ilable at ScienceDirect

Toxