A Randomized, Double-Blind, Placebo-Controlled Trial of a Highly Purified Equine F(ab)2 Antibody...

This article appeared in a journal published by Elsevier. The attachedcopy is furnished to the author for internal non-commercial researchand education use, including for instruction at the authors institution

and sharing with colleagues.

Other uses, including reproduction and distribution, or selling orlicensing copies, or posting to personal, institutional or third party

websites are prohibited.

In most cases authors are permitted to post their version of thearticle (e.g. in Word or Tex form) to their personal website orinstitutional repository. Authors requiring further information

regarding Elsevier’s archiving and manuscript policies areencouraged to visit:

http://www.elsevier.com/authorsrights

http://www.elsevier.com/authorsrights

Author's personal copy

TOXICOLOGY/ORIGINAL RESEARCH

A Randomized, Double-Blind, Placebo-Controlled Trial of aHighly Purified Equine F(ab)2 Antibody Black Widow

Spider Antivenom

Richard C. Dart, MD, PhD; Gregory Bogdan, PhD; Kennon Heard, MD; Becki Bucher Bartelson, PhD;Walter Garcia-Ubbelohde, MD; Sean Bush, MD; Tom Arnold, MD; Richard C. Clark, MD; Gregory W. Hendey, MD;

Christopher Holstege, MD; Elizabeth A. Spradley, MNM

Study objective: Black widow spider antivenom has never been tested in a randomized clinical trial, to ourknowledge. We explore various efficacy measures for a novel F(ab)2 antivenom in patients with moderate tosevere pain caused by black widow spider envenomation.

Methods: A randomized, placebo-controlled, double-blind, clinical trial was conducted in 12 academic emergencydepartments. We included patients at least 10 years old with moderate to severe latrodectism. Subjects receivedeither a single intravenous infusion of antivenom or placebo. Pain was assessed with the visual analog scale. Theprimary efficacy outcome was the difference in pre- and posttreatment visual analog scale score. Prospectivelydefined secondary outcomes included treatment failures and time to clinically important decrease in pain.

Results: Twenty-four subjects were enrolled between October 2005 and October 2006; 13 were randomized toantivenom and 11 to placebo. The median change in visual analog scale at 150 minutes posttreatment was�50.0 mm (Interquartile Range [IQR] �67, �41 mm) in the antivenom treatment group and �46.0 mm (IQR�51, 0 mm) in the placebo treatment group (P�.14). There were 7 treatment failures (64%; 95% confidenceinterval 35% to 92%) in the placebo group and 3 (23%; 95% confidence interval 0.2% to 46%) in the antivenomgroup (P�.06). The median time to a clinically important decrease in pain after treatment was shorter in theantivenom group compared with the placebo group (30 minutes [IQR 30, 60 minutes] versus 90 minutes [IQR30, 90 minutes]; P�.03). No serious adverse events or deaths were reported.

Conclusion: Although the overall reduction in pain was similar for antivenom- and placebo-treated subjects,antivenom reduced pain more rapidly than placebo. No significant adverse events occurred in either group. [AnnEmerg Med. 2013;61:458-467.]

Please see page 459 for the Editor’s Capsule Summary of this article.

A feedback survey is available with each research article published on the Web at www.annemergmed.com.A podcast for this article is available at www.annemergmed.com.

0196-0644/$-see front matterCopyright © 2012 by the American College of Emergency Physicians.http://dx.doi.org/10.1016/j.annemergmed.2012.10.008

INTRODUCTIONBackground

In 2008, poison centers received 2,303 calls reporting blackwidow spider bites, of which 950 cases were treated in a healthcare facility.1 Many emergency physicians in endemic areas arefamiliar with the treatment of envenomation and may notcontact poison centers for assistance; thus, the actual number ofbites is likely underreported.

The clinical effects of black widow spider venom include pain atthe bite site, which can become severe and migrate to other parts ofthe body. The pain can mimic an acute abdominal condition ormyocardial infarction. A syndrome of clinical effects caused byblack widow spider bite is termed latrodectism: pain, anxiety,

agitation, and autonomic dysfunction manifested by tachycardia,hypertension, and diaphoresis.2-4

Antivenom for black widow spider bite was first licensed in theUnited States in 1936 by Merck & Co, Inc. as antiveninLatrodectus mactans (black widow spider antivenin). It is producedby partial purification of serum Immunoglobulin G (IgG) fromhorses immunized with L mactans venom. Although consideredeffective, it has not been tested in a prospective clinical trial. Itssafety has been questioned because of 2 deaths from anaphylaxisassociated with its use.2,5 Currently, Merck has limited inventory ofits antivenom and has established limitations on distribution.6,7

Alternative treatments include opioid analgesics,benzodiazepines, muscle relaxants, and calcium gluconate.

458 Annals of Emergency Medicine Volume , . : April


Previous research suggests that calcium gluconate and musclerelaxants have limited effectiveness and often fail to producesustained relief. Opioids and benzodiazepines, althougheffective, carry a risk of adverse events and typically requirerepeated administration in patients with a severeenvenomation.8,9 Patients treated with opioid analgesics andbenzodiazepines are likely to require repeated parenteraladministration and are therefore hospitalized more often thanpatients treated with antivenom.2

ImportanceAlthough thousands of people are envenomated by

black widow spiders (L mactans, L hesperus) in the UnitedStates each year,1 few studies have addressed the treatment ofenvenomation, and treatment options remainsuboptimal.2,4,8,10,11

Antivenin Latrodectus (Black Widow) Equine ImmuneF(ab)2 (Analatro) is commercially available in Mexico but hasnot been approved for use in the United States. This antivenomis a highly purified equine F(ab)2 antibody preparation. PurifiedF(ab)2 products are expected to be less immunogenic thanwhole antibody products and may convey a safety advantage inlatrodectism treatment.12-15

Goals of This InvestigationWe conducted a randomized, placebo-controlled, phase 2

clinical trial to assess various measures for comparison of safety andefficacy of a novel F(ab)2 antivenom in patients with moderate to

severe pain caused by black widow spider envenomation. Theprimary outcome was the magnitude of decrease in the self-reported visual analog scale pain score 150 minutes after infusion ofantivenom. Two secondary outcomes were the incidence oftreatment failure and the time to clinically important decrease inpain.

