Severe anaphylaxis caused by intravenous anti‐cancer drugs

Abstract Background The incidence and risk factors of severe anaphylaxis by intravenous anti‐cancer drugs are unclear, whereas those of milder reactions have been reported. Study Design Electronic medical charts of cancer patients who have undergone intravenous chemotherapy between January 2013 and October 2020 in a university hospital were retrospectively reviewed. Non‐epithelial malignancies were also included in the analysis. "Severe anaphylaxis" was judged using Brown's criteria: typical presentation of anaphylaxis and one or more of hypoxia, shock, and neurologic compromise. (UMIN000042887). Results Among 5584 patients (2964 males [53.1%], 2620 females [46.9%], median age 66 years), 88,200 person‐day anti‐cancer drug administrations were performed intravenously, and 27 severe anaphylaxes were observed. The causative drugs included carboplatin (14 cases), paclitaxel (9 cases), and cisplatin, docetaxel, trastuzumab, and cetuximab (1 case each). The person‐based lifetime incidence of severe anaphylaxis for patients who received at least one intravenous chemotherapy was 0.48% (27/5584, 95% confidence interval (CI) 0.30%–0.67%) and the administration‐based incidence was 0.031% (27/88,200, 95% CI 0.019%–0.043%). Among 124 patients who received at least 10 carboplatin administrations, 10 patients experienced carboplatin‐induced severe anaphylaxis (10/124, 8.1%, 95% CI 3.0%–13.1%). Carboplatin caused severe anaphylaxis after at least 9‐min interval since the drip started. Thirteen out of 14 patients experienced carboplatin‐induced severe anaphylaxis within a 75‐day interval from the previous treatment. Paclitaxel infusion caused severe anaphylaxis after a median of 5 min after the first drip of the day at a life‐long incidence of 0.93% (9/968, 95% CI 0.27%–1.59%). Conclusion We elucidated the high‐risk settings of chemotherapy‐induced severe anaphylaxis.


| BACKGROUND
Anaphylaxis is an acute, severe, and potentially lifethreatening allergic reaction that is caused by a variety of foreign substances. Food, venom, and medications including beta-lactam antibiotics and non-steroidal antiinflammatory drugs are common triggers of anaphylaxis. 1,2 Intravenous anti-cancer drugs such as platinum agents and taxanes also often induce hypersensitivity reactions. [3][4][5] A patient with mild allergic reaction may have localized rash, itchiness, or rhinitis. 1,2 However, severe forms of druginduced anaphylaxis are iatrogenic critical events that demand life-supporting maneuvers. 6 Therefore, severe anaphylaxis should be paid careful attention. However, few studies have focused on chemotherapy-induced severe anaphylaxis, except for case reports. 7 Despite the recommended use of dexamethasone and H1/H2 antihistamine prophylactic pretreatment prior to chemotherapies, 5,8 some patients still experience severe anaphylaxis. Therefore, we need to know the incidence and risk factors of severe anaphylaxis caused by intravenous anti-cancer drugs to prepare for acute-onset critical events and to design novel regimens.
In 1999, Markman et al. studied 205 patients who were treated with carboplatin, observing that 24 (12%) of them developed hypersensitivity reactions of various severities. 8 Their observation also demonstrated that allergic reactions mainly developed at the 8th or later carboplatin treatment. 8 Another well-known risk factor is carboplatin infusion after a 1-year or longer interval. [9][10][11] In contrast, based on a report by Weiss et al. in 1990, paclitaxel hypersensitivity reactions of any grade occurred in 27 of 301 patients (9%) and caused mainly by the first or second exposure to paclitaxel. 12 However, it is not clear whether these rules can be applied for severe anaphylaxis because limited numbers of severe cases were reported in these studies. Furthermore, recently recommended prophylactic premedication might modify the epidemiology. In addition, numerous new anti-cancer agents, especially molecular-targeted drugs, have become available in the last two decades. Therefore, a comprehensive updated risk assessment is necessary to elucidate the incidence and high-risk situations of chemotherapy-induced severe anaphylaxis. The aim of this study was to assess the incidence of severe anaphylaxis caused by anti-cancer medications and to elucidate high-risk situations.

| Overview
This was a single-center retrospective chart review study registered with the identification number UMIN000042887 and was approved by the Yokohama City University Hospital Institutional Ethical Review Board (ID: B201200069). Instead of informed consent, chance of opt out was offered for patients; however, none of the patient claimed opt out for this retrospective study.

