Anaphylactic and nonanaphylactic reactions to SARS-CoV-2 vaccines: a systematic review and meta-analysis

Background Currently there is no systematic review and meta-analysis of the global incidence rates of anaphylactic and nonanaphylactic reactions to SARS-CoV-2 vaccines in the general adult population. Objectives To estimate the incidence rates of anaphylactic and nonanaphylactic reactions after COVID-19 vaccines and describe the demographic and clinical characteristics, triggers, presenting signs and symptoms, treatment and clinical course of confirmed cases. Design A systematic review and meta-analysis. Preferred Reporting Items for Systematic Reviews and Meta-Analyses [PRISMA] statement was followed. Methods Electronic databases (Proquest, Medline, Embase, Pubmed, CINAHL, Wiley online library, and Nature) were searched from 1 December 2020 to 31 May 2021 in the English language using the following keywords alone or in combination: anaphylaxis, non-anaphylaxis, anaphylactic reaction, nonanaphylactic reaction, anaphylactic/anaphylactoid shock, hypersensitivity, allergy reaction, allergic reaction, immunology reaction, immunologic reaction, angioedema, loss of consciousness, generalized erythema, urticaria, urticarial rash, cyanosis, grunting, stridor, tachypnoea, wheezing, tachycardia, abdominal pain, diarrhea, nausea, vomiting and tryptase. We included studies in adults of all ages in all healthcare settings. Effect sizes of prevalence were pooled with 95% confidence intervals (CIs). To minimize heterogeneity, we performed sub-group analyses. Results Of the 1,734 papers that were identified, 26 articles were included in the systematic review (8 case report, 5 cohort, 4 case series, 2 randomized controlled trial and 1 randomized cross-sectional studies) and 14 articles (1 cohort, 2 case series, 1 randomized controlled trial and 1 randomized cross-sectional studies) were included in meta-analysis. Studies involving 26,337,421 vaccine recipients [Pfizer-BioNTech (n = 14,505,399) and Moderna (n = 11,831,488)] were analyzed. The overall pooled prevalence estimate of anaphylaxis to both vaccines was 5.0 (95% CI 2.9 to 7.2, I2 = 81%, p =  < 0.0001), while the overall pooled prevalence estimate of nonanaphylactic reactions to both vaccines was 53.9 (95% CI 0.0 to 116.1, I2 = 99%, p =  < 0.0001). Vaccination with Pfizer-BioNTech resulted in higher anaphylactic reactions compared to Moderna (8.0, 95% CI 0.0 to 11.3, I2 = 85% versus 2.8, 95% CI 0.0 to 5.7, I2 = 59%). However, lower incidence of nonanaphylactic reactions was associated with Pfizer-BioNTech compared to Moderna (43.9, 95% CI 0.0 to 131.9, I2 = 99% versus 63.8, 95% CI 0.0 to 151.8, I2 = 98%). The funnel plots for possible publication bias for the pooled effect sizes to determine the incidence of anaphylaxis and nonanaphylactic reactions associated with mRNA COVID-19 immunization based on mRNA vaccine type appeared asymmetrical on visual inspection, and Egger’s tests confirmed asymmetry by producing p values < 0.05. Across the included studies, the most commonly identified risk factors for anaphylactic and nonanaphylactic reactions to SARS-CoV-2 vaccines were female sex and personal history of atopy. The key triggers to anaphylactic and nonanaphylactic reactions identified in these studies included foods, medications, stinging insects or jellyfish, contrast media, cosmetics and detergents, household products, and latex. Previous history of anaphylaxis; and comorbidities such as asthma, allergic rhinitis, atopic and contact eczema/dermatitis and psoriasis and cholinergic urticaria were also found to be important. Conclusion The prevalence of COVID-19 mRNA vaccine-associated anaphylaxis is very low; and nonanaphylactic reactions occur at higher rate, however, cutaneous reactions are largely self-limited. Both anaphylactic and nonanaphylactic reactions should not discourage vaccination.


