Incidence of Immediate Hypersensitivity Reaction and Serum Sickness Following Administration of Crotalidae Polyvalent Immune Fab Antivenom: A Meta-analysis

Authors

  • Tammi H. Schaeffer DO,

    1. From the Rocky Mountain Poison & Drug Center, Denver Health and Hospital Authority (THS, VK, LMR, EJL), Denver, CO; and the Department of Emergency Medicine, University of Colorado School of Medicine (THS, EJL), Aurora, CO.
    Search for more papers by this author
  • Vaishali Khatri MPH,

    1. From the Rocky Mountain Poison & Drug Center, Denver Health and Hospital Authority (THS, VK, LMR, EJL), Denver, CO; and the Department of Emergency Medicine, University of Colorado School of Medicine (THS, EJL), Aurora, CO.
    Search for more papers by this author
  • Liza M. Reifler MPH,

    1. From the Rocky Mountain Poison & Drug Center, Denver Health and Hospital Authority (THS, VK, LMR, EJL), Denver, CO; and the Department of Emergency Medicine, University of Colorado School of Medicine (THS, EJL), Aurora, CO.
    Search for more papers by this author
  • Eric J. Lavonas MD

    1. From the Rocky Mountain Poison & Drug Center, Denver Health and Hospital Authority (THS, VK, LMR, EJL), Denver, CO; and the Department of Emergency Medicine, University of Colorado School of Medicine (THS, EJL), Aurora, CO.
    Search for more papers by this author

Errata

This article is corrected by:

  1. Errata: Erratum Volume 19, Issue 3, 368, Article first published online: 21 March 2012

  • This investigator-initiated study was supported by a grant from Protherics Inc., manufacturer of FabAV, to the Denver Health Hospital Authority. No author or other employee of the Denver Health Hospital Authority received direct or indirect compensation as a result of this grant or study. Industry representatives did not participate in the design, execution, or writing of the study, nor did they control the decision to publish the results.

  • The authors have no potential conflicts of interest to disclose.

  • Supervising Editor: Shahriar Zehtabchi, MD.

and reprints: Tammi H. Schaeffer, DO; e-mail: tammi.schaeffer@rmpdc.org.

Abstract

ACADEMIC EMERGENCY MEDICINE 2012; 19:121–131 © 2012 by the Society for Academic Emergency Medicine

Abstract

Objectives:  Crotalidae polyvalent immune Fab (ovine) (FabAV) is commonly used in the treatment of symptomatic North American crotaline snake envenomation. When approved by the U.S. Food and Drug Administration in 2000, the incidences of immediate hypersensitivity reactions and serum sickness were reported as 0.14 and 0.18, respectively. The objective of this meta-analysis was to evaluate the incidence of immediate hypersensitivity reactions and serum sickness reported in studies of patients treated with FabAV therapy after North American crotaline envenomation.

Methods:  The authors searched PubMed, Ovid MEDLINE, and EMBASE from January 1, 1997, to September 20, 2010, for English-language medical literature and cross-referenced bibliographies of reviewed articles. The published abstracts of the major toxicology conferences were also searched. All prospective and retrospective cohort studies with patients receiving FabAV therapy for North American crotaline envenomations were eligible for data abstraction. Two content experts reviewed full-text articles and extracted relevant study design and outcome data. Proportions of immediate hypersensitivity and serum sickness for each study were analyzed in a random-effects model to produce an overall estimate of immediate hypersensitivity and serum sickness incidence associated with FabAV administration.

Results:  The literature search revealed 11 unique studies of patients who received FabAV that contained information on immediate hypersensitivity reactions and serum sickness. The meta-analysis produced a combined estimate of the incidence of immediate hypersensitivity of 0.08 (95% confidence interval [CI] = 0.05 to 0.11) and a combined estimate of the incidence of serum sickness of 0.13 (95% CI = 0.07 to 0.21).

