk of perioperative anaphylaxis at 8.64 per 100,000 persons per year, 9-fold that of cefuroxime

RECOMMENDATIONS Institutional Patients with reported allergy to a beta-lactam antibiotic and at least one other class of antibiotics should be referred for allergy investigation, before elective surgery, in line with NICE CG183: Drug allergy: diagnosis and management. If antibiotic allergy is suspected despite negative skin tests, challenge testing should be considered Broad beta lactam avoidance advice should be discouraged and patients should be further investigated to clarify the drug(s) to avoid and to identify safe alternatives. Individual Ninety per cent of anaphylaxis due to antibiotics presents within ten minutes of administration. When perioperative antibiotics are indicated they should be administered as early as possible, where practical at least 5-10 minutes before induction of anaesthesia, providing this does not interfere with their efficacy. The anaesthetist should consider co- amoxiclav or teicoplanin amongst the likely culprits when anaphylaxis occurs after their administration. Avoid test doses of antibiotic A test dose of antibiotic should not be used, as it will not prevent or reduce the severity of anaphylaxis. Administration of antibiotics several minutes before induction of anaesthesia will likely improve detection, may simplify treatment and will help investigation when reactions occur Teicoplanin was administered frequently because of a history of penicillin allergy. With the knowledge that penicillin allergy is unfounded in >90% of cases, effective de-labelling of penicillin allergy would decrease overall risk of anaphylaxis.

Key Findings Antibiotics most common cause of perioperative anaphylaxis in UK 35% of all cases (NMBAs ≈25% of cases) Antibiotic anaphylaxis in 5.15 per 100,000 patients /yr receiving at least one antibiotic perioperatively 3.96 per 100,000 antibiotic administrations /yr develop antibiotic anaphylaxis Co-amoxiclav caused17.3% and teicoplanin 13.5% of the 35% total ie 89% of all cases Perioperative anaphylaxis incidence: Teicoplanin -16.39 per 100,000 administrations /yr, ie16-fold that of cefuroxime Co-amoxiclav -8.64 per 100,000 administrations /yr ie 9-fold that of cefuroxime vancomycin was 3rd and piperacillin-tazobactam 4th place - distant Most common first clinical feature was hypotension - 42% of all antibiotic cases presented within 5 min of exposure in 76% of cases Two fifths of the patients with a pre-operative history of penicillin allergy reacted to teicoplanin Allergy clinics did not identify the antibiotic culprits in 25% cases due to incomplete investigations ie skin tests and drug provocation challenges In two thirds of cases inappropriate advice on future avoidance was given by allergy clinics Allergy clinics may be underdiagnosing antibiotic allergy, potentially placing patients at risk NMBAs, second leading cause, ≈25% of cases The onset of anaphylaxis was within 5 minutes in 74% of cases, within 10 minutes in 92% and in all cases within 30 minutes Previous history of penicillin allergy appears to be a risk factor for anaphyalxis to Teicoplanin and therefore: This evidences the need for robust programmes to investigate and de-label, where appropriate, patients with reported history of penicillin allergy.

Thank you