| Data collection
The electronic medical charts of the Yokohama City University Hospital were retrospectively reviewed. In this hospital, prophylactic medications were routinely provided prior to high-risk agents, such as carboplatin and paclitaxel. All out-patients and in-patients who underwent intravenous chemotherapy for any malignancy between 1 January 2013 and 30 October 2020 were identified. Since our study focused on intravenous anti-cancer medications, anti-cancer drugs though other routes, such as oral, hypocutaneous, or intra-arterial administration were excluded from our analysis. Non-epithelial malignancies such as sarcoma and leukemia were also included in our analysis.
Background data such as age, sex, type of malignancies, and chemotherapy regimens were systematically extracted. Charts of patients who were given the International Classification of Disease Code (9th or 10th edition) of anaphylaxis, anaphylactic shock, or shock on the day of the chemotherapy were scrutinized in detail.

| Diagnosis of severe anaphylaxis
"Severe anaphylaxis" was judged using Brown's criteria, 6 typical presentation of anaphylaxis and one or more of hypoxia with saturation of oxygen by pulse oximetry (SpO 2 ) of 92% or lower, hypotension with systolic blood pressure below 90 mmHg, and neurologic compromise such as confusion, collapse, loss of consciousness, and incontinence. This criterion annotated 139 severe anaphylaxis cases (12.2%) out of 1149 patients with systemic hypersensitivity reactions due to a variety of triggers who Conclusion: We elucidated the high-risk settings of chemotherapy-induced severe anaphylaxis.

K E Y W O R D S
anaphylaxis, drug hypersensitivity, medical oncology, retrospective studies visited or were transported to an emergency department. 6 When an anti-cancer drug was intravenously started and then led to an allergic reaction within 30 min, the drug was considered to be the causative agent for our analysis. 1,8 Allergic reactions to monoclonal antibodies are often termed infusion reactions as distinguished from anaphylaxis. However, as long as the clinical presentation satisfied Brown's criteria, we counted allergic reaction caused by the monoclonal antibody as severe anaphylaxis in our analysis, following recent guidelines. 13

| Statistics
We calculated two types of incidences in this study. For administration-based incidence, one person-day treatment was considered as one administration. For example, when two cycles of gemcitabine (days 1, 8, and 15) were provided to a patient, this patient is considered to have 6 day-persons of gemcitabine treatment regardless of the cycle count. On the other hand, person-based lifetime incidence was calculated regardless of the total number of administrations for each patient. The Agresti-Coull method was used to estimate the 95% confidence interval (CI) of incidence, and the Mann-Whitney test was used to compare nonparametric numbers in two groups.
The most frequently used drug, fluorouracil, was administered 14,893 times to 1066 individual patients. Drugs infused more than 2000 times such as gemcitabine, cisplatin, and solvent-based paclitaxel are listed in Table 2.
The first sign or symptom of carboplatin-induced anaphylaxis was observed at a median of 17.5 min (IQR 12.5-24.75) after the first carboplatin drip of the day (

T A B L E 3 (Continued)
as paclitaxel, it took longer interval from the first drip of the day to the reaction onset (p < 0.001, Figure 1B). The interval from the previous carboplatin treatment was then analyzed. A median interval from the previous carboplatin administration to the administration leading to the event was 28 days. Thirteen out of 14 patients experienced carboplatin-induced severe anaphylaxis within a 75day interval, whereas one patient experienced severe event at an interval of 295 days. There was no severe anaphylaxis observed after a 360-day or longer interval (Table 3).
No case of severe anaphylaxes was induced by the other platinum treatments in this study, including oxaliplatin (4321 administrations, 694 patients, Table 2) and nedaplatin (505 administrations, 153 patients).