Conclusions:  In this systematic literature review and meta-analysis, the combined estimates of the incidence of immediate hypersensitivity reactions and serum sickness from FabAV in the treatment of symptomatic North American crotaline envenomations appear to be lower than previously reported, at 0.08 and 0.13, respectively.

Snakebites are common envenomations in the United States. Envenomations are mostly from pit vipers (family Viperidae, subfamily Crotalinae, genera Crotalus, Agkistrodon, and Sistrurus). Data from the National Electronic Injury Surveillance System–All Injury Program suggest that approximately 9,000 patients are treated for snakebites in U.S. emergency departments (EDs) each year.1 In addition, data from the United States National Center for Health Statistics indicate that there are at least five deaths annually.2 In 2009, more than 3,000 snakebites were reported to United States poison centers.3 Since its approval by the U.S. Food and Drug Administration (FDA) in 2000, Crotalidae polyvalent immune Fab (ovine) (CroFab, FabAV) has become the mainstay of antivenom therapy in the treatment of North American crotaline snakebite and is the only antivenom currently available for use in this population. Forty-four percent of patients whose cases were reported to U.S. poison centers in 2007 were treated with antivenom, an increase from 30% in 2000.4 While it is unclear why exactly antivenom use has been increasing, this is likely multifactorial. Reasons may include practitioner familiarity and comfort, as well as the presumed safety of FabAV biologics.

As with any biologic product, immediate hypersensitivity and serum sickness reactions can occur following antivenom administration. FabAV was first approved for noninvestigational human use based on the results of two human clinical trials.5–7 All prospective clinical trials maintain active surveillance for adverse drug reactions throughout these studies. Taken together, these trials reported 6 of 41 patients with immediate hypersensitivity reactions; serum sickness was reported in 6 of the 34 patients who completed mandatory follow-up.5–7

Since its approval by the FDA, many patients have been treated with FabAV and our hypothesis was that the incidence of immediate hypersensitivity and serum sickness might be different than originally published. For this meta-analysis, we conducted a comprehensive search of the English-language medical literature to identify all published cohort studies involving FabAV administration to summarize the incidence of immediate hypersensitivity reactions and serum sickness as reported in these studies.

Methods

Study Design

This was a systematic literature review and meta-analysis of prospective and retrospective cohort studies.8 We used the PRISMA guidelines for performing a meta-analysis.9

Article Selection

We searched PubMed, Ovid MEDLINE, and EMBASE to identify all published articles containing primary data about North American human crotaline snake envenomations treated with FabAV. All searches were initially performed on July 28, 2008, and repeated on December 28, 2010, using the search strategy as outlined (Appendix 1). Abstracts from two major toxicology conferences (North American Congress of Clinical Toxicology and European Association of Poisons Centres and Clinical Toxicologists), Cochrane CENTRAL database, and ClinicalTrials.gov were searched and reviewed manually for snakebite and/or FabAV studies. The results of the search were not restricted as to article type.

All citations identified from the search strategy were imported into a ProCite database (Version 5.0) and duplicates were removed. We then performed an internal search to deselect articles outside the scope of the search. The terms searched for exclusion were rat(s), mouse, mice, rabbit(s), cellular, in vivo, and in vitro.

Titles and abstracts were then reviewed by a single medical toxicologist (EJL). Articles judged to have no likelihood of containing data about FabAV administration for North American crotaline envenomation (e.g., reports that clearly related only to the treatment of African, Asian, South American, or Australian snakebite) were excluded from further review. We included any articles identified from our institutional article files as well as studies from bibliographies of initially selected articles. At this stage, all articles underwent a detailed review, and any studies that were not cohort or clinical trial studies were excluded from data abstraction.

Article Analysis

Full-text versions of the resulting articles were independently reviewed by two board-certified medical toxicologists with experience in the management of snakebites (EJL, THS). Data from each study were abstracted to a standardized data collection form. This form consisted of reported data concerning immediate hypersensitivity reactions and serum sickness, summary data about any severe envenomation cases, and information for each severely envenomated patient reported (reported separately). Both reviewers completed abstraction training, which consisted of 1 hour of instruction followed by 11 practice abstractions. The reviewers were not blinded to the intent of the study. The reviewers abstracted all reports of immediate hypersensitivity and serum sickness as well as signs and symptoms the authors attributed to immediate hypersensitivity and serum sickness. To remain consistent with prior publications, the minimum observation time for signs or symptoms of serum sickness was 6 days.10 Discrepancies between the reviewers were identified and resolved by discussion and concordance. Data were entered into Microsoft Excel (Version 2003, Microsoft Corp., Redmond, WA) and double verified to ensure accuracy of data entry.

Due to the variation of the study designs of the resulting articles, three quality assessment tools were used (Table 1): the Newcastle-Ottawa Quality Assessment Scale for observational prospective studies, the Jadad scale for randomized controlled trials, and a chart review tool for retrospective studies.11–13 The items evaluated on the Newcastle-Ottawa Quality Assessment Scale included selection of cohorts, comparability of study design and analysis, assessment of outcome, and adequacy of follow-up.11 A study can achieve a maximum of nine stars and was considered to have a low risk of bias if the maximum number of stars was reached. The Jadad scale assessed the quality of randomized controlled trials by evaluating markers that are directly and indirectly related for controlling bias.12 Any study that is found to have an endorsement frequency between 15 and 85% is included, and anything outside of this range is said to provide little discriminative power. The third quality assessment was based on a study by Gilbert et al.13 This study evaluated the methods of case selection, abstractor training, monitoring and blinding, and inter-rater agreement. We applied the criteria they report to the chart review studies included.

Table 1. 
Quality Assessment of Included Studies
Prospective Cohort Studies11*
ArticleCohortRepresentativeSelection of Nonexposed CohortAscertainment of ExposureOutcome Not Present at StartComparability by Design/AnalysisOutcome AssessmentDuration of Follow-upAdequacy of Follow-upTotal No. of Stars
Dart et al. (2001)6 7
Offerman et al. (2002)15 7
Bush et al. (2002)17 7
Randomized Controlled Trials12
ArticleRandomizedDouble-blindWithdrawals/ Drop-outsDefined ObjectivesOutcome MeasuresInclusion/ ExclusionSample SizeInterventionsControl GroupAE AssessmentStatistical AnalysisEndorsement Frequency (%)
Dart et al. (1997)52011111111184.6
Retrospective Studies13
ArticleAbstractors TrainedInclusion/Exclusion DescribedVariables DefinedStandardized Abstraction FormAbstractors’ Performance MonitoredAbstractors BlindedInter-rater ReliabilityInter-rater Agreement Tested
  1. AE = adverse effects.

  2. *A star (★) is awarded for each item that the study met in the Newcastle-Ottawa Scale; maximum possible score of nine stars for the lowest risk of bias.

  3. †Any study with an endorsement frequency below 15% and above 85%

Ruha et al. (2002)16NoYesYesYesNoNoNoNo
Lavonas et al. (2004)23NoYesYesNoNoNoNoNo
Corneille et al. (2006)22NoYesYesNoNoNoNoNo
Pizon et al. (2007)20YesYesYesYesYesYesYesYes
Cannon et al. (2008)19NoYesYesYesNoNoYesYes
Nazim et al. (2008)24NoYesYesNoNoNoNoNo
Lavonas et al. (2011)10YesYesYesYesNoNoNoNo
Adherence (%)28.610010057.114.314.328.628.6

Data Analysis

Study outcome data were analyzed using Comprehensive Meta-Analysis Software (CMA, Version 2, Biostat, Englewood, NJ), and graphical representation of study rates and combined rates were created in SAS (Version 9.2, SAS Institute, Cary, NC). Two outcomes, the proportions of immediate hypersensitivity and serum sickness, were analyzed across studies. To accommodate the studies that reported zero cases with the outcomes of interest, proportions were transformed by adding 0.5 to the numerator. Using CMA, a test of heterogeneity of study outcomes was conducted. Regardless of heterogeneity test significance, random-effects models were used to estimate the across-study proportion and 95% confidence intervals (CIs). Random-effects methods were used due to the small number of studies and the possibility of systematic differences in subject selection across studies. Individual and combined estimates of outcome proportions and 95% CIs were exported from CMA to plot in SAS. Funnel plots were created to evaluate for potential publication bias. A sensitivity analysis excluding premarketing trials was also performed.

Results

The initial literature search found 3,898 citations (Figure 1), and 2,005 articles were retained after initial duplicate and title filters. Among the excluded citations, 1,804 were excluded because they did not involve the use of FabAV for treatment of a North American crotaline envenomation. There were 205 unique publications regarding FabAV that met inclusion criteria. Of these, 14 articles discussing cohort studies involving FabAV administration to defined populations were identified.5,6,10,14–24 Of these 14 publications, three were excluded: one did not systematically report adverse reactions,21 one contained combined data from two other included trials,18 and one was a presentation abstract of a subsequently published trial.14 Exclusion of these three studies left 11 studies available for review.

Figure 1.

 Summary of selection process for the systematic review.

Examination of the dates and locations of these 11 studies revealed probable overlap between four reports.15,17,19,20 Although overlap existed and was removed, some content remained in our meta-analysis sample sizes for the studies. The authors of two studies were contacted and provided information allowing us to identify five duplicate cases and remove them from summarization of outcome and sample in one of the studies during the meta-analysis.15,17 The other two studies had directly indicated the overlap, and the sample sizes of each were adjusted appropriately.19,20 Three of the 11 studies did not mention any occurrence of serum sickness.19,20,22

Four referenced abstracts were identified from the North American Congress of Clinical Toxicology annual meetings from 1997 to 2010.25–28 These abstracts were reviewed, and it was found that the data were published later in full articles. These unpublished abstracts are captured in our search strategy in Figure 1.

The included studies selected are summarized in Table 2. Three of the 11 reports used prospective data collection, seven used retrospective data collection, and one used both prospectively and retrospectively collected cases. There were a total of 661 unique patients from these 11 studies, of whom 504 received only FabAV and no other antivenom.

Table 2. 
Studies on the Incidence of Immediate Hypersensitivity Reactions and/or Serum Sickness From FabAV Therapy
ArticleStudy DesignPopulationInclusion CriteriaNo. of FabAV SubjectsNo. of Immediate Hypersensitivity CasesNo. of Serum Sickness CasesLength of ObservationGeographical LocationLimitations
  1. FabAV = Crotalidae polyvalent immune Fab (ovine).

  2. *For analysis, patients with only FabAV administered were included.

