Adverse Events Associated with the Clinical Use of Bee Venom: A Review

Bee venom is used to treat various diseases but can cause a tickling sensation and anaphylaxis during clinical treatment. Adverse events (AEs) associated with bee venom may vary depending on the dosage, method, route of administration, and the country, region, and user. We summarized the AEs of bee venom used in various ways, such as by the injection of extracts, venom immunotherapy (VIT), live bee stings, or external preparations. We conducted a search in eight databases up to 28 February 2022. It took one month to set the topic and about 2 weeks to set the search terms and the search formula. We conducted a search in advance on 21 February to see if there were omissions in the search terms and whether the search formula was correct. There were no restrictions on the language or bee venom method used and diseases treated. However, natural stings that were not used for treatment were excluded. A total of 105 studies were selected, of which 67, 26, 8, and 4 were on the injection of extracts, VIT, live bee stings, and external preparation, respectively. Sixty-three studies accurately described AEs, while 42 did not report AEs. Thirty-five randomized controlled trials (RCTs) were evaluated for the risk of bias, and most of the studies had low significance. A large-scale clinical RCT that evaluates results based on objective criteria is needed. Strict criteria are needed for the reporting of AEs associated with bee venom


Introduction
Bee venom treatment uses the pharmacological effect of bee sting toxins and is widely used worldwide [1]. In addition to musculoskeletal diseases, bee venom is used for therapeutic purposes such as for uterine ovarian disease [2], cancer [3], and atopic dermatitis [4].
Bee venom treatment is performed in various ways, such as through apitoxin, bee venom acupuncture, venom immunotherapy (VIT), and live bee stings [5]. Among studies on the adverse events of bee venom, studies summarizing adverse events according to the type of paper have been conducted along with randomized controlled trials (RCTs) [6]. However, no studies have reported the side effects of bee venom treatment.
The toxin component of bee venom is presented to T cells by antigen-presenting cells in the skin and eventually causes an allergic reaction by producing IgE [7]. The most serious adverse event of bee venom treatment is anaphylaxis; however, the incidence is not high [8]. If anaphylaxis occurs, epinephrine may be treated preferentially [9]. However, owing to practical and ethical issues, strong evidence on the diagnosis and management of anaphylaxis is lacking [10]. Anaphylaxis can present similarly to acute asthma, local angioedema, fainting, and anxiety/panic seizures [11].

Risk of Bias in Included Studies
Among the 37 RCTs, 1 study that used bee venom for the intervention group and wasp venom for the control group and 1 study that used different doses of bee venom for the intervention and control groups were excluded. For the remaining 35 RCTs, the interventions, control group treatment contents, evaluation index, results, and effective values were summarized. Subsequently, the risk of bias (RoB) was evaluated based on the content of the included studies.
All 35 studies in the first domain of random allocation and double blindness were evaluated as "some concerns." In all studies, participants were randomly assigned. However, there was no information on blinding after the random assignment. All 35 studies were evaluated as "some concerns" in the second domain because there were dropouts, the sample size was not sufficient, or the caregiver was not blinded to the group assignment of the participants. In all 35 studies, the results of the study participants were evaluated as "low risk" because they appeared to be universally available to all participants. In the fourth domain, 8 studies were "low risk" because there was an objective outcome measurement method, but 27 studies were "high risk" because only scales based on the subjective symptoms of participants were used. All 35 studies were evaluated as having "some concerns" because no implementation plan or protocol was mentioned. The details are presented in Table 7 and Figures 2 and 3.  Toxins 2022, 14, x FOR PEER REVIEW 15 of 32       An [31] cancerous pain from lung cancer Wei [52] rheumatoid arthritis (1) p < 0.01 (2) p < 0.01 (3) p < 0.01 (4) p < 0.05 Chen [56] Rheumatoid arthritis   Qin [110] shoulder-hand syndrome after CVA (2) (A) 100% (B) 0% (1) p < 0.05 (2) p < 0.05 Wu [112] lumbar disc herniation