MATERIALS AND METHODSStudy Design

This was a multicenter, randomized, double-blind, placebo-controlled, clinical trial comparing the analgesic efficacy of a singleintravenous infusion of Antivenin Latrodectus (Black Widow)Equine Immune F(ab)2 (Analatro). The antivenom dose consistedof 3 vials reconstituted in 50 mL normal saline solution. Placebowas 50 mL normal saline solution. The study was approved by theinstitutional review board of all participating sites, and all patientsprovided informed consent before enrollment.

The coordinating clinical site for the trial generated therandomization schedule, facilitated study drug packaging,performed site monitoring, and conducted statistical analysis.Site investigators and study team members obtained informedconsent and enrolled subjects at each site, completed case reportforms, and performed telephone follow-up on days 2, 10, and21. The pharmacist at each site was responsible for study drugpreparation and ongoing drug accountability.

SettingThe trial was conducted between October 2005 and October

2006 at 12 academic emergency departments (EDs) in areas of theUnited States endemic for black widow spider species (L mactans, Lhesperus). Study procedures were completed in the ED, withscripted follow-up questionnaires completed by telephone on days2, 10, and 21 after the subject’s discharge from the treatmentfacility.

Selection of ParticipantsEligible subjects were aged 10 years or older who presented for

treatment within 72 hours of symptom onset related to presumedblack widow spider bite. Subjects required a clinical diagnosis oflatrodectism by a study investigator, with concurrence of diagnosisby an independent (non–study-affiliated) physician. Each studyinvestigator has extensive clinical experience in the diagnosis andmanagement of black widow spider bite. Subjects had to presentwith a visual analog scale pain score of at least 40 mm duringscreening and have a baseline score of at least 40 mm immediatelybefore infusion of antivenom or saline solution control. Spiderexposure was classified as one of 4 categories: putative: a spider wasneither found nor suspected to be in the area of the patient beforesymptoms onset; presumptive: a spider may have been in the areaof the patient before symptoms onset but was not found; plausible:a spider was seen in the area of the patient before symptoms onsetor at time of the bite but was not clearly identified as a black widowspider; and confirmed: the spider was brought to the hospital andidentified as a black widow spider.

Editor’s Capsule Summary

What is already known on this topicBlack widow spider envenomations are painful,although rarely life threatening. There is no readilyavailable safe, effective, specific therapy.

What question this study addressedIs a purified F(ab)2 antivenom that is in early stagesof clinical development safe and effective forpatients with black widow spider envenomation?

What this study adds to our knowledgeIn this 24-patient randomized trial of symptomatic,clinically diagnosed black widow spider bites,treatment with the new antivenom shortened thetime to a clinically important reduction in painseverity and decreased pain at each point by anamount that did not reach statistical significance,likely because of low power.

How this is relevant to clinical practiceThis study will not directly affect practice becausethe antivenom is not yet available for clinical use.

Dart et al Equine F(ab)2 Antibody Black Widow Spider Antivenom

Volume , . : April Annals of Emergency Medicine 459


Exclusion criteria included a history of asthma or a knownhypersensitivity to fentanyl, morphine, diazepam, or equineserum; a history of significant cardiac, respiratory, hepatic, orrenal disease; psychiatric disorder or chronic pain syndrome thatin the investigator’s assessment would obscure the efficacy orsafety endpoints; inability to give informed consent or inabilityto provide assent if a minor; pregnancy; a distracting injury withacute pain; or inability or unwillingness to provide a telephonenumber to be contacted for follow-up questionnaires.

InterventionsEnrolled subjects were randomized to receive either a single

10-minute intravenous infusion of 3 vials of AntiveninLatrodectus (Black Widow) Equine Immune F(ab)2 (not lessthan 600 lethal dose 50% per vial; batch P-4J-08; InstitutoBioclon S.A. de C.V., Mexico City, Mexico) reconstituted in 50mL normal saline solution or a single dose of 50 mL normalsaline solution (0.9% sodium chloride USP).

A block randomization scheme was computer generatedbefore study initiation. Blocks were produced containing 4study kits: 2 kits containing antivenom and 2 containing salinesolution control. Each kit was numbered sequentially withineach block (eg, 1, 2, 3, 4) before delivery to an investigative site.When a patient was enrolled, the investigational pharmacist atthe treatment site selected a prelabeled study kit sequentiallyfrom the site’s block for assignment to the subject.

Antivenom and saline solution control treatment kits wereindistinguishable according to exterior packaging. Both salinesolution control and antivenom infusions were transparent,colorless fluids and were tested to ensure that they wereindistinguishable during clinical use. A tab was placed over the portof the intravenous bag to maintain the integrity of the study blind.All subjects and clinicians were blinded to treatment assignment.The investigational pharmacist at the treatment site was unblindedbecause of his role in study drug preparation.

Subjects could request to receive bolus injections of fentanyl forpain relief both before treatment infusion and during the 150-minute postinfusion observation period. To allow a 30-minutewashout between fentanyl doses and pain assessments, the doses offentanyl were planned for administration immediately after visualanalog scale score determination. The maximum fentanyl dose was1.5 �g/kg per hour. No other analgesics were permitted.