Dart et al. (1997)5Prospective, experimental, nonrandomized, open-labelAdultMinimal/moderate North American crotalid envenomation within 6 hours
≥10 years old
Symptom progression
100048 hours in hospital
7- and 14-day follow-up visits
Tucson, AZ
Phoenix, AZ
San Diego, CA
Georgia
Unblinded, noncomparative design
Observed patients with only minimal or moderate severity
Dart et al. (2001)6Prospective, experimental, randomized, open-labelAdultMinimal/moderate North American crotalid envenomation within 6 hours
≥10 years old
Symptom progression
316636 hours in hospital
2-, 4-, 7-, and 14-day follow-up visits
Albuquerque, NM
Denver, CO
Gainesville, FL
Thomasville, GA
Phoenix, AZ
San Diego, CA
Tucson, AZ
Possible contamination of one FabAV batch
Exclusion of patients with severe envenomation
Offerman et al. (2002)15Retrospective and prospective, open-labelPediatricCrotaline envenomation symptoms at presentation
FabAV administration
1200≥24 hours in hospitalSan Diego, CA
Loma Linda, CA
Not reported
Bush et al. (2002)17Prospective observational, open-labelPediatric and adultPatients with rattlesnake envenomation occurring between April 2001 and October 200192124 hours in hospital
≥21-day
Loma Linda, CANot blinded
No comparison group
Ruha et al. (2002)16RetrospectivePediatric and adultPatients between March 1and September 9, 2001, with rattlesnake envenomation
Treated with FabAV
2802Mean hospital stay of 4.2 days
48- to 72-hour follow-up visit; 21-day phone call follow-up
Phoenix, AZNot blinded
No comparison group
Lavonas et al. (2004)23RetrospectivePediatric and adultA. contortrix snake envenomation
FabAV was administered
Symptom progression at FabAV administration
3211Hospital stay range 7 hours to 8 daysNorth CarolinaRepresentative of only 8% of copperhead victims
Patients receiving FabAV were more severely affected
No comparison group with supportive care
Corneille et al. (2006)22RetrospectiveAdultSnakebite between 1995 and 2005
Treated with ACP, FabAV, or expectant management (without antivenin)
Received both ACP and FabAV*
680Not reportedHospital stay range 1–16 daysTexasNo standard grading scale for envenomation severity
Differential patient length of stays
Inconsistent snake species identification and documentation
Pizon et al. (2007)20RetrospectivePediatricPatients ≤ 13 years treated with CroFab241Not reportedHospital stay range 1–7 daysPhoenix, AZNo comparison to other treatments
No follow-up information
Represents crotaline envenomations in one geographic area
Cannon et al. (2008)19RetrospectiveAdultDiagnosis of rattlesnake bite between 2000 and 2004
Treated with FabAV
735Not reportedNot reportedPhoenix, AZChart review data collection
Underestimating the incidence of hypersensitivity reactions
Nazim et al. (2008)24RetrospectivePediatric and adultED snake and spider bite victims between January 2002 and January 20061210Not reportedWest VirginiaNot reported
Lavonas et al. (2011)10RetrospectivePediatric and adultTreated with FabAV at study hospital during study period
Had sufficient documentation to allow for a 7-point severity score
205142Not reported17 U.S. hospitals in areas of indigenous venomous snakesRetrospective study design
Inclusion criteria, FabAV dosing, concomitant therapies and clinical observations were not standardized
Variation in the level of training/experience of chart reviewers

Immediate Hypersensitivity Reactions

The combined estimate of immediate hypersensitivity incidence was 0.08 (95% CI = 0.05 to 0.11; Figure 2). The estimated incidence of immediate hypersensitivity among all included studies ranged from 0.01 (95% CI = 0.00 to 0.11) to 0.19 (95% CI = 0.09 to 0.37).

Figure 2.

 Incidence of immediate hypersensitivity reactions in patients who received FabAV (random effects model). The incidence of immediate hypersensitivity reactions in the individual studies is represented by a square (proportion) through which runs a horizontal line (95% CI).

Immediate hypersensitivity reactions were directly reported by 7 of the 11 studies.6,10,17,19,20,23,24 In four of the seven studies, a rash was the most commonly reported immediate hypersensitivity reaction.6,17,19,24 The other types of reactions that were reported included urticaria, wheezing, dyspnea, hypotension, tongue and lip swelling, tachycardia, dizziness, chest pain, and nausea.6,10,17,19,23,24

Overall, six studies reported subjects who were able to complete their entire course of FabAV therapy after an immediate hypersensitivity reaction occurred.6,10,17,19,20,23 Three studies reported a total of five patients who had their FabAV discontinued before completion following immediate hypersensitivity reactions.10,17,19

Serum Sickness

The combined estimate of serum sickness incidence across the seven studies that reported serum sickness as an adverse event was 0.13 (95% CI = 0.07 to 0.21; Figure 3). Of the 11 studies identified, eight reported some form of patient follow-up,5,6,10,15–17,23,24 while three studies did not comment on or report on serum sickness. In one of the eight studies, only one subject was followed long enough to assess serum sickness.24 The lack of systematic follow-up in this study could lead to a potentially biased estimate of serum sickness. Thus, this study was excluded from the analysis. The range of the estimate incidence was 0.06 (95% CI = 0.00 to 0.54) to 0.23 (95% CI = 0.11 to 0.43).