Discussion
We conducted a literature search using eight databases: PubMed, Cochrane, EMBASE, CINAHL, CNKI, NDSL, OASIS, KISS, KoreaMED, and KMBASE. However, there were many cases in which access to Chinese-based databases was not possible, so an additional literature search could not be performed. Ultimately, 105 studies were included. There were forty-nine, twenty-eight, six, five, four, three, two, two, one, one, one, and one studies from China, Korea, Germany, Australia, Poland, Turkey, Spain, Czech Republic, Greece, Belgium, France, and Japan, respectively. As for the paper type, there were 37 RCTs, 33 CRs, 15 CSs, 14 cohort studies, and 6 nRCTs.
When classified according to the stimulation method of bee venom, there were 67, 26, 8, and 4 studies on extract injections, VIT, live bee stings, and external preparations, respectively. Twenty-seven studies described the injection capacity of bee venom, but few studies specifically described the dose that was injected into how many acupoints.
Twenty-eight studies reported no adverse events, thirty-four specifically reported adverse events, and the remaining forty-three studies partially or failed to describe adverse events. Seven of the forty-six studies did not describe specific symptoms of adverse events but described adverse events such as "skin problem" and "systemic reaction". Based on Mueller's classification, twenty-nine cases were grade I and two cases were grade II with the patients complaining of abdominal pain, chest pain, and vomiting. There was also one case of grade III, with the patient presenting with weakness and dyspnea, and eleven cases of grade IV, with patients suffering from hypotension and cyanosis. According to Spilker's classification, 26 cases were "mild" with no functional disruption to daily activities, 4 cases were "moderate" with symptoms disappearing over time when additional treatment was applied, and 11 cases were "severe" with immediate treatment required or after-effects. Regarding "mild" symptoms, there were cases where it was accompanied by "moderate" to "severe" symptoms.
Bee venom injections are performed using refined bee venom. In this process, active ingredients can be extracted separately and allergens can be removed. Moreover, the capacity and concentration of the bee venom injections can be easily controlled [117]. Depending on the venom to be purified, snakes [118] and jellyfish [119] can be used instead of bees. However, as an invasive treatment, there may be a risk of infection, depending on the injection site. In addition, since the unification of terms, such as bee venom acupuncture and bee venom pharmacopuncture, has not been achieved, it is necessary to establish appropriate terminology.
VIT is a prophylactic method that aims to reduce hypersensitivity in individuals with hypersensitivity to venom [120]. If adverse events occur during follow-up, additional treatment such as the oral administration of omalizumab, an anti-IgE, may be introduced [121]. However, in the case of VIT, since it is targeted at people who have already experienced adverse events or hypersensitivity, it seems that the definition of an adverse event should be different.
Live bee stings may have similar effects to bee venom injections but are clinically impractical because they require live bees [122]. Bees vary slightly in composition and concentration, depending on the type and growth area [123]. In addition, live-bee dermatitis may occur if infected [124]. Since this method directly uses bees to sting, criteria that detail the infection control process, effective bee type, recommended time, and/or the number of stings are needed.
Bee venom is sometimes used as an external preparation, and honey, royal jelly, and bee venom are used to treat and prevent oral diseases [125], while cream containing bee venom is used to improve wrinkles [126]. A direct correlation between bee venom allergy and bee products has not been revealed, but some people are allergic to honey or propolis [127]. In the case of external preparations, more clinical studies are needed to determine the correlation between the concentration of ingredients, the amount of application, and allergies.
Of the 105 studies included in this review, only 63 reported specifically on the adverse events that occurred. VIT seems to be used in many Western countries, whereas bee venom injections and live bee stings seem to be used in many Eastern countries. Since the method of bee venom stimulation differs by country and culture, it is thought that the reporting method for the adverse events that occur may be different. There are criteria such as Mueller's and Spilker's classification, but these criteria do not appear to be essential in the reporting of venom treatment. Since bee venom has the potential to cause anaphylaxis, reports of side effects must be included in venom clinical trials.
RoB evaluation was conducted on 35 RCT studies. Each domain explains a randomization process, deviations from intended interventions, missing outcome data, a measurement of the outcome, selection of the reported result and overall bias. As shown in the RoB results, there were no studies with a low RoB. To supplement this study, objective and diverse scales are required in large RCTs to evaluate the effectiveness of bee venom, and a rigorous reporting framework for adverse events should be presented.
From the 105 studies reviewed, there were 10 studies in which Mueller's grade IV adverse events occurred (3 extract injection studies, 6 VIT studies, and 1 live bee sting study). Only 2 studies were conducted in advance. The most serious adverse events that can occur with bee venom treatment were anaphylaxis and unrecoverable sequelae. Since there is a possibility of anaphylaxis, it is recommended that a person with medical knowledge manages patients undergoing a bee venom procedure. Further research is needed on the relationship between skin test results and serious adverse events. However, to reduce the occurrence of serious adverse events in clinical practice, skin tests should be conducted prior to treatment. In addition, since skin tests are used to adjust the concentration and capacity of the active ingredient, the live bee sting type is not recommended. In the selected papers, the capacity of bee venom was expressed in various ways, such as mL and cc. When researchers write papers or conduct experiments, it is necessary to use general units such as mg/kg, or to specify capacity units and concentrations of effective ingredients according to the purpose of the study. This study focuses on the adverse events of bee venom. If an additional comparative study on the effect, adverse events rate, and fatality rate according to the stimulation type is conducted, the clinician may use bee venom in consideration of the effect and adverse events.