Subjects were deemed to have failed treatment and enteredthe rescue arm if any of the following criteria were met:● the visual analog scale pain score at 60, 90, or 120 minutes

after the start of the infusion was greater than the baselinevisual analog scale pain score;

● a clinically important decrease in the visual analog scale painscore of at least 13 mm was not achieved between baseline andthe end of the 150-minute postinfusion observation period; or

● the physician concluded that the subject’s clinical condition hadworsened or that continued participation was not appropriate.

Subjects not achieving adequate pain control after havingreceived the randomized treatment entered a rescue arm andreceived the Food and Drug Administration–approved widow

spider antivenom (Merck antivenin Latrodectus mactans) ifclinically indicated.

Methods of MeasurementThe primary efficacy outcome was the difference in pre- and

posttreatment pain intensity (change in visual analog scale score),comparing the pain score measured immediately before study druginfusion to the pain score measured at the end of the 150-minuteposttreatment observation period. Visual analog scale scores weredetermined at baseline and then every 30 minutes after the start ofthe infusion to the end of the 150-minute observation period,which was selected because previous case series suggest that patientsrespond to antivenom within 120 minutes.2

A visual analog scale was selected as the assessment toolfor patient self-report of pain levels. It has been validated as areliable tool for assessments of acute pain through its use instudies of multiple types of pain in the ED patientpopulation16-18 and is a widely accepted, easily administered,semiquantitative measurement tool for assessing painintensity in persons with a wide variety of acutely painfulconditions.16,19-21

The paper-based visual analog scale consisted of the instruction“Put a mark on the line at the point that describes how much painyou are having right now” with “No Pain” and “Worst PossiblePain” anchoring the left and right ends of a 100 mm horizontalline, respectively. All visual analog scale scores were self-reported bythe subject with this assessment tool, which included instructionprovided in both English and Spanish. Each visual analog scalescore was independently measured with a ruler to the nearest 0.5mm by 2 trained observers, and the 2 scores were averaged andthen rounded to the nearest 1.0 mm. Vital signs and clinicalsymptoms were also assessed at 30-minute visual analog scaleintervals. Blood samples for antivenom measurement were collectedat baseline (before treatment infusion) and at 30 and 150 minutesafter treatment initiation.

The secondary efficacy endpoint was the incidence oftreatment failures. Subjects were deemed to have failedtreatment according to specific criteria established in theprotocol: if they entered the rescue arm, if they returned to ahealth care facility within 24 hours after discharge and receivedprescription pain medication administered in the hospital, or ifthey were treated with prescription pain medication thatprolonged their stay in the hospital after completion of thepostinfusion observation period but before discharge. Theclinician caring for the patient (not the research investigator)made all decisions about the prescription of analgesics or theadministration of additional treatment after study completion.

The third efficacy endpoint was the time to the first clinicallyimportant decrease in pain, defined as a decrease of at least 13mm in posttreatment visual analog scale pain score comparedwith baseline visual analog scale measurement. If no significantdecrease in pain score was observed during the posttreatmentobservation period or if the subject entered the rescue arm, 150minutes was assigned.

Equine F(ab)2 Antibody Black Widow Spider Antivenom Dart et al



The primary safety endpoint was the proportion of subjectsexperiencing 1 or more treatment-related adverse events, whichincluded any adverse event classified as “definitely,” “probably,”or “possibly” related to treatment. Adverse events were codedwith the Medical Dictionary for Regulatory Activities(MedDRA) (version 11; International Conference onHarmonization of Technical Requirements for Registration ofPharmaceuticals for Human Use - Maintenance and SupportServices Organization (MSSO); Chantilly, VA) and presentedby system organ class and preferred term.

Observation of all subjects continued until discharge (ortransfer) from the ED (minimum of 150 minutes postinfusionof study drug and up to an additional 150 minutes postinfusionof rescue drug for treatment failures). Telephone follow-upsusing scripted questionnaires were conducted at 2, 10, and 21days after study treatment to determine adverse events,medication use, and recurrent signs and symptoms.

Data Collection and ProcessingData were recorded in original source medical records and

transcribed onto standardized, paper-based case report forms bytrained research coordinators or investigators at each site. Eachcase report form was source data verified and monitored by atrained clinical researcher who worked for the studycoordinating center. All data were double entered into adatabase with edit checking.

Plasma samples were analyzed for F(ab)2 by themanufacturer of the study drug, using a previously describedmethod.22 This immunoassay has a reported limit of detectionof 3 ng/mL, a limit of quantification of 40 ng/mL, and acoefficient of variation of less than 20%. Plasma samples werecollected for F(ab)2 at baseline, 30 minutes, and 150 minutes(unless the subject entered the rescue arm).

Primary Data AnalysisBecause this was a phase 2 study, our primary focus was

measuring direction and magnitude of effect rather thanconducting formal statistical comparisons between placebo andantivenom. Therefore, our prespecified statistical analysis wasperformed with a 1-sided hypothesis tests for the 3 outcomes.Median change in visual analog scale pain scores at 150 minutespostinfusion were compared between the antivenom andplacebo groups with a 1-tailed Mann-Whitney U test. Missingvisual analog scale pain scores were replaced with the lastobservation carried forward method. The proportion oftreatment failures was compared with a 1-tailed Fisher’s exacttest. The time to clinically important decrease in pain wascompared with a 1-tailed Mann-Whitney U test.

Although not prespecified in the study protocol, a mixedmodel was used to compare the mean visual analog scale timeresponse curves for the 2 treatment groups. The mixed modelincluded fixed effects for treatment group, time, and theirinteraction. An unstructured variance-covariance matrix wasassumed. Given significance of the interaction term, contrasts oftreatment group means at each 30-minute visual analog scale

interval were computed with 2-sided tests (a Fisher’s protectedmultiple comparisons strategy). Total fentanyl dosage afterrandomization was compared with a 2-sided Mann Whitney Utest. Per protocol, a 2-tailed Fisher’s exact test was used tocompare the frequencies of subjects experiencing adverse eventsclassified as definitely, probably, or possibly related to treatmentintervention for the primary safety endpoint. Data wereanalyzed with SAS (version 9.1; SAS Institute, Inc., Cary, NC).

