Symptomatology and Management of Adult Anaphylaxis according to Trigger: A Cross-Sectional Study

Anaphylaxis is an acute life-threatening allergy, most commonly provoked by food, venom, or drugs [1, 2]. The diagnosis of anaphylaxis is currently based on clinical criteria [3]. It is estimated that 1.6% of the US population has experienced anaphylaxis. Causes of anaphylaxis vary depending on the regions of the world as some studies have demonstrated venom to be the principal trigger in anaphylaxis [4], and others, drugs [5]. Data regarding differences in symptomatology of anaphylaxis related to the various triggers are limited, especially in adults [6].

Early recognition of anaphylaxis in relation to a patient’s history is critical to improve outcomes [7]. A German study [5], based on the European Anaphylaxis registry, explored the variation in symptomatology and anaphylaxis based on different triggers in both adults and children. The findings revealed that respiratory symptoms were more prominent in cases of food-induced anaphylaxis, while a Korean study [8] highlighted an association between drug-induced anaphylaxis (DIA) and cardiovascular symptoms. However, there is a gap in understanding how these patterns manifest in a Canadian context, where demographic and healthcare system differences may influence symptom presentation and management. For this reason, we aimed to assess whether symptoms, comorbidities, and medication use differ between those presenting with the most common triggers of anaphylaxis in the adult population.

Data for this study were from the Cross Canada Anaphylaxis Registry (C-Care), an extensive database of patients who have experienced anaphylaxis across three Canadian provinces including 8 emergency departments (EDs) and one electronic medical service (EMS) [9]. Anaphylaxis was defined as a systemic reaction involving at least two organ systems and/or hypotension according to the Food Allergy and Anaphylaxis Network/National Institute of Allergy and Infectious Disease criteria [10]. We collected data prospectively and retrospectively through a standardized data entry form [11]. Informed written consent was obtained for participating patients. For prospectively recruited patients, the treating ED physician along with a trained member of the research team identified cases of anaphylaxis and, after obtaining consent, completed the data entry form. Retrospective cases of anaphylaxis were identified through chart review of all ED visits based on a previously validated algorithm using International Classification of Diseases, Tenth Revision (ICD-10) codes [12]. Anaphylaxis was defined as the involvement of two or more organ systems after exposure to a possible allergen, or hypotension after exposure to a known allergen. Patients who did not meet the definition of anaphylaxis did not satisfy the inclusion criteria and were thus not eligible for recruitment into the registry [2, 13]. Data were collected regarding demographics (age and sex), types of triggers (food, drug, venom, unknown, and others), the severity of symptoms (mild, moderate, and severe), presence of comorbidities (asthma, exercise, environment, food allergy) as well as the regular use of medications such as non-steroidal anti-inflammatory drugs (NSAIDs) and beta-blockers. The severity of anaphylaxis was classified as mild, moderate, or severe, using a modified Muraro et al. [14] grading system. Mild anaphylaxis included nasal congestion/sneezing, rhinorrhea, throat tightness, mild wheezing, generalized pruritus, flushing, urticaria, angioedema, mild abdominal pain, nausea/vomiting, tachycardia, and anxiety [15]. Moderate anaphylaxis included harsh cough, moderate wheezing, dysphagia, shortness of breath, crampy abdominal pain, repeated vomiting, diarrhea, and light-headedness. Severe anaphylaxis included cyanosis, respiratory arrest, hypotension, circulatory collapse, dysrhythmia, bowel incontinence, cardiac arrest or serious bradycardia, confusion, and unconsciousness [16].

In Quebec, adults were recruited from 4 EDs involved in the C-Care for adults: the Montreal General Hospital, Montreal Children’s Hospital, Royal Victoria Hospital, and Hôpital du Sacré-Coeur de Montreal. Adults managed by Quebec’s Outaouais region emergency medical services were also included. In addition, adults presenting with anaphylaxis to St Joseph’s Healthcare Hamilton, London Health Sciences center in Ontario and the Foothills Medical Center in Alberta were also included in the C-Care dataset. Written consent was obtained by participants to participate in this study.