Background
Globally, as of 28 May 2021, there have been 168,040,871 confirmed cases of coronavirus disease 2019 (COVID- 19), including 3,494,758 deaths, reported to the World Health Organization [1]. Immunization is an important strategy to halt COVID-19 pandemic. Two COVID-19 vaccines were granted emergency use authorization [EUA] by the United States Food and Drug Administration: the Pfizer-BioNTech COVID-19 vaccine and the Moderna COVID-19 vaccine [2,3]. Other vaccines had also been used or authorized in multiple parts of the world. Vaccination hesitancy is an important obstacle to achieve herd immunity [4][5][6]. One of the main reasons for vaccine hesitancy is the concerns about vaccine safety, adverse effects, or toxicity.
Adverse reactions to vaccines are commonly reported, but most are not immunologically mediated. Nonimmunologically mediated reactions are typically related to the pharmacology of the vaccine and consist of toxic reactions, side effects, and medication interactions [7]. Immunologically mediated reactions include IgEmediated and T-cell mediated reactions as well as other immunologic mechanisms that occur as a result of allergen exposure. These responses, particularly severe anaphylaxis, rarely occur [8,9]. Symptoms can begin within minutes of exposure to the allergen [generally within 30 min and less commonly within several hours]; and range from urticaria, swelling, and gastrointestinal upset, to respiratory distress and cardiovascular collapse associated with severe anaphylaxis [9]. Delayed reactions typically occur hours to days after exposure to the antigen trigger, and most typically manifest as a cutaneous reaction [10].
Maintaining public confidence to minimize vaccine hesitancy is crucial as disease outbreaks among the unvaccinated population are likely to occur when herd immunity is compromised. We aimed to estimate the incidence rates of anaphylactic and nonanaphylactic reactions after mRNA COVID-19 vaccines (Pfizer/ BioNTech and Moderna) and describe the demographic and clinical characteristics, triggers, presenting signs and symptoms, treatment and clinical course of confirmed cases. We primarily focused this assessment on the Pfizer/BioNTech and Moderna mRNA vaccines given that reports of anaphylactic and nonanaphylactic reactions from adenovirus vaccines beyond clinical trials were limited.

Design
We followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines [PRISMA] in conducting this systematic review and meta-analysis [11]. The following electronic databases were searched: PROQUEST, MEDLINE, EMBASE, PUBMED, CINAHL, WILEY ONLINE LIBRARY, and NATURE with Full Text. We used the following keywords: anaphylaxis OR non-anaphylaxis OR anaphylactic reaction OR nonanaphylactic reaction OR anaphylactic/anaphylactoid shock OR hypersensitivity OR allergy reaction OR allergic reaction OR immunology reaction OR immunologic reaction AND angioedema OR loss of consciousness OR generalized erythema OR urticaria OR urticarial rash OR cyanosis OR grunting OR stridor OR tachypnoea OR wheezing OR tachycardia OR abdominal pain OR diarrhea OR nausea OR vomiting OR tryptase. The search was limited to papers was 53.9 (95% CI 0.0 to 116.1, I 2 = 99%, p = < 0.0001). Vaccination with Pfizer-BioNTech resulted in higher anaphylactic reactions compared to Moderna (8.0, 95% CI 0.0 to 11.3, I 2 = 85% versus 2.8, 95% CI 0.0 to 5.7, I 2 = 59%). However, lower incidence of nonanaphylactic reactions was associated with Pfizer-BioNTech compared to Moderna (43.9, 95% CI 0.0 to 131.9, I 2 = 99% versus 63.8, 95% CI 0.0 to 151.8, I 2 = 98%). The funnel plots for possible publication bias for the pooled effect sizes to determine the incidence of anaphylaxis and nonanaphylactic reactions associated with mRNA COVID-19 immunization based on mRNA vaccine type appeared asymmetrical on visual inspection, and Egger's tests confirmed asymmetry by producing p values < 0.05. Across the included studies, the most commonly identified risk factors for anaphylactic and nonanaphylactic reactions to SARS-CoV-2 vaccines were female sex and personal history of atopy. The key triggers to anaphylactic and nonanaphylactic reactions identified in these studies included foods, medications, stinging insects or jellyfish, contrast media, cosmetics and detergents, household products, and latex. Previous history of anaphylaxis; and comorbidities such as asthma, allergic rhinitis, atopic and contact eczema/ dermatitis and psoriasis and cholinergic urticaria were also found to be important.
published in English between 1 December 2020 and 31 May 2021. Based on the title and abstract of each selected article, we selected those discussing and reporting rates of anaphylactic and nonanaphylactic reactions to mRNA SARS-CoV-2 vaccines. We also utilized backward snowballing to increase the yield of potentially relevant articles.

Inclusion and exclusion criteria
We retained randomized controlled trials, cohort studies, case reports and case series and excluded other studies. We excluded studies that did not report data on anaphylactic and nonanaphylactic reactions to SARS-CoV-2 mRNA vaccines; or studies that never stated details on identified cases with anaphylactic and nonanaphylactic reactions. We evaluated studies that included all adults as our population of interest who received 1 or more doses of mRNA COVID-19 vaccines and reported evidence on anaphylactic and nonanaphylactic reactions during the period from December 1, 2020 through May 31, 2021.