RESULTSCharacteristics of Study Subjects

Disposition of the 26 subjects who consented to participatein this study is shown in Figure 1. One subject was excludedbecause of a positive pregnancy test result and another forfailing to achieve moderate to severe pain levels (visual analogscale pain score �40 mm) at screening and baseline painassessments. All randomized subjects were included in both thesafety and efficacy analyses.

Demographic and baseline envenomation characteristics weresimilar between the antivenom and placebo groups, except for alower proportion of women in the placebo group (Table 1). Mostsubjects reported similar symptoms at presentation (Table 2).Fatigue, weakness, localized area of erythema, mild hypertension,and nausea or vomiting were reported by at least half the subjects.

Main ResultsThe visual analog scale pain score decreased in both the

antivenom and placebo groups at 150 minutes (Figure 3, Table3). The median decrease in the antivenom group wasnumerically greater than that in the placebo group but did notachieve statistical significance.

The time course of the visual analog scale was compared forthe 2 treatment groups (Figure 2). The interaction of time andtreatment group was significant (P�.001), affirming theobservation that visual analog scale pain levels appeared todecrease more quickly in the antivenom treatment group thanin the placebo treatment group. Treatment group meansdiffered significantly only at the 60-minute point (meandifference�–26.7; standard error�12.9).

There were 7 treatment failures in the placebo group (64%;95% confidence interval [CI] 35% to 92%) and 3 in theantivenom group (23%; 95% CI 0.2% to 46%) (P�.06).Within the placebo group, 4 subjects failed treatment because ofno or insufficient improvement in visual analog scale painscores, 2 failed because they returned to the hospital for painmedication within 24 hours of discharge, and 1 subject receivedprescription pain medication after the postinfusion observationperiod before discharge. No other subjects returned to thehospital. Reasons for treatment failures in the antivenom groupincluded no or insufficient improvement in visual analog scalepain scores (n�2) and prescription of a pain medication afterthe observation period but before discharge (n�1).

The median time to first clinically important decrease in visualanalog scale pain score (a posttreatment improvement of �13 mmfrom baseline) was significantly shorter in the antivenom group (30




minutes; IQR 30, 60 minutes) compared with the placebo group(90 minutes; IQR 30, 90 minutes) (P�.03).

Few subjects received fentanyl after infusion of study drug (3in the antivenom group; 2 in the placebo group). The mediantotal fentanyl dose was zero for both treatment groups.

No deaths or serious adverse events occurred during this study.A total of 90 treatment-emergency adverse events were observed,involving 22 of 24 subjects. No significant differences wereobserved in occurrence or relatedness of adverse events within the 2treatment groups. No adverse events were judged by the treatingphysician to be definitely or probably related to antivenom orplacebo.

In the antivenom treatment group, 6 (46.2%) subjects had atleast 1 adverse event that was potentially related to treatment,whereas in the placebo group, 4 (36.4%) subjects had at least 1adverse event that was potentially related. The most commonpotentially related events were gastrointestinal disorder (20.8%),

skin and subcutaneous tissue disorders (20.8%), and generaldisorders and administration site conditions (20.8%).

Table 4 lists the number and percentage of subjectsexperiencing any treatment-emergency adverse event byMedDRA system organ class and preferred term. Pruritus(37.5%) was the most frequently reported event, followed bynausea (29.2%) and arthralgia (25.0%).

Analysis of plasma samples showed median F(ab)2concentrations at time 0 were 0 �g/mL in both groups. Medianconcentrations at time 30 were 0 �g/mL (range 0 to 1.21 �g/mL) and 1.46 �g/mL (range 0.71 to 3.09 �g/mL) in theplacebo and antivenom groups, respectively. Seven placebosubjects had samples collected at 150 minutes (the remainingsubjects had entered the rescue arm), and no samples hadmeasureable F(ab)2 concentrations. In the antivenom group, themedian concentration of F(ab)2 at 150 minutes was1.23 �g/mL (range 0.58 to 3.09 �g/mL).

Figure 1. Disposition of subjects.




The serum F(ab)2 measurement results yielded unexpectedbaseline values in 2 subjects, 1 placebo (time 0�1.63 �g/mL)and 1 antivenom (time 0�1.99 �g/mL). Exact times of bloodsample draws were not recorded, so it could not be confirmedwhether these samples were collected after treatment infusion.The antivenom subject reported receiving Merck antiveninLatrodectus for his spider bite approximately 56 hours beforeconsenting for trial participation, likely explaining the presenceof detectable F(ab)2 at baseline. The placebo subject was notexpected to show F(ab)2 at any of the measurement time pointsyet had detectable levels at baseline and 30 minutes (time0�1.63 �g/mL; time 30�1.21 �g/mL). All other subjectsshowed expected proportional increases of F(ab)2 concentrationat points subsequent to antivenom infusion or nondetectionsubsequent to placebo infusion.

There are several potential explanations for these unexpectedfindings. The F(ab)2 assay is experimental, and it is possible thatthere was an error during sample handling or during analysis thatproduced false-positive results. It is also possible that the subjectrandomized to placebo received antivenom instead. Alldocumentation at the investigator’s site (drug accountability recordsand the subject’s case report form) indicated that this subject didreceive placebo. However, we cannot exclude that an error occurredand the subject inadvertently received antivenom, which wouldresult in misclassification of a subject who did not respond to

treatment. The most conservative analysis would be to consider thesubject’s results a treatment failure in the antivenom group.