Descriptive statistics were used to present demographics, reaction characteristics, and medication use subcategorized depending on the anaphylaxis trigger. All statistical data were analyzed using R 4.2.3 binary for macOs 10.3 (R Foundation for Statistical Computing, Vienna, Austria). Categorical data were presented as percentages, and continuous data as a median with interquartile range. Univariate and multivariate regression models were used to evaluate symptoms involving all patients with the outcome of DIA, venom-induced anaphylaxis (VIA), peanut-induced anaphylaxis (PIA), shellfish-induced anaphylaxis, tree-nut induced anaphylaxis (TIA) and nut-induced anaphylaxis (NIA). The term NIA was used when the patient did not specify whether tree nuts or peanuts were the trigger. We assessed comorbidities associated with severe reactions, stratified by triggers listed above and evaluated the association of each trigger with treatment through regression models involving all patients with the medications used as outcome and triggers as independent variables. To determine factors associated with each of the triggers listed above, univariate logistic regression was performed for variables of interest. Variables with p < 0.10 were retained. A bidirectional stepwise logistic regression was performed on the remaining variables using the MASS package, which optimizes the model based on the Akaike information criterion [17]. Eventually, in multivariable analyses, a p value <0.05 was considered significant.

This study was approved by the McGill Ethics Board with IRB: 10–203 Glen. This study was approved by the Ethics Board of participating hospitals, including the Hamilton Integrated Research Ethics Board, Schulich School of Medicine and Dentistry Ethics Board, Sacré-Coeur Research Ethics Board and the Health Ethics Research Board of Alberta.

From April 2011 to November 2023, 1,135 adults presenting with anaphylaxis to EDs were recruited. Among these patients, the median age was 35.5 (interquartile range 25.3–51.1) and 38.4% were males (Table 1). The median age was highest in patients with soy induced anaphylaxis, being 52.1 years (41.2–56.3), followed by VIA with 51.8 years (39.5–58.1). The median age was lowest in patients with milk induced anaphylaxis at 23.6 (20.6–29.9) followed by PIA 24.6 (19.6–33.3). Participating adults had various other comorbidities, including asthma (15.1%). Common symptoms included pruritus (61.8%), urticaria (55.3%), and throat tightness (54.4%) (Table 2). Among all anaphylaxis cases, 23.7% presented with no cutaneous symptoms. Anaphylaxis was graded as mild in 7.6%, moderate in 73.4%, and severe in 18.9% (Table 2). Most of the patients (53.7%) presented with food-induced anaphylaxis (FIA), 20.2% presented with DIA, and 6.3% presented with VIA. Among food-induced reactions, the most common causes were peanut (13.5%), shellfish (10.2%), tree-nut (8.7%), and undefined nuts (4.1%). Other less common triggers of anaphylaxis included fish, milk, egg, soy, and wheat (Table 1). Within DIA, over 10.5% of drugs were antibiotics, with amoxicillin (4.7%) being the most common. The most common contact routes with drugs were ingestion for 77.3% of DIA and parenteral for 18.8% of reactions. Our cohort had 168 admissions to the hospital following anaphylaxis. DIA accounted for at least 35.7% of these admissions. Our cohort had 18 admissions to the ICU, with DIA accounting for 44.4% of these admissions.

When adjusting for sex and age at reaction, hypotension was more likely associated with DIA (aOR = 1.20, 95% CI = 1.11–1.30, p < 0.01) and VIA (aOR = 1.08, 95% CI = 1.03–1.13, p = 0.042) (Table 3). Hypotension was more likely associated with intravenous (IV) DIA (aOR = 2.50, 95% CI = 1.32–4.41, p = 0.043). Throat tightness was more likely associated with TIA (aOR = 1.04, 95% CI = 1.01–1.06, p < 0.01) (Table 3).