Data extraction
Six authors (S.A., A.A., A.R., R.T., M.A.A., and Z.A.) critically reviewed all of the studies retrieved and selected those judged to be the most relevant. The abstracts of all citations were examined thoroughly. Data were extracted from the relevant research studies using key headings, which are noted in Tables 1 and 2, simplifying analysis, and review of the literature. Articles were categorized as randomized controlled trial, case report, case series or cohort studies. Studies were also categorized as a national or international online survey as we sought to describe reports of anaphylactic and nonanaphylactic reactions after mRNA COVID-19 vaccination made to the United States' Vaccine Adverse Event Reporting System (US-VAERS) and Canadian Ontario's vaccine safety surveillance system; or categorized as self-reported by vaccine recipients, mostly by healthcare workers, via an online questionnaire conducted to compare the safety of available mRNA COVID-19 vaccines.
The following data were extracted from selected studies: authors; publication year; study location; study design and setting; age; proportion of male patients; past history of allergies or allergic reactions and/or anaphylaxis; vaccine brand and dose [if 1st dose, 2nd dose or both]; time from vaccination to anaphylactic or nonanaphylactic reaction onset; signs and symptoms; anaphylaxis level of severity if occurred; treatment setting and therapy received for reaction; if epinephrine was received; treatment outcome; if skin test was performed and test result; assessment of included study risk of bias (tool used; finding) and; remarks on notable findings.

Quality assessment
The quality assessment of the studies was undertaken mainly based on the Newcastle-Ottawa Scale [NOS] to assess the quality of the selected studies [12]. This assessment scale has two different tools for evaluating case-control and cohort studies. Each tool measures quality in the three parameters of selection, comparability, and exposure/outcome, and allocates a maximum of 4, 2, and 3 points, respectively. Highquality studies are scored greater than 7 on this scale, and moderate-quality studies, between 5 and 7. Revised Cochrane risk of bias tool (RoB 2.0) was used to assess the risk of bias in randomized controlled studies [13]. Quality assessment was performed by five authors (H.M.A., A.S.A., T.T.A., G.A., and A.M.A.) independently, with any disagreement to be resolved by consensus.

Data analysis
The primary outcome of interest was to determine the incidence rates of anaphylactic and nonanaphylactic reactions induced after the administration of mRNA COVID-19 vaccines [namely Pfizer-BioNTech and Moderna]. Secondary outcome was to analyze identified evidence describing the demographic and clinical characteristics, triggers, presenting signs and symptoms, treatment and clinical course of confirmed cases with anaphylactic and nonanaphylactic reactions induced by those two mRNA vaccines. Taking a conservative approach, a random effects meta-analysis with the Hunter-Schmidt model was used [14], which produces wider confidence intervals [CIs] than a fixed effect model. Results were illustrated using forest plots. Statistical heterogeneity was evaluated using the Cochran's chisquare (χ 2 ) and the I 2 statistic [15]. An I 2 value of > 50% is suggestive of significant heterogeneity [16]. To detect the source of heterogeneity, subgroup analysis was performed based on vaccine brand administered [Pfizer-BioNTech or Moderna]. Publication bias was evaluated using funnel plots and the Egger's correlation test, with p < 0.05 indicating statistical significance [17]. All analyses were done using R version 4.1.0 with the packages metafor and meta.