A sensitivity analysis was performed to explore the potentialeffect of misclassification. For this analysis, the subject wasmoved from the placebo group to the antivenom group and theoutcome was considered a treatment failure. In this analysis, themedian change in visual analog scale score for the placebo groupwas –46.5 mm (IQR �51, �18), and for the antivenom groupit was �48.5 mm (IQR �67, �41) (P�.53). The proportionof treatment failures in the placebo group changed to 6 of 10(60.0%; 95% CI 30% to 90%) and the proportion in theantivenom group changed to 4 of 14 (28.6%; 95% CI 5% to52%). Finally, in the sensitivity analysis the median time toclinically important change in visual analog scale scores was 75minutes (IQR 30, 90 minutes) for the placebo group and 30minutes (IQR 30, 60 minutes) for the antivenom group, andthis difference was no longer statistically significant. Thisanalysis decreased the significance of our effect but did notchange the direction of a treatment effect.

There were a total of 23 protocol deviations reported in 11subjects (5 in the antivenom group and 6 in the placebo group).One subject in the placebo group was categorized as having atreatment failure at 60 minutes postinfusion but was not movedto the rescue arm until after the 120-minute postinfusionassessment point. Two protocol violations (1 placebo, 1

Table 1. Demographic and envenomation characteristics.

Variable Statistic/Response Antivenom (n�13) Placebo (n�11) Total (n�24)

Age, y Mean (SD) 31.2 (14.4) 33.0 (16.6) 32.0 (15.1)Median 29.0 29.0 29.0Range 13.0–54.0 11.0–60.0 11.0–60.0

Sex, No. (%) Male 6 (46) 10 (91) 16 (67)Female 7 (54) 1 (9) 8 (33)

Race, No. (%) White 7 (54) 3 (27) 10 (42)Black 0 1 (9) 1 (4)Hispanic 5 (38) 6 (55) 11 (46)Other 1 (8) 1 (9) 2 (8)

Screening VAS pain score, mm Mean (SD) 77.5 (15.8) 70.6 (17.9) 74.3 (16.8)Median 77.0 68.0 76.5Range 48.0–98.0 46.0–97.0 46.0–98.0

Certainty of black widow spider envenomation, No. (%)* Putative 1 (8) 0 1 (4)Presumptive 2 (15) 2 (18) 4 (17)Plausible 5 (38) 7 (64) 12 (50)Confirmed 5 (38) 2 (18) 7 (29)

Bite location, No. (%) Upper extremity 6 (46) 5 (45) 11 (46)Lower extremity 5 (38) 3 (27) 8 (33)Head or trunk 2 (15) 1 (9) 3 (13)Unknown 0 2 (18) 2 (8)

Time to presentation, h Mean (SD) 10.3 (17.99) 6.3 (12.73) 8.5 (15.61)Median 5.5 2.7 4.0Range 1.0–69.0 0.5–44.3 0.5–69.0

Time to symptom onset, h Mean (SD) 1.2 (1.73) 0.4 (0.72) 0.8 (1.39)Median 0.3 0.2 0.2Range 0–5.0 0–2.5 0–5.0

*Putative: A spider was neither found nor suspected to be in the area of the patient before symptoms onset. Presumptive: A spider may have been in the area of thepatient before symptoms onset but was not found. Plausible: A spider was seen in the area of the patient before symptoms onset or at the time of the bite but wasnot clearly identified as a black widow spider. Confirmed: The spider was brought to the hospital and identified as a black widow spider.




antivenom) involved the administration of greater than theallowable amount of fentanyl (per protocol, maximum allowablefentanyl dose was 1.5 �g/kg per hour). The placebo patientreceived a total of 2.58 �g/kg per hour between times 60 and150 minutes. The antivenom patient received a total of 1.87�g/kg per hour between times 60 and 150 minutes. Sevenprotocol violations involved pain assessments less than 30minutes after fentanyl dosing. One control patient was assessedat 15 minutes and another at 20 minutes post–fentanyladministration. One control patient had 3 violations (intervalsbetween fentanyl and pain assessment of 23, 12, and 20minutes). One antivenom subject was assessed at 5 minutespost–fentanyl administration and another treatment patient hadfentanyl administered between visual analog scale score 1 and

study drug administration. Her pain assessment was performed38 minutes after she received the fentanyl dose.

LIMITATIONSInterpretation of data for safety and efficacy conclusions is

limited by the small sample size of the phase 2 trial. The safetyoutcome measures required that subjects maintain contact withthe site research personnel through day 21 postdischarge; therewere 2 instances in which subjects were lost to follow-up afterED discharge. Efficacy may have been affected by imbalance inbaseline characteristics. One clear difference was the time fromreported envenomation to presentation at the treating healthcare facility. The 4-hour difference in time to presentation mayreflect a difference in disease severity or could alter the efficacyof the treatment or the natural progression of disease during thestudy. These issues must be addressed in the larger phase 3 trial.

Protocol noncompliance introduces possible inconsistencieswithin the subject data considered in efficacy analysis. The mostcommon violation was early pain assessment relative to fentanyladministration. These violations occurred in both the placeboand antivenom groups and may have contributed additionalvariability that would bias our study toward finding nodifference between groups.

Our study is limited by the inherent difficulty in diagnosinglatrodectism. The diagnosis is straightforward when a patientpresents with symptoms and describes a bite from a spider thatis clearly identified as a black widow. However, clinicians mayencounter patients who have the symptoms of latrodectism butwhose history of the bite or identification of the spider is notdefinitive. It is possible that some of our patients receivedmisdiagnoses and therefore would not benefit from antivenomtreatment. Because our study was randomized, the risk formisclassification should bias the results toward the null.However, with a small sample size, if there was unbalancedmisclassification in the diagnosis of latrodectism our resultscould be biased. Future studies may benefit from definitivediagnosis of envenomation by venom detection.