In our cohort, 216 patients (18.9%) had severe reactions. The highest prevalence of severe reactions was in DIA (30.1%), VIA (27.8%), and TIA (18.9%) whereas the lowest prevalence of severe reactions was in NIA (10.6%) and PIA (14.6%) (Table 2). When adjusted for age at reaction and male sex, alcohol was more likely associated (aOR = 1.51, 95% CI = 1.04–2.19, p = 0.035) with severe undefined NIA. Exercise was not significantly associated with severe reactions in any of the triggers studied. No other significant associations were found between known comorbidities such as alcohol consumption and asthma with other triggers (Table 4).

In the outpatient setting, 27.8% of patients received epinephrine auto-injectors (EAI), 42.1% received antihistamines and 8.7% received corticosteroids. In the hospital setting, 43.3% received EAI, 73.7% received antihistamines and 74.8% received corticosteroids (Table 5). When adjusted for age at reaction and male sex, PIA was more likely associated with outpatient EAI (aOR = 3.25, 95% CI = 2.03–5.20, p < 0.01). TIA was more likely associated with inpatient EAI (aOR = 2.05, 95% CI = 1.16–3.64, p = 0.014). DIA was less likely associated with outpatient antihistamine (aOR = 0.68, 95% CI = 0.48–0.89, p < 0.01) whereas TIA was more likely associated with outpatient antihistamine (aOR = 1.81, 95% CI = 1.03–3.19, p < 0.01). PIA was less likely associated with inpatient antihistamine (aOR = 0.46, 95% CI = 0.29–0.74, p < 0.01). Similarly, TIA was less likely associated with inpatient antihistamine (aOR = 0.48, 95% CI = 0.26–0.83, p < 0.01) (Table 6).

To our knowledge, we describe the largest cohort of adult anaphylaxis cases to date and the clinical characteristics of those reactions. Whereas other studies have found drugs to be the most common cause of anaphylaxis in adults [18, 19], we found food, mainly peanut, to be the major culprit. It is possible that increased prevalence of food allergy worldwide accounts for our findings [20]. Additionally, we found peanuts to be the most common trigger within FIA, followed by shellfish and tree-nuts. Our findings contrast with the finding of shellfish and wheat being the most common causes of FIA in adults in other studies [21]. These differences might be related to differences in sociodemographic (e.g., differences in age) and clinical characteristics (e.g., difference in the presence of comorbid conditions) between our population and populations assessed in other studies [22].

Multiple studies of VIA in adults have found a significant association between VIA and cardiovascular symptoms [23, 24]. This is in line with our findings, as we have demonstrated a positive association between hypotension and VIA and DIA. Indeed, it was reported that elderly patients with preexisting cardiovascular disease are more likely to die from VIA than FIA [25]. Similarly, studies suggest a high risk of hypotension with IV drugs versus oral or cutaneous exposure [26, 27] and with venom allergy versus other anaphylaxis triggers [28]. This association explains the higher prevalence of severe reactions among VIA and DIA when compared to other triggers.

We established a very weak but positive association between TIA and throat tightness. There is a paucity of data regarding this association. This association could potentially be explained by pollen food syndrome (PFS) as almond was the most common cause of TIA in our cohort, with 29.6% of patients presenting following almond consumption. Almonds have been reported to be associated with more severe forms of PFS [29] characterized by a severe reaction, involving the face and upper GI tract [30]. In our cohort, over 64.2% of TIA patients presented with angioedema, skin flushing and throat tightness, with absence of gastrointestinal manifestations, indicating possible PFS manifestations and 23.4% of reactions were severe. Further studies are needed to establish this possible association.