| Paclitaxel and taxanes
Intravenous solvent-based paclitaxel was infused 8894 times in 1034 patients ( Table 2). Paclitaxel induced nine severe anaphylaxes including six cases of shock. Seven of the nine patients had severe anaphylaxis during the first administration of paclitaxel. The second and third administrations caused one severe event each (Table 3). Because 66 of the 1034 patients in our database received their first paclitaxel treatment before January 2013, the remaining 968 patients in this database received their first paclitaxel administration during our observation period. Among them, the incidence of paclitaxel-induced severe anaphylaxis during the first administration was 0.72% (7/968, 95% CI 0.12%-1.32%) ( Figure 2B). For this population, the incidence of lifelong paclitaxel anaphylaxis during the first up to the third treatment was 0.93% (9/968, 95% CI 0.27%-1.59%).
The median duration from the first paclitaxel drip of the day to the first sign or symptom of anaphylaxis was 5 min (range 1 to 15, Table 3).
Nanoparticle albumin-bound paclitaxel (nabpaclitaxel) was notably being increasingly administered, with 87 instances in 2013 and 718 in 2019. This drug was administered 3693 times in 434 cases, none of whom had severe anaphylaxis (Table 2).  Table 2). The first-course docetaxel caused severe anaphylaxis in a 60-year-old gastric patient 5 min after the initiation of the infusion (Table 3).
No cases of severe anaphylaxis were caused by cabazitaxel, a recently developed taxane, which was used in 13 patients.

| HER2 monoclonal antibody
HER2 monoclonal antibody, trastuzumab, was administrated 2903 times for 166 patients who had breast cancer, Paget's disease, or gastric cancer ( Table 2). A 55-year-old woman with breast cancer who did not have any prophylactic premedication suffered from anaphylactic shock with desaturation 14 min after trastuzumab infusion (Table 3). Intra-muscular adrenaline was shot within a minute since blood pressure dropped below 90 mmHg. She was recovered from shock in 2 min. She was safely retreated with trastuzumab with prophylactic dexamethasone, diphenhydramine, and ranitidine 10 days after this event and successfully underwent total 45 doses of trastuzumab. She was the only patients who received the same drug after the severe anaphylaxis in our institute.

| EGFR monoclonal antibody
Cetuximab, an EGFR monoclonal antibody, was infused 5240 times in 249 patients with colon, oral, or head and neck cancers ( Table 2). A 71-year-old male with head and neck carcinoma experienced a severe anaphylactic event 8 min after intravenous treatment with cetuximab (Table 3).
All patients were successfully treated with combinations of oxygen, hydration, intramuscular adrenaline, hydrocortisone, chlorpheniramine, or famotidine, and no one died. Intravenous aminophylline, inhaled salbutamol, and nebulized salbutamol were also used in a few patients. Intra-muscular adrenaline was provided in 11 patients, of whom eight had active shock. Four of the eight patients who received an adrenaline shot recovered from shock within 5 min whereas each of the other four cases took 8, 12, 19 for the shock recovery, and 56 min, respectively.