Another potential protocol violation is suggested by thedetection of F(ab)2 in a placebo-treated subject’s plasma sample.Although we have no evidence (other than the detection of

Table 2. Baseline envenomation symptoms.

Variable

No. (%)

Antivenom(n�13)

Placebo(n�11)

Total(n�24)

Pain migrated proximally frombite site to larger musclegroups

12 (92) 11 (100) 23 (96)

Muscle cramping or pain 12 (92) 11 (100) 23 (96)Chest, abdominal, back,

neck, or limb pains notattributable to othercauses

13 (100) 10 (91) 23 (96)

Pain/stinging followed bydiscomfort/diaphoresis atbite site

11 (85) 10 (91) 21 (88)

Fatigue 9 (69) 6 (55) 15 (63)Weakness 7 (54) 6 (55) 13 (54)Bite site with localized area

of mild erythema7 (54) 6 (55) 13 (54)

Mild hypertension 6 (46) 7 (64) 13 (54)Nausea/vomiting 6 (46) 6 (55) 12 (50)Generalized diaphoresis 6 (46) 2 (18) 8 (33)Increased respiration rate 3 (23) 5 (45) 8 (33)Conjunctival erythema 4 (31) 4 (36) 8 (33)Mild tachycardia 3 (23) 3 (27) 6 (25)Periorbital edema 3 (23) 0 3 (13)

Table 3. Efficacy assessments by treatment group.

Variable Statistic/Response Antivenom (n�13) Placebo (n�11) P Value

VAS pain score change from baseline to 150 min (using LOCF), mm Mean (SD) �55.0 (26.9) �34.6 (35.6) .14*Median �50.0 �46.0Range �98.0 to 1.0 �98.0 to 26.0

Treatment failure, No. (%) No 10 (77) 4 (36) .06†

Yes 3 (23) 7 (64)Time to 13 mm or greater decrease in VAS pain score, min Mean (SD) 46.2 (33.8) 76.4 (45.2) .03*

Median 30.0 90.0Range 30.0 to 150.0 30.0 to 150.0

LOCF, Last observation carried forward.*One-sided Mann-Whitney U test.†One-sided Fisher’s exact test.




F(ab)2) that this subject received antivenom, the possibility ofmisclassification bias is a threat to the validity our results.

Overall results are suggestive of a quicker time to pain relief.The ultimate utility of the medication must be measured againstcosts, risk, and long-term outcomes, which were beyond thescope of this phase 2 trial. In the current study, the safety profileis good, which aligns with the positive safety profile of thecommercialized product in other countries.

DISCUSSIONWe were not able to demonstrate that antivenom produced

greater overall pain reduction compared with placebo at 150minutes or fewer treatment failures than placebo; we did findthat antivenom reduces pain more quickly than placebo inpatients with moderate to severe latrodectism. The majority ofthe treatment effect for antivenom-treated patients wasdetectable at 30 minutes, whereas the placebo-treated patientshad more gradual decrease in pain that was statistically similarto the treatment group by the end of the study. For patientswho are experiecing severe latrodectism, rapid pain relief is aclinically relevant outcome. Additionally, treatment withantivenom had an adverse event profile similar to that ofplacebo, and no serious adverse events were found. We believe

that the results of this study warrant a larger clinical trial of thisantivenom to further evaluate efficacy outcomes.

Existing information about black widow spider bite islimited. Our patients appeared similar to those reported inretrospective studies. The largest retrospective study involved163 patients in Arizona.2 Most patients (61%) were men. Theaverage age of all patients was 31.6 years. The average time frombite until onset of symptoms was 1.1 hours (SD 1.6 hours). Aswith our patients, the most common presenting complaint wasgeneralized muscular abdominal, back, and leg pain, although avariety of complaints were recorded. Thirty-six patients (22%)were diaphoretic on presentation. Most adult victims presentedwith normal vital signs, but 45 (31%) were hypertensive.2

Each assessment technique we tested appeared useful, withdirectional trends preferring the antivenom group. However,important limitations of each technique were identified. Forexample, the change in visual analog scale score has been used todetermine the efficacy of analgesics in many ED studies.23,24 In ourcohort, 4 (36.4%) placebo control subjects demonstrated markedimprovement in visual analog scale scores by the end of thetreatment arm, including 1 subject who reported a 98-mmimprovement in visual analog scale score. This improvement in thecontrol group was unexpected. Potential explanations of this effectinclude a strong placebo effect, a strong response to cointervention(eg, fentanyl), spontaneous resolution of pain, or another cause ofthe pain (misdiagnosis of latrodectism), which then spontaneouslyremitted.

A second measure of analgesic efficacy was to compare theproportion of patients who did not reach a threshold of pain reliefby the end of the study period (treatment failures).25 This measureis favored by many studies because the threshold can be set at a levelthat is clinically meaningful to patients16 and because it also limitsthe effect of patients who have a strong placebo response. Inaddition to providing clinical benefit to the patient, a decreased rateof treatment failures could translate to more efficient use of EDresources through shorter patient stay and fewer return hospitalvisits because of rebound pain, as indicated by the more rapid andlasting treatment response witnessed in the antivenom group. The

Figure 2. Mean and 95% CI of visual analog scale painscores over time.

Figure 3. Visual analog scale pain scores over time.




proportion of treatment failures in the placebo group was 64%compared with 23% in the antivenom group. Because thisdifference approached statistical significance (P�.06) and theabsolute difference in the effect (41%) would be clinically relevant,the findings should be validated in a larger sample.