A study by Shaker et al. [31] reports exercise and asthma to be risk factors for severe anaphylaxis. Another study by Worm et al. [5] shows mild exercise to be inversely associated with severe reactions while vigorous exercise was strongly associated with severe reactions. This differs from our findings as we have not found any association between exercise and asthma with severity of reactions in the triggers studied. We found a strong association between alcohol and severe reactions in nut-triggered anaphylaxis. Alcohol is a known cofactor for severe reactions. There are different hypothesis regarding the negative effect of alcohol on anaphylaxis. It was suggested that the vasodilation effect of ethanol can lead to a more severe shock from food allergy [32]. Additionally, it was reported that elevation of acetaldehyde following alcohol consumption could induce histamine release by airway mast cells [33] and may thus increase reaction severity. The specific association with nuts may be in relation to nuts commonly being used as flavoring options in alcoholic drinks, such as Amaretto, Amadeus, and Galliano. Although we do not know specifically which alcoholic drink the patients consumed, it is possible that the combination of nuts with the histamine-releasing effects of alcohol led to a more severe anaphylactic reaction. It is also possible that alcohol is a cofactor that leads to a more severe form of PFS following tree-nut ingestion.

PIA was positively associated with outpatient EAI. This association can be explained by higher awareness of peanut allergy and the use of EAI among patients with peanut allergy [34].

Additionally, we found that TIA was positively associated with inpatient EAI. To our knowledge, there are currently no publication reporting this association. However, we had previously determined a positive association between tree-nut and upper respiratory symptoms, making the diagnosis of anaphylaxis more evidence, hence increasing use of first-line management of EAI.

Our study has strengths as it is a cross-sectional study including over 8 EDs and one EMS across Canada with a large sample size of adults with anaphylaxis that were confirmed to fulfill criteria by two independent reviewers. However, it also has potential limitations. First, this is retrospective data based on chart review. In addition, data are collected only from EDs, leading to a lack of data regarding less severe anaphylactic reactions. Adults with mild/moderate anaphylaxis were less likely to present to EDs than those with severe anaphylaxis [33], thus biasing towards inclusion of more severe reactions.

In conclusion, our study presents a large cross-sectional cohort of adult anaphylaxis in North America, highlighting food, particularly peanut, as the major cause. While our findings highlight associations between specific anaphylaxis triggers and symptom presentation, these associations do not imply a universal early warning system for diagnosing anaphylaxis. Therefore, the role of validated diagnostic criteria such as those from Sampson (2005) or the World Allergy Organization (2020) remains crucial in guiding swift diagnosis and appropriate treatment. It is important that clinical vigilance remains focused on the comprehensive assessment of each patient, including rapid administration of epinephrine where indicated.

Author Jocelyn Moissan was not available to confirm co-authorship, but the corresponding author Roy Khalaf affirms that author Jocelyn Moissan contributed to the paper, had the opportunity to review the final version to be published and guarantees author Jocelyn Moissan’s co-authorship status and the accuracy of the author contribution and conflict of interest statements.

This study was approved by the McGill Ethics Board with IRB: 10–203 Glen. Written informed consent was obtained for participation in this study. Members of Ethics Board: McGill Ethics Board: Dr Carlo Cicero, Ms Elizabeth Craven. HiREB (Hamilton Integrated Research Ethics Board). Sacré-Coeur de Montréal Research Ethics Board. Schulich School of Medicine and Dentistry Ethics Board: Dr. Jonathan Dreyer, Dr. Gauri Ghate, Dr. Marcia Edmonds. Health Ethics Research Board of Alberta: Dr. Albert Mehl.

The authors do not have any conflict of interest to declare.

This study was not supported by any sponsor or funder.

R.K., C.P., and M.B.-S. wrote the manuscript. M.B.-S. was the principal investigator, guiding the manuscript, gathering the data and reviewing the manuscript. C.M., A.B., M.K., A.E.C., J.M., R.L., E.S.C., R.D.G., A.O., J.G., D.K.C., J.U., E.H., J.M., X.Z., J.L.P.P., E.A., E.S., and J.R. helped with data collection and reviewed the manuscript.

Edited by: H.-U. Simon, Bern.

Data involved in this study are not publicly available as their containing information could compromise the privacy of research participants. Further inquiries can be directed to the corresponding author.