Study characteristics and quality
A total of 1,734 publications were identified (Fig. 1). After scanning titles and abstracts, we discarded 609 duplicate articles. Another 264 irrelevant articles were excluded based on the titles and abstracts. The full texts of the 416 remaining articles were reviewed, and 390 irrelevant articles were excluded. As a result, we identified 26 studies that met our inclusion criteria . The The Brighton Collaboration case definition uses combinations of symptoms to define levels of diagnostic certainty. Brighton level 1 represents the highest level of diagnostic certainty that a reported case is indeed a case of anaphylaxis; levels 2 and 3 are successively lower levels of diagnostic certainty. Level 4 is a case reported as anaphylaxis but that does not meet the Brighton Collaboration case definition. Level 5 is a case that was neither reported as anaphylaxis nor meets the case definition [42] b Time to symptom onset missing in 2 Pfizer-BioNTech reports c NIAID/FAAN clinical criteria for the diagnosis of anaphylaxis must meet 1 of the following criteria: (1) acute onset with involvement of skin and/or mucosal tissue and either (a) respiratory compromise or (b) reduced blood pressure or associated symptoms of end organ dysfunction; (2) 2 or more of the following occur after exposure to a likely allergen for that patient: (a) involvement of skin or mucosal tissue, (b) respiratory compromise, (c) reduced blood pressure or associated symptoms, or (d) persistent gastrointestinal symptoms; and (3) reduced blood pressure after exposure to a known allergen for that patient [43] § An elevated tryptase level was defined as either above the upper limit of normal or > (2        Tables 1 and 2. Among the included studies, 12 reported anaphylaxis reactions [22,25,27,29,[33][34][35][36][38][39][40], 17 reported nonanaphylactic reactions [18-24, 26-28, 30-32, 35, 37, 39, 41], and 3 studies reported both anaphylactic and nonanaphylactic reactions (27,35,40). The majority of studies reporting incidence rates of anaphylaxis and nonanaphylactic reactions were conducted based on registry databases [22,24,30,33,39,40], and most were from hospital and/or emergency department admission databases [18, 19,   systematic review and meta-analysis, 89.1% [23] of whom were part of the US-VAERS [24,30,39,40] and Ontario's vaccine safety surveillance system to report anaphylactic and nonanaphylactic events [33]. Anaphylactic and nonanaphylactic reactions were reported by vaccine recipients via an online self-administered survey [20,22,27,29,37]; and only few studies in which anaphylactic events received through electronic health records were reviewed by allergists and immunologists [22] or physicians with unspecified medical specialty [39,40]. Only few nonanaphylactic reactions were reported by dermatologists, other physicians, and other HCWs [30]. There were 8 case report [18,25,28,31,[34][35][36]38], 5 cohort [20,22,29,37,41], 4 case series [21][22][23]26], 2 randomized controlled trial [19,32] and 1 randomized cross-sectional [27] studies. These studies were conducted in United States (n = 15), Italy (n = 3), United Kingdom (n = 2), Canada (n = 1), France (n = 1), South Korea (n = 1), Spain (n = 1), Portugal (n = 1), and Czech Republic (n = 1). Only 3 studies were performed with a multi-centre design [19,22,35]. The median NOS score for these studies was 6 (range, 5-7). Among the 26 included studies, 22 studies were moderate-quality studies (i.e., NOS scores were between 5 and 7) [18-25, 27-39, 41] and 4 studies demonstrated a relatively high quality (i.e., NOS scores > 7) [22,26,39,40]; Tables 1 and  2.
Essential to exclude the presence of mastocytosis and to characterize the anaphylactic events that might occurred during vaccination, only four studies reported measurement of tryptase levels in patients who suffered anaphylactic reactions (22,34,36,38); Tables 1 and 2.
Two patients who had anaphylaxis following the first dose of COVID-19 vaccine successfully received the second dose without premedication or split-dosing in consultation with an allergist. These two only exhibited mild nausea and minor lip or tongue tingling on second dose administration [34,35] (Table 1). A total of 39 study participants with reported history of nonanaphylactic reactions to COVID-19 mRNA vaccines were challenged with a second dose [21-23, 28, 35], and none were given the vaccine by split-dosing; and only 12 participants received premedication [22] and most patients then received the second dose of vaccine without a relapse of large local reactions (21-23, 28, 35). However, 3 patients had recurrent reactions similar to those after the initial dose and 3 patients had recurrent reactions that were of a lower grade than after the initial dose [22] (Table 2). reactions in COVID-19 mRNA vaccine recipients were limited to case reports, case series and cohort studies; and majority of studies reporting incidence rates of anaphylactic and nonanaphylactic reactions were conducted based on registry databases, and may not reflect the overall incidence rates of anaphylactic and nonanaphylactic reactions as the vast majority were from hospital and/or emergency department admission databases. We found a wide variation in the prevalences of reported anaphylactic and nonanaphylactic reactions across different studies, because they assessed different populations and settings, and used different methods. Thus, it is important that any studies in the future use a consistent and accurate definitions of anaphylactic and nonanaphylactic reactions. Use of the correct epidemiological method to define prevalence would help to identify any true difference between countries, and help to provide an overall estimate of prevalence. Across the included studies, the most commonly identified risk factors for anaphylactic and nonanaphylactic reactions to SARS-CoV-2 mRNA vaccines were being female and previous history of atopy. Based on the published studies included in our review, evidence suggests females are much more susceptible to anaphylactic and nonanaphylactic reactions. Formation of a polyethylene glycol (PEG)-conjugated lipid derivative is hypothesized to cause COVID-19 mRNA vaccineassociated anaphylactic and nonanaphylactic reactions [44]. Sensitization to PEG is more common in women due to the relatively frequent exposure to PEG-containing products, such as cutaneous exposure to cosmetics or the use of medications such as contraceptive injections and explains female predominance in the reported cases of vaccine-associated anaphylactic and nonanaphylactic reactions [45]. Another possible explanation includes hormonal differences such as the role of estrogen which may be an important factor in allergic immunological responses [45].

