Incidence of serum sickness following Indian polyvalent antivenom therapy in a cohort of snake-envenomed patients in rural Sri Lanka

Abstract Introduction Serum sickness is a poorly reported delayed adverse reaction following snake antivenom therapy. We aimed to assess the frequency of serum sickness associated with administering Indian polyvalent antivenom in Sri Lanka. Methods We recruited patients from the Anuradhapura snakebite cohort who were admitted to a rural tertiary care hospital in Sri Lanka over one year period. Patients were interviewed over the phone 21 to 28 days post-envenoming to collect data on clinical effects: fever/chills, arthralgia/myalgia, rash, malaise, headache, abdominal pain, and nausea/vomiting. The presence of three or more symptoms between the 5th to 20th days after snake envenoming was defined as serum sickness. Results We were able to contact 98/122 (80%) patients who received antivenom and 423/588 (72%) who did not receive antivenom during the study period. The treated patients received a median dose of 20 vials (interquartile range: 20–30) of Indian polyvalent antivenom and of them, 92 (92%) received premedication. However, 67/98 (68%) developed acute adverse reactions to antivenom, including 19/98 (19%) developing anaphylaxis. Only 4/98 (4%) who received antivenom met the criteria for serum sickness, compared to none who did not receive antivenom therapy. All patients who developed serum sickness were envenomed by Russell’s vipers, were premedicated, and received VINS Bioproducts antivenom. Three of them were treated with hydrocortisone in the acute stage, as premedication or as a treatment for acute adverse reactions of antivenom. Although all four patients sought medical advice for their symptoms, only one was clinically suspected to be serum sickness and treated, while the others were treated for infections. Conclusions We confirmed that Indian polyvalent antivenom use in Sri Lanka is associated with high rates of acute adverse reactions. In contrast to studies of other antivenoms only a small proportion of patients developed serum sickness.


Introduction
Snakebite is a neglected tropical disease affecting mainly rural farming communities in South Asia, South East Asia, Sub-Saharan Africa, and Latin America [1][2][3]. Snake envenoming results in a range of local and systemic effects, including venom-induced consumption coagulopathy, neuromuscular paralysis, acute kidney injury, and myotoxicity [4]. The survivors of snakebites may be left with debilitating long-term, physical, and psychological effects [5,6].
Antivenom is the mainstay of treatment for snakebites [7]. A major concern with antivenom therapy is that it is frequently associated with acute and delayed adverse reactions [8,9]. Reactions are either immediate such as pyrogenic or anaphylactic reactions, or delayed reactions such as serum sickness. In Sri Lanka, 40% to 75% of patients develop acute adverse reactions following Indian polyvalent antivenom therapy, an equine F(Ab') 2 antivenom developed against the four major snakes in India (Naja spp., Daboia russelii, Echis carinatus, and Bungarus caeruleus), and anaphylaxis has been reported in as high as 43% of patients [10][11][12][13].
Reports of serum sickness following antivenom therapy vary across different settings globally, ranging from 5 to 56% [8,[14][15][16]. Serum sickness is an adverse drug reaction that occurs 5 to 20 days after the administration of heterologous foreign proteins or serum, such as antivenom, vaccines (e.g., anti-rabies, tetanus toxoid, and diphtheria), and immunomodulators (e.g., infliximab and rituximab) [17][18][19]. With serum sickness, the host immune system produces antibodies against these heterologous proteins, which results in the formation of circulating antigen-antibody complexes and results in a Type III hypersensitivity reaction. The mononuclear phagocytic cells are involved in clearing these immune complexes from circulation. However, in certain instances, mononuclear phagocytic cells fail to remove immune complexes from the circulation, saturating the circulation with the immune complexes, and resulting in their deposition in various tissues such as vascular endothelium, synovial tissues, and renal glomeruli. Complement activation results in the stimulation of histamine secretion which leads to increased vascular permeability. The result is the development of signs and symptoms of serum sickness such as fever, chills, urticaria, rash, arthralgia, myalgia, abdominal pain, nausea, vomiting, headache, and lymphadenopathy [8,14,[20][21][22].
There has not been a uniform definition of serum sickness across studies, with a range of different signs and symptoms being included, and varying times of onset of these clinical effects [8]. In settings in which serum sickness is uncommon, this may also be due to poor reporting because serum sickness is delayed, often when the patient has been discharged from the hospital and lost to follow-up. In addition, serum sickness can mimic other common conditions, such as influenza-like or febrile illnesses, which are usually self-limiting [23].
There are few data on the prevalence of serum sickness, even with antivenoms that have high acute reaction rate, such as in Sri Lanka. There is little information on the impact of serum sickness on the patient's life as an intermediate effect associated with snakebite and its treatment. We aimed to investigate the incidence of serum sickness amongst patients in a snakebite cohort from Sri Lanka, who received Indian polyvalent antivenom therapy.