| DISCUSSION
Although many studies have warned about anaphylaxis caused by anti-cancer drugs, 7,8,14 the risk factors for chemotherapy-induced severe anaphylaxis remained unclear. This is partly because most studies on this topic have been published as case repots, which are unsuitable for estimates of incidence and risk factors. By contrast, retrospective database analysis is a reasonable strategy for this clinical question given the low incidence of severe anaphylaxis. We analyzed 88,200 person-day treatments for 5584 patients and clarified that the repeated carboplatin infusion and the first few administrations of solvent-based paclitaxel are distinct high-risk situations for severe anaphylaxis, despite the use of prophylactic medications. In addition to carboplatin and taxanes, numerous cytotoxic agents have been frequently used (Table 2); however, most of these agents did not cause severe events at our hospital. In a few cases, monoclonal antibodies were the causative agents.
Carboplatin is believed to induce IgE-mediated type I hypersensitivity. 15,16 It typically occurs more than 10 min after the first drip of the 11th to 20th course of the drug ( Table 3, Figures 1 and 2). This finding was compatible with numerous previous reports about carboplatin hypersensitivity of any grade. 3,[7][8][9] However, the interval between the previous carboplatin administration and severe anaphylaxis remains controversial. O'Cearbhaill et al. reported that the median interval between prior platinum regimen and hypersensitivity reaction was 20 months. 9 Gadducci et al. assessed 69 patients of whom nine experienced anaphylactic events and noted that carboplatin re-treatment after an interval longer than 23.4 months was a strong risk for hypersensitivity of any severity (odds ratio 7.1, p = 0.013). 10 A database analysis by Schwartz et al. involving 126 patients suggested that a prolonged platinum-free interval longer than 24 months substantially increases severe carboplatin hypersensitivity of any severity (p = 0.009). 11 Nonetheless, no patient in our hospital experienced severe anaphylaxis caused by carboplatin after a 1 year interval or longer (Table 3). By contrast, 13 out of 14 patients in our study had events within 75 days from the previous carboplatin treatment (Table  3). Although we could not detect a definite clue for this discrepancy, the high-risk situations of severe anaphylaxis might be different from those of milder hypersensitivity. In any case, a patient with repeated carboplatin infusion should be closely monitored even after a short interval from the last treatment. Various protocols have been advocated for patients who experience carboplatin hypersensitivity. 17,18 However, re-administration of the trigger medication that caused severe anaphylaxis cannot be justified. 4,7,19 Interestingly, only gynecological patients experienced carboplatin-induced severe anaphylaxis in our study, despite carboplatin being commonly used for a variety of cancers. This could be because 6th-or lattercarboplatin administration was the risk factor for severe anaphylaxis and 6th-or latter-carboplatin was mainly provided for gynecological patients (Figure 2A).
In our study, the first to third administration of taxanes caused severe anaphylaxis whereas fourth-or latter-infusion did not, and most of these taxanes-induced severe anaphylaxes occurred within 5 min. These findings consistent with those of previous studies regarding hypersensitivity of any grade. 8,14 Although some researchers mentioned that paclitaxel hypersensitivity is milder than that caused by carboplatin, 5 we should do recognize that paclitaxel-induced severe anaphylaxis is not rare ( Table 3). Paclitaxel re-administration may be considered once mild symptoms subside. 8,20 However, paclitaxel re-administration after severe anaphylaxis seems too risky. 4,7,19 Polyethylated castor oil solvent of paclitaxel is known to directly cause hypersensitivity via the non-IgE mechanism. 5,21,22 By contrast, nab-paclitaxel is believed to rarely cause hypersensitivity. 5,21,22 There was no nab-paclitaxelinduced severe anaphylaxis in our hospital, although this should be confirmed through another larger database.
Monoclonal antibodies can cause acute hypersensitivity, including anaphylaxis mediated by IgE, similar to anaphylactoid reactions, infusion reactions, and cytokine release syndrome. 23 Cetuximab, a chimeric mousehuman monoclonal antibody against EGFR is known to induce anaphylaxis when IgE antibodies target galactose. 23,24 A post-marketing survey of trastuzumab, a humanized anti-HER2 monoclonal antibody, covered 25,000 cases identified 74 patients (0.3%) with severe infusion-related events mostly occurring shortly after infusion. Nine events leading to death were induced by the first administration of trastuzumab. 25 Although some might consider that hypersensitivity reaction to monoclonal antibody is not anaphylaxis but an infusion reaction, two of our cases presented typical symptoms of severe anaphylaxis and had excellent reaction to adrenaline (Table 3, cases 19 and 20). 1,13 As long as medical staff are prepared for severe anaphylaxis, it is easy to diagnose because most patients present with sudden-onset typical symptoms during intravenous treatment (Table 3). In addition, the first-choice medical therapeutic option, intramuscular epinephrine (0.3 mg = 0.3 ml of 0.1%), is effective for most patients. 2,26 Nonetheless, experts strongly believe that epinephrine is underused despite the recommendations of guidelines. 2,26 Most clinicians are used to administering 1 mg intravenous epinephrine for cardio-pulmonary arrested patients; 27 however, they usually do not have enough chance to be trained administering intramuscular epinephrine for anaphylactic patients. Doctors and nurses who provide intravenous anticancer drugs should know the risk situation illustrated in our analysis and be familiar with intramuscular adrenaline shots to deal with iatrogenic critical events.
The limitations of our study include the low evidence level due to the retrospective design, inconsistent prophylaxis regimens, and possibly omitted events 30 min after the first drip.
In conclusion, we conducted a retrospective database analysis of 88,200 person-days of treatment for 5584 patients. Repeated carboplatin and the first few administrations of paclitaxel and docetaxel are high-risk situation. The lifelong incidence of severe anaphylaxis was as high as 8.1% among patients who received at least 10 carboplatin administrations. Carboplatin typically caused severe anaphylaxis after 10-min or longer interval since the first drip. Long interval since the last carboplatin treatment was not a risk of severe anaphylaxis. Paclitaxel-induced severe anaphylaxis was typically caused by the first administration within 5 min from the first drip and the lifelong incidence was 0.93%. Monoclonal antibodies also caused a few severe anaphylaxes.