Time to clinically important pain relief was our final measure ofanalgesic efficacy. This measure is similar to the concept of averagepain, which was highly correlated with patient satisfaction in astudy of postoperative pain.26 In the current study, visual analogscale pain score decreased in both groups during the 150-minuteobservation period, with more rapid symptom relief in theantivenom group. The median time to clinically important decreasein pain in the antivenom group was 30 minutes compared with 90minutes in the placebo group. This result is consistent with otherfindings showing that antivenom can significantly shorten thecourse of severe black widow spider envenomation.2

In summary, although antivenom did not reduce the overallseverity of pain from Latrodectus envenomation, the use ofantivenom led to significantly faster pain relief than placebo anddid not produce more adverse events. A larger phase 3 clinicaltrial is needed to more definitively determine efficacy and safety

of this investigational antivenom for treatment of moderate tosevere pain associated with latrodectism.

Supervising editors: Lewis S. Nelson, MD; Michael L.Callaham, MD

Author affiliations: From the Denver Health and HospitalAuthority/Rocky Mountain Poison & Drug Center, Denver, CO(Dart, Bogdan, Heard, Bucher Bartelson, Spradley); theUniversity of Colorado School of Medicine Department ofEmergency Medicine, Aurora, CO (Dart and Heard); theInstituto Bioclon Sociedad Anonima de Capital Variable,Tlalpan, Mexico Distrito Federal, Mexico (Garcia-Ubbelohde);Loma Linda University, Loma Linda, CA (Bush); the Universityof Louisiana Health Sciences Center, Shreveport, LA (Arnold);the University of California–San Diego, San Diego, CA (Clark);the University of California–Fresno, Fresno, CA (Hendey); andthe University of Virginia, Charlottesville, VA (Holstege).

Dr. Nelson and Dr. Callaham were the supervising editors onthis article. Dr. Dart did not participate in the editorial reviewor decision to publish this article.

Table 4. Number and percent of subjects experiencing at least one treatment emergent adverse event by treatment group.

System Organ Class Preferred Term

No. (%)

Antivenom Incidence(n�13)

Placebo Incidence(n�11)

Total Incidence(n�24)

Gastrointestinal disorders Abdominal pain 2 (15) 2 (18) 4 (17)Abdominal pain upper 1 (8) 0 1 (4)Diarrhea 0 1 (9) 1 (4)Nausea 3 (23) 4 (36) 7 (29)Vomiting 0 3 (27) 3 (13)

General disorders and administration site conditions Chest pain 2 (15) 1 (9) 3 (13)Chills 3 (23) 1 (9) 4 (17)Infusion site pain 1 (8) 0 1 (4)Malaise 0 1 (9) 1 (4)Pain 2 (15) 0 2 (8)Pyrexia 2 (15) 2 (18) 4 (17)

Infections and infestations Urinary tract infection 1 (8) 0 1 (4)Musculoskeletal and connective tissue disorders Arthralgia 3 (23) 3 (27) 6 (25)

Flank pain 1 (8) 0 1 (4)Myalgia 1 (8) 0 1 (4)Pain in extremity 2 (15) 3 (27) 5 (21)

Nervous system disorders Dizziness 1 (8) 1 (9) 2 (8)Headache 3 (23) 1 (9) 4 (17)Hypoesthesia 2 (15) 0 2 (8)

Psychiatric disorders Anxiety 0 1 (9) 1 (4)Renal and urinary disorders Chromaturia 1 (8) 0 1 (4)

Renal pain 1 (8) 0 1 (4)Reproductive system and breast disorders Breast tenderness 1 (8) 0 1 (4)Respiratory, thoracic, and mediastinal disorders Cough 1 (8) 0 1 (4)Skin and subcutaneous tissue disorders Blister 0 1 (9) 1 (4)

Hyperhidrosis 0 2 (18) 2 (8)Pruritus 6 (46) 3 (27) 9 (38)Pruritus generalized 0 1 (9) 1 (4)Rash 3 (23) 2 (18) 5 (21)Rash pruritic 0 1 (9) 1 (4)

Vascular disorders Flushing 1 (8) 0 1 (4)Hot flush 1 (8) 0 1 (4)




Author contributions: RCD and GB conceived of the study, designedthe trial, obtained research funding, and supervised the conduct ofthe trial and data collection. BBB provided statistical analysis. RCD,KH, and BBB drafted the article, and all authors contributed to itsrevision. RCD takes responsibility for the paper as a whole.

Funding and support: By Annals policy, all authors are requiredto disclose any and all commercial, financial, and otherrelationships in any way related to the subject of this articleas per ICMJE conflict of interest guidelines (seewww.icmje.org). Financial support for conduct of the trial wasprovided to the coordinating clinical site (Rocky MountainPoison & Drug Center [RMPDC]) by Rare DiseaseTherapeutics, Nashville, TN; and Instituto Bioclon S.A. de C.V.,Talapalan, Mexico. Drs. Dart, Bogdan, Heard, and BucherBartelson, and Ms. Spradley are employed by RMPDC andreceived standard salary throughout the duration of theconduct of the trial. Dr. Garcia-Ubbelohde is employed byInstituto Bioclon S.A. de C.V. and received standard salarythroughout the duration of the conduct of the trial. Drs. Bush,Arnold, Clark, Hendey, and Holstege received funds fromRMPDC to cover administration and procedural costsassociated with the conduct of the trial at their respectiveinvestigative sites. Dr. Heard was also supported by awardK08DA020573 from the National Institute on Drug Abuse. NoRMPDC employee received direct compensation for their rolein this trial. No funding was influenced or modified by dataoutcomes. The content is solely the responsibility of theauthors and does not necessarily represent the official viewsof the National Institute on Drug Abuse or the NationalInstitutes of Health.

Publication dates: Received for publication November 4,2011. Revisions received April 30, 2012; August 14, 2012;and September 5, 2012. Accepted for publication October 4,2012. Available online February 4, 2013.