Anuradhapura snakebite cohort
This study included patients recruited to the Anuradhapura snakebite cohort over one year. The Anuradhapura snakebite cohort prospectively enrols snakebite patients admitted to the Teaching Hospital Anuradhapura, the third largest hospital in Sri Lanka and located in an area of the country with the highest incidence of snakebite and envenoming [24,25]. This cohort recruits all confirmed snakebite victims greater than 16 years of age. Medically qualified research assistants assess patients on admission and at 1, 4, 8, 12, 24 h after, and then daily until discharge. They prospectively record all clinical manifestations and complications, laboratory investigations, and medical interventions. Antivenom administration, including the time, dose, batch number, pre-antivenom medication, development of acute adverse reactions according to Brown's grading, and the medications used to manage adverse reactions are documented [26]. A pre-formatted clinical instrument is used to collect the data, which are then prospectively entered into a relational database. In recruited patients, snakes are authenticated by either (1) expert identification of the specimen by a herpetologist (AS), (2) a species-specific sandwich enzyme-linked immunosorbent assay (ELISA) of blood samples, or (3) a previously validated syndromic approach based on the Sri Lankan venomous snake identification algorithm [27,28].

Patients
For this analysis, we recruited patients from the Anuradhapura snakebite cohort greater than 18 years of age who had a snakebite from 1 July 2021 to 30 June 2022 and received antivenom during their hospital stay. The patients who were admitted for snakebite during the same period, but did not receive antivenom during the hospital stay, were used as the comparison group. The study and comparison group patients were contacted over the telephone three weeks after the snake bite. If the patients were not contactable on the 21st day post-snakebite, we tried to contact them daily until the 28th day post-bite, before considering them as failures to follow up.

Data collection
The patients were interviewed over the phone by a medical doctor (SWa) guided by a standardised questionnaire in the local language, and the data were recorded on clinical research forms. This was then entered into a relational database by the same investigator. Patients were asked about the development of an erythematous rash/urticaria, fever/chills, arthralgia or myalgia, malaise, headache, and gastrointestinal symptoms, including nausea/vomiting and abdominal pain, that began 5 to 20 days following the snakebite and treatment with antivenom. The presence of three or more symptoms was defined as serum sickness based on a previous study [14]. Medical treatments received during the postdischarge period were also recorded.
Demographic data, data on the patient's medical history and regular medications, acute admission following the snakebite, including features of envenomation, complications, investigations, acute management, antivenom therapy, pre-antivenom medication, acute adverse reactions following antivenom therapy, and their treatment, were extracted from the Anuradhapura snakebite cohort database.

Ethics approval
The Anuradhapura snakebite cohort has human ethics approval from the Ethics Review Committee of the Faculty of Medicine and Allied Sciences of the Rajarata University of Sri Lanka, for prospective patient enrolment (ERC/2012/036, ERC/2013/019) and to review patients physically and over the telephone (ERC/2017/047, ERC/2017/48). Written informed consent is obtained from all the patients on admission and proxy consent (verbally taken from the guardian) when the patients are severely ill or unconscious. Written informed consent is taken before discharge when the patient clinically improved.

Results
During the study period, 775 patients were recruited to the Anuradhapura snakebite cohort. Seven (1%) died following the snakebite and 58 (7%) survivors were younger than 18 years of age. Of the 710 remaining patients, 122 (17%) received Indian polyvalent antivenom. Of these 710 patients, 521 (73%) were successfully contacted over the phone 21 to 28 days after the snakebite and are included in this study ( Table 1). The median patient age was 44 years (interquartile range: 32-55 years) and 314 (60%) were males. Of the contacted patients, 98 (19%) received Indian polyvalent antivenom during their hospital stay.
Of the 98 patients who received polyvalent antivenom (median, 20 vials; interquartile range, 20-30), 92 patients (94%) were given premedications to prevent acute adverse reactions to antivenom. These include subcutaneous epinephrine, intravenous hydrocortisone, and any of the following medications alone or in combination via intravenous or oral route; chlorphenamine (chlorpheniramine) maleate, metoclopramide, promethazine, ondansetron, ( Table 2). Sixtyseven of the 98 (68%) patients developed acute adverse reactions to antivenom, including 19 (19%) with Brown grade III reactions (severe anaphylaxis). For the treatment of acute adverse reactions, 54/67 received intravenous or intramuscular epinephrine, 39/67 received intravenous hydrocortisone, and 38/69 received intravenous or oral chlorphenamine maleate. Table 3 provides the frequency of the different clinical features of serum sickness that were reported from the 5th to the 20th day following the snakebite in patients who received and did not receive antivenom. Only 4/98 (4%) patients who received antivenom reported three or more symptoms of serum sickness and hence met the criteria for serum sickness. None of the 423 patients who did not receive antivenom met the criteria. All four patients were bitten by Russell's vipers, were given Indian polyvalent antivenom (VINS Bioproducts), and three of them had acute adverse reactions. Three were also administered either corticosteroids or antihistamines to prevent or manage acute adverse reactions (Table 4). Flucloxacillin was administered in 67/98 (68%) of the antivenom patients, and 214/423 (51%) of non-antivenom patients. Although all four patients with serum sickness received flucloxacillin, none of 214 patients receiving flucloxacillin without antivenom developed a reaction.
All four patients who developed serum sickness sought treatment for their symptoms, three in government hospitals and one from a general practitioner. Two were treated for viral fever as an outpatient; one was treated for cellulitis at the bite site as an inpatient, while one patient was treated as an inpatient for suspected serum sickness due to antivenom.

Discussion
We found that only 4% of patients reported clinical features consistent with serum sickness following Indian polyvalent antivenom treatment in rural Sri Lanka, which is unlikely to have a significant impact on snakebite survivors. No patient who did not receive antivenom met the criteria for serum sickness. All patients who developed features suggestive of serum sickness had Russell's viper bites and received VINS Bioproducts antivenom. Three of those who met the criteria for serum sickness received 20 vials of antivenom, and one had 30 vials. Three of them initially developed acute adverse reactions to antivenom, despite all receiving premedications, including epinephrine.
The definition and diagnostic criteria for serum sickness have not been validated to date, which is the most likely reason for the vast difference in the rates of serum sickness   reported between studies À 4% to 56% [8,14]. Serum sickness is defined as an immune-complex-mediated hypersensitivity reaction that classically presents with fever, rash, polyarthritis, or polyarthralgia [22]. Most studies have used a combination of these clinical features as part of their definition, including the present study here and an Australian study [14]. Ruha et al. [29] reported 7% of patients having serum sickness, based on rash, pruritis, fever, myalgia, and arthralgia for the diagnosis. Lavonas et al. [30] used the same criteria, but included the presence of vomiting, and reported 4.9% having serum sickness. Bush et al. [31] used fever, arthralgia, and pruritus as diagnostic criteria for serum sickness, and reported 6.7% of patients developing serum sickness, after Crotalidae polyvalent immune Fab (ovine) antivenom for Southern Pacific rattlesnake (Crotalus helleri) envenomation. LoVecchio et al. [15] reported the highest rate of serum sickness in 56% of cases following the administration of crotalid Wyeth polyvalent antivenom (equine, whole IgG), for rattlesnake envenomation. However, LoVecchio et al. defined serum sickness as the development of an unexplained rash within 3 to 21 days of antivenom administration, likely increasing the number of patients meeting the definition of serum sickness [15].
The proportion of patients with serum sickness in our study was low compared to other studies [15,[29][30][31], including the Australian study, which used the same definition for serum sickness [14]. This was despite over two-thirds of our patients developing acute adverse reactions and none of the four patients with serum sickness receiving more than 30 vials of antivenom. In the Australian study, only 18% of the patients had acute adverse reactions and there was no association between the development of serum sickness and the volume of the antivenom administrated [14]. The differences in reported rates of serum sickness are likely to be due to its ambiguous definition, but other factors are also likely to be involved.
One-third of the patients who received antivenom in our study received intravenous hydrocortisone as premedication to prevent acute adverse reactions. Half of those who developed acute adverse reactions received intravenous hydrocortisone in managing the acute adverse reactions. Nonetheless, three of the four patients with serum sickness in our study had received intravenous hydrocortisone as a premedication or as a treatment for acute adverse reactions. Whether the administration of intravenous hydrocortisone to patients was associated with a lower incidence of serum sickness, requires further exploration. Serum sickness-like reactions are a late immune response that appears 5-10 days after exposure. The common triggers for serum sickness-like reactions include antibiotics and vaccines [32][33][34][35]. More than half of all the patients with snakebite, including those not receiving antivenom, received flucloxacillin to prevent or treat secondary bacterial infections at the bite site. Although the four patients with serum sickness were treated with flucloxacillin, 214 patients not receiving antivenom were treated with flucloxacillin, and none developed serum sickness. Nonetheless, to differentiate serum sickness from serum sickness-like reaction we would have needed to measure serum compliment C3 and C4 concentrations. In serum sickness, serum C3 and C4 concentrations will be low, whereas in serum sickness-like reactions C3 and C4 concentrations will be normal [36]. This study was conducted over the phone and there were no facilities to measure complement concentrations and therefore no way of excluding possible serum sickness-like reactions in these four patients.

Strengths and limitations
This study used data from a prospective snakebite cohort, so data on the acute presentation, complications, and treatment were available for analysis. The study was telephone interview-based and relied on the patients' descriptions of illness, hence information bias associated with this approach cannot be excluded. In particular, we could not assess the serum concentrations of C3 and C4 in the patients with features of serum sickness to confirm the diagnosis.

Conclusion
Despite high rates of acute adverse reactions, the occurrence of the clinical features of serum sickness was uncommon among the snakebite patients who received Indian polyvalent antivenom in this study. Therefore, serum sickness is not a major adverse reaction of Indian polyvalent antivenom in rural Sri Lanka.