Figure 3.

 Incidence of serum sickness in patients who received FabAV (random effects model). The incidence of serum sickness in the individual studies is represented by a square (proportion) through which runs a horizontal line (95% CI).

Five of the studies reported cases with rash/urticaria as a frequent sign or symptom of serum sickness. The other symptoms that were reported in these studies included fever, myalgias, epigastric pressure, thrombocytopenia, anorexia, hives, and arthralgia.6,10,16,17,23 Four studies reported treatment for serum sickness. Of these four studies, three reported administration of antihistamine,6,10,16 and three reported administration of steroid treatment.6,10,23 None of the studies reported any pretreatment with medication prior to FabAV administration. One study reported serum sickness recovery with rest and no treatment for symptoms.17 None of the studies reported rehospitalization or death of a patient as a result of serum sickness.

A sensitivity analysis excluding the two premarketing trials was conducted. This removed the premarketing studies from the overall analysis and removed the study with potential fecal coliform contaminations. The estimate of hypersensitivity reduces to 0.06 (95% CI = 0.04 to 0.09), and the estimate of serum sickness reduces to 0.08 (95% CI = 0.04 to 0.16). To examine the possibility of publication bias, funnel plots of hypersensitivity and serum sickness outcomes were constructed (Data Supplement S1, available as supporting information in the online version of this paper). The funnel plot for hypersensitivity is not characteristic of publication bias as the only suggestion of underrepresentation was in the area showing studies with smaller sample size and large effects. These studies are not expected to be subject to publication bias. While not suggestive of publication bias, the plot for serum sickness does not have enough information to be able to be interpreted to any significant degree.

Discussion

In 1954, Wyeth introduced equine-derived whole immunoglobulin (IgG) antivenom for the treatment of North American crotaline envenomation. While it showed effectiveness in the treatment of venom effects from these snakes, the incidences of both immediate hypersensitivity reactions and serum sickness were high, ranging from 23% to 56%.29–31

Antigen-specific antibody fragments were investigated in the hopes of decreasing the immediate hypersensitivity reactions noted with a full IgG protein. These fragments were derived from the IgG, in the case of FabAV, of sheep immunized with the venom of four different crotaline snakes: Crotalus atrox (Western Diamondback rattlesnake), C. adamanteus (eastern diamondback rattlesnake), C. scutulatus (Mojave rattlesnake), and Agkistrodon piscivorus (cottonmouth or water moccasin). The recovered antibodies are cleaved by papain to yield two antigen-binding fragments: Fab and the more immunogenic Fc portion. Column-affinity purification is then used to remove the Fc portion, any Fab molecules that lack avidity for snake venom antigens, and other excess protein. As a result, the Fab is purified and is specific to the toxin from which it is derived. Highly purified FabAV is expected to result in fewer immediate hypersensitivity reactions than the previous generation of whole IgG preparation.32 Since the process of isolation of Fab fragments was developed, their use in pharmaceuticals has increased significantly. Although Fab fragments were investigated in the treatment of such toxins as colchicine and desipramine, besides FabAV, it has been most used in the treatment of digoxin toxicity with digoxin-specific antibodies.33,34

Our summary of all available FabAV studies suggests that the combined estimate incidence of immediate hypersensitivity reactions is 0.08 (95% CI = 0.05 to 0.11). As mentioned earlier, the package insert reports that the estimate of hypersensitivity based on the two clinical trials was 0.14 (95% CI = 0.07 to 0.28). Overall incidence of immediate hypersensitivity, estimated using the outcomes of the two clinical trials plus outcomes of nine other observational trials, suggests that the incidence may not be as high as initially reported. This is also noted when comparing the combined estimate across seven studies of serum sickness incidence of 0.13 (95% CI = 0.07 to 0.21), compared to the initial estimate of 0.18 (95% CI = 0.08 to 0.34) reported in the package insert.