Conclusions
This study reviewed the adverse effects of bee venom stimulation. Most of the RoB evaluations of RCT studies were not significant, and large-scale RCT studies with a system for reporting adverse events of bee venom are required. A skin test is needed to reduce the occurrence of adverse events, and a person who can cope with anaphylaxis should perform a bee venom procedure. It was confirmed that many studies omitted reports of adverse events. In order to analyze the occurrence and fatality rate of adverse events according to the stimulation type, it is essential to include a report of adverse events when using bee venom. . The search was conducted using "bee venom acupuncture" and "adverse events" as keywords. There were no restrictions on the country or the language of the issue. The search was conducted up to 28 February 2022.

Types of Studies
CRs, CSs, and nRCTs were included. Experimental, animal, and protocol studies were excluded.

Types of Participants
There were no special restrictions on the diseases treated and patient characteristics.

Types of Interventions
All treatments using bee venom were included in the intervention group. Nonintervention cases were excluded even if bee venom was used. In the case of RCTs, group classification according to the capacity of bee venom was included. Studies that included individuals who were accidentally stung by a bee (i.e., the sting was not part of their treatment) were excluded. There were no restrictions on the comparison group.

Types of Outcome Measures
Contents related to adverse events were also extracted. The symptoms were classified into skin problems, systemic reactions, and others. The severity of the symptoms was classified as mild, moderate, and severe according to Spilker's classification (Table 5) [128] and grades I to IV according to Mueller's classification (Table 6) [129]. Causality was classified as certain, probable, possible, unlikely, unclassified, and unclassifiable according to the WHO-UMC causality scale (Table 8) [130]. No adverse events were described as "non-reported," and no adverse events were "none".

Selection of Studies
Two authors (JY and GL) independently searched each of the eight databases based on the abstracts. The full text was checked for papers for which the abstract was insufficient. The entire process was summarized according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines ( Figure 4) [131].  The study selection process is summarized in Figure 4. Duplicate studies and those that did not meet the selection criteria were excluded.

Data Extraction
One author (JY) extracted the data, and the other (GL) inspected the extracted data. The number of participants, type of bee venom treatment method, outcomes, and the information related to adverse events were recorded.

Assessment RCTs
Two reviewers evaluated the bias of RCT studies using the RoB evaluation [132]. The bias evaluation item consisted of five categories: (1) randomization process, (2) deviations from intended interventions, (3) missing outcome data, (4) measurement of the outcome, and (5) selection of the reported result. The first domain is whether random assignments and double blindness are properly performed, and the second domain is whether the dropout rate is high, the sample size is sufficient, or the caregiver is aware of the group assignment of the participants. The third domain concerned whether the results of the study were all available to the study participants. The fourth domain relates to whether the method of measuring results is the same and appropriate between groups, and the fifth domain relates to whether the research results were conducted using a pre-protocol. In addition, overall bias was evaluated by synthesizing five evaluation items. In each item, if there is no RoB, it is marked as "low risk", if the RoB was high as "high risk", and if there is no information on the item, it was marked as "some concerns".

Data Availability Statement:
The data used in this study are available from the corresponding author upon request.

Conflicts of Interest:
The authors declare no conflict of interest regarding the publication of this paper.