Address for correspondence: Richard C. Dart, MD, PhD, [email protected].

REFERENCES1. Bronstein AC, Spyker DA, Cantilena LRJ, et al. 2008 Annual

report of the American Association of Poison Control Centers’National Poison Data System (NPDS): 26th annual report. ClinToxicol (Phila). 2009;47:911-1084.

2. Clark RF, Wethern-Kestner S, Vance MV, et al. Clinicalpresentation and treatment of black widow spider envenomation:a review of 163 cases. Ann Emerg Med. 1992;21:782-787.

3. Maretic Z. Latrodectism: variations in clinical manifestationsprovoked by Latrodectus species of spiders. Toxicon. 1983;21:457-466.

4. Binder LS. Acute arthropod envenomation. Incidence, clinicalfeatures and management. Med Toxicol Adverse Drug Exp. 1989;4:163-173.

5. Murphy CM, Hong JJ, Beuhler MC. Anaphylaxis with Latrodectusantivenin resulting in cardiac arrest. J Med Toxicol. 2011;7:317-321.

6. Food and Drug Administration. Biologic product shortages [Website]. Available at: http://www.fda.gov/BiologicsBloodVaccines/SafetyAvailability/Shortages/default.htm. Accessed June 8, 2012.

7. American Society of Health-System Pharmacists. Black widowantivenin (Latrodectus mactans) [Web site]. Available at:https://www.ashp.org/DrugShortages/Current/bulletin.aspx?id�670. Accessed June 8, 2012.

8. Key GF. A comparison of calcium gluconate and methocarbamol(Robaxin) in the treatment of latrodectism (black widow spiderenvenomation). Am J Trop Med Hyg. 1981;30:273-277.

9. Bloom JW, Chernick DA, Davidson AB, et al. Reversal of centralbenzodiazepine effects by flumazenil after intravenous conscioussedation with diazapam and opioids: report of a double-blindmulticenter study. Clin Ther. 1992;14:910-923.

10. Daly FF, Hill RE, Bogdan GM, et al. Neutralization of Latrodectusmactans and L. hesperus venom by redback spider (L. hasseltii)antivenom. J Toxicol Clin Toxicol. 2001;39:119-123.

11. Suntorntham S, Roberts JR, Nilsen GJ. Dramatic clinical responseto the delayed administration of black widow spider antivenin.Ann Emerg Med. 1994;24:1198-1199.

12. Smith TW, Butler VPJ, Haber E, et al. Treatment of life-threateningdigitalis intoxication with digoxin-specific Fab antibody fragments:experience in 26 cases. N Engl J Med. 1982;307:1357-1362.

13. Clark RF, McKinney PE, Chase PB, et al. Immediate and delayedallergic reactions to Crotalidae polyvalent immune Fab (ovine)antivenom. Ann Emerg Med. 2002;39:671-676.

14. Dart RC, Seifert SA, Boyer LV, et al. A randomized multicentertrial of crotalinae polyvalent immune Fab (ovine) antivenom forthe treatment for crotaline snakebite in the United States. ArchIntern Med. 2001;161:2030-2036.

15. Ruha AM, Curry SC, Beuhler M, et al. Initial postmarketingexperience with Crotalidae polyvalent immune Fab for treatmentof rattlesnake envenomation. Ann Emerg Med. 2002;39:609-615.

16. Gallagher EJ, Bijur PE, Latimer C, et al. Reliability and validity of avisual analog scale for acute abdominal pain in the ED. Am JEmerg Med. 2002;20:287-290.

17. Gallagher EJ, Liebman M, Bijur PE. Prospective validation ofclinically significant changes in pain severity measured on avisual analog scale. Ann Emerg Med. 2001;38:633-638.

18. Todd KH, Funk KG, Funk JP, et al. Clinical significance of reportedchanges in pain severity. Ann Emerg Med. 1996;27:485-489.

19. Bijur PE, Silver W, Gallagher EJ. Reliability of the visual analogscale for measurement of acute pain. Acad Emerg Med. 2001;8:1153-1157.

20. Gonzales VA, Vazquez IM, Robles GD, et al. Measuring painintensity in patients with latrodectism: the visual analog scale(VAS). Clin Toxicol. 2005;43:707-708.

21. Maxwell C. Sensitivity and accuracy of the visual analogue scale:a psycho-physical classroom experiment. Br J Clin Pharmacol.1978;6:15-24.

22. Vazquez H, Chavez-Haro A, Garcia-Ubbelodhe W, et al.Pharmacokinetics of a F(ab’)2 scorpion antivenom in healthyhuman volunteers. Toxicon. 2005;46:797-805.

23. Garra G, Singer AJ, Leno R, et al. Heat or cold packs for neckand back strain: a randomized controlled trial of efficacy. AcadEmerg Med. 2010;17:484-489.

24. Kostic MA, Gutierrez FJ, Rieg TS, et al. A prospective, randomizedtrial of intravenous prochlorperazine versus subcutaneoussumatriptan in acute migraine therapy in the emergencydepartment. Ann Emerg Med. 2010;56:1-6.

25. Drendel AL, Gorelick MH, Weisman SJ, et al. A randomizedclinical trial of ibuprofen versus acetaminophen with codeine foracute pediatric arm fracture pain. Ann Emerg Med. 2009;54:553-560.

26. Jensen MP, Martin SA, Cheung R. The meaning of pain relief in aclinical trial. J Pain. 2005;6:400-406.



A Randomized, Double-Blind, Placebo-Controlled Trial of a Highly Purified Equine F(ab)2 Antibody...

Documents

Transcript of A Randomized, Double-Blind, Placebo-Controlled Trial of a Highly Purified Equine F(ab)2 Antibody...