In our methods, we excluded non–North American crotalids to better reflect experience in the United States and to decrease the possibility that reactions could have been due to higher free-antivenom levels from incomplete binding of antigen antivenom binding of nontarget snakes.

Immediate Hypersensitivity Reactions

The term immediate hypersensitivity reaction has historically been used to describe hypersensitivity reactions mediated by immunoglobulin E (IgE). This immune-mediated response is seen in patients who have been previously sensitized to a given agent with the end result of anaphylaxis. Often clinically indistinguishable from anaphylaxis, anaphylactoid reactions are not immune-mediated, but are triggered by direct mast cell activation and do not require prior sensitization. Although it is impossible to prove without immunologic testing, our assumption is that most patients reviewed in our study had not received antivenom prior to the current exposure and were unlikely to be sensitized to sheep proteins. Even though these reactions were likely to be mostly anaphylactoid, the term immediate hypersensitivity is used to encompass both types of reactions.

Prospective collection of medical adverse events may yield an increase in reportable adverse events.35 Three of the four studies with the highest incidences of immediate hypersensitivity either were prospective or had a component of prospective data collection.5,15,17 The rate of the fourth prospective, experimental study fell within the middle of the range of the other studies.5 While it is doubtful that a serious immediate hypersensitivity reaction would not be captured, it is difficult to determine if minor, less significant reactions went unreported in the retrospective cohort studies from which the majority of our data are derived.

The prospective premarketing trials of digoxin-specific antibody showed essentially no hypersensitivity adverse events, while the later retrospective studies showed varying adverse events. However, none of these adverse events were described as life-threatening.36–39 Although the incidence of immediate hypersensitivity reactions is higher with FabAV than with digoxin-specific Fab, the mildness and manageability of the reactions to FabAV mirror those of digoxin-specific antibody. Furthermore, the tolerability of both the digoxin-specific antibody and FabAV, with the consistent lack of significant immediate hypersensitivity reactions, has made these agents a staple in the treatment of the poisoned patient. Although rare, airway complications have been reported with FabAV use.10,40

Serum Sickness

Type III hypersensitivity, or serum sickness, is a complement-mediated reaction that occurs 4 to 10 days after the introduction of foreign proteins or haptens. The frequency and severity of serum sickness are associated with the dose of protein/hapten and can vary by the specific antigen.29,41 Symptoms can include rash, fever, lymphadenopathy, and arthritis or arthralgias. In rare, more severe cases, there can be renal, cardiovascular, neurologic, or pulmonary manifestations.

As was the case with immediate hypersensitivity reactions, prospective data collection may have identified more serum sickness cases than studies with retrospective design. Unlike immediate hypersensitivity reactions, which may be observed by the clinician at the hospital bedside, serum sickness generally develops after a patient is discharged from the hospital. Obtaining consistent follow-up on snake envenomation patients is challenging under the best of circumstances; even in the premarketing studies, 20% of patients were lost to follow-up.5,6 Although it has never been formally studied, it is likely that follow-up rates are biased by the presence of serum sickness symptoms. Given that adequate follow-up is only available for 29% of the FabAV-treated patients in these 11 cohort studies, this problem weakens our ability to obtain a true estimate of serum sickness incidence.

It is likely that the incidence of serum sickness reported in the FabAV prescribing information was inflated by a manufacturing defect that affected the study drug used in the second premarketing trial.6 Of the six patients in the second premarketing trial who developed serum sickness, five patients (83%) received FabAV from a batch contaminated with fecal coliform fragments.6 The sensitivity analysis demonstrates even lower rates of immediate hypersensitivity and serum sickness when excluding the premarketing studies. This difference could be multifactorial, in that it may be the result of the fecal coliform contamination or even that they were prospective studies.