Discussion
Consistent with previous studies, our review found that foods, drugs or therapeutic agents, contrast media, stinging insects or jellyfish, cosmetics, detergents, household products, and latex are the most common triggers of anaphylactic and nonanaphylactic reactions induced by mRNA COVID-19 vaccines [21-23, 38, 46]. Previous history of anaphylaxis; and comorbidities such as asthma, allergic rhinitis, atopic and contact eczema/ dermatitis and psoriasis and cholinergic urticaria were also described by other investigators as elicitors for anaphylactic and nonanaphylactic reactions to mRNA COVID-19 vaccines [47,48]. Nevertheless, assessment by an allergist is NOT required for people with a history of unrelated allergies, including to allergies to foods, drugs, insect venom or environmental allergens and COVID-19 vaccines can be administered in these individuals without any special precautions [49].
The occurrence of anaphylaxis is of a particular concern in the use of newly authorized COVID-19 vaccines. The diagnosis of anaphylaxis requires professional judgment, such as case-by-case interpretation using validated diagnostic criteria for anaphylaxis [42,43] as anaphylaxis occurrence may have a profound effect on the quality of life of the sufferer and their family [50]. It is important to identify those who might be at an increased risk of anaphylaxis in order to reduce morbidity, and provide successful management plans. Given the importance of anxiety as a contributor to the quality of life impact of COVID-19 vaccines allergy, our finding that fatal vaccine anaphylaxis incidence is relatively nil may be important information for vaccine-allergic people and their carers. Although anaphylaxis after mRNA SARS-CoV-2 vaccination is very rare, its immediate onset [usually within minutes] and life-threatening nature require that all HCWs and facilities providing vaccinations have procedures in place for anaphylaxis management. Patients who have experienced anaphylaxis should be referred to an allergy/immunology specialist to confirm the diagnosis, confirm or determine the causes(s), and determine relevant patient risk factors for severe or fatal anaphylaxis (e.g., comorbidities or concurrent medications that could be modified to reduce the patient's risk of recurrences in the future). In someone with a suspected or confirmed allergy to a SARS-CoV-2 vaccine or one of its components for whom an additional dose is required, choices include deferral of the second vaccine dose, selection of an alternative vaccine with a different platform and excipients, and the administration of the same vaccine using a graded vaccine administration protocol [51].
Healthcare workers need to differentiate the occurrence of anaphylaxis from the development of vasovagal reactions and anxiety-related symptoms [52]. While it is important to recognize and treat anaphylaxis, it is equally important not to label these other conditions as anaphylaxis, particularly when there are no objective findings. Localized cutaneous reactions were common following the mRNA vaccines and include urticarial and morbilliform eruptions which may reflect immediate hypersensitivity but have rarely been associated with anaphylaxis. There are infrequent reports of erythema multiforme, pernio/chilblains, cosmetic filler reactions, zoster, herpes simplex flares, and pityriasis rosea-like reactions, mainly occurring in high-risk patient groups [27,30]. Ultimately, the identified cutaneous reactions are largely self-limited and should not discourage vaccination. Existing reports should reassure patients of the overall compelling safety profiles and benignity of skin reactions following the mRNA SARS-CoV-2 vaccines. Mimics of anaphylaxis should not discourage vaccination against the COVID-19 pandemic and a high rate of SARS-CoV-2 vaccines uptake across all sectors of worldwide societies is a priority public health goal. These findings should provide reassurance to HCWs and to vaccine recipients and promote confidence in the safety of COVID-19 vaccines.

Limitations
There are several limitations to our findings. First, the quality of data submitted by reporters on anaphylactic and nonanaphylactic reactions mostly through registry databases and online questionnaires varies widely and never used standardized data collection methods. Second, most of the included studies relied on the clinical history for the diagnosis of anaphylactic and nonanaphylactic reactions. Third, some studies do not provide denominator data to calculate the incidence rates of anaphylactic and nonanaphylactic reactions; therefore, estimating accurate incidence rates by interpreting these studies databases was not possible. Fourth, the quality and completeness of the reports included in few studies might not be optimal, thus making the assessment of causality challenging. The fifth limitation is the exclusion of non-English language studies.

Conclusion
The prevalence of COVID-19 mRNA vaccineassociated anaphylaxis is very low; and nonanaphylactic reactions occur at higher rate, however, cutaneous reactions are largely self-limited. Both anaphylactic and nonanaphylactic reactions should not discourage vaccination.