With the advent of safer Fab products, practitioners appear to be more and more comfortable with their use, as evidenced by the increased use of digoxin-specific Fab despite an apparent decrease in digoxin toxicity.42 With the estimates of reactions to crotalidae FabAV being lower than previously described, practitioners may be more likely to treat with FabAV if one of their personal deciding factors is hypersensitivity incidence. Nonetheless, the indications for use of FabAV after snakebite are clinical and laboratory parameters, and it should not be used without meeting those criteria despite a low risk for immediate hypersensitivity and serum sickness.7

Limitations

All meta-analyses have limitations that must be taken into consideration during analysis and interpretation. This meta-analysis is limited by the inclusion of both retrospective and prospective data, as well as each study’s limitations (Table 2). It is difficult to calculate a more precise incidence of immediate hypersensitivity because of the small number of studies and the small sample sizes in this series. We attempted to limit publication bias by conducting an electronic search, as well as a manual search of potentially relevant references of retrieved articles and unpublished article abstracts. These factors may lead to additional variability that cannot be explained in this analysis. Dart et al.6 met Jadad criteria and was left in the analysis. All of the observational studies met Newcastle-Ottawa criteria, except for the “unexposed cohort” which none of the studies included.

Chart review in emergency medicine was shown to have the lowest percentage of adherence to systematic methods.13 However, some variables such as a standardized abstraction form may have been used but not specifically mentioned as such in the article. This may explain the low adherence percentage for some of the variables that were reviewed. While it is difficult to know how this affected quality, chart review is commonly used in emergency medicine and toxicology and is a frequent study type.

We did not address whether the immediate hypersensitivity reactions were from the antivenom rather than the venom itself. Although temporally all reactions occurred after the FabAV was administered, some of them hours from envenomation, laboratory evidence is the only way to confirm whether the immediate hypersensitivity reactions came specifically from the antivenom.

We also did not address the variables that may affect immediate hypersensitivity and serum sickness, such as severity and snake species. While important, our goal was to report the overall incidence, and evaluating specific parameters and variables were beyond the scope of this paper.

Conclusions

The combined estimate of incidence of immediate hypersensitivity reaction to Crotalidae polyvalent immune Fab (ovine) across 11 cohort studies was 0.08 (95% CI = 0.05 to 0.11), and the combined estimate of incidence of serum sickness from Crotalidae polyvalent immune Fab (ovine) across seven studies was of 0.13 (95% CI = 0.07 to 0.21). These estimates of incidence are lower than reported in the early premarketing studies alone.

The authors thank Sara Mlynarchek, MPH, Jamie Kokko, MPH, Becki Bucher-Bartelson, PhD, and Greg Bogdan, PhD, for their time and effort in the initial abstraction, statistical analysis, and review.

Appendix

APPENDIX 1

Search Strategies and Results

DatabaseSearch TermsDate of Final SearchCitations Retrieved
  1. Medical Subject Headings (MeSH): Crotalid Venoms/po [Poisoning], Crotalid Venoms/to [Toxicity], Snake Venoms/po [Poisoning], Snake Venoms/to [Toxicity], Snake Bites/dt [Drug Therapy], Snake Bites/th [Therapy]

  2. EMBASE: “Crotalid venoms” AND (intoxication OR toxicity), “Snake venoms” AND (intoxication OR toxicity), “Snake bites” AND (“drug therapy” OR therapy); NA indicates not applicable.

PubMed (01/01/1999–12/28/2010)Crotalid Venoms, Snake Venoms, Snake Bites, Viperidae, Agkistrodon, Crotalus (MeSH key word) or [CroFab or Crotaline immune Fab or FabAV].mp
Limit: English language, human
12/28/2010935
Ovid MEDLINE (1999–2010)Crotalid Venoms, Snake Venoms, Snake Bites, Viperidae, Agkistrodon, Crotalus (MeSH key word) or [CroFab or Crotaline immune Fab or FabAV].mp
Limit: English language, human
12/28/20101,021
EMBASE* (1999–2010)Crotalid venoms, snake venoms, snake bites, Viperidae, Agkistrodon, Crotalus, Crotaline immune Fab, FabAV
Limit: English language, human
12/28/20101,938
TotalN/AN/AN/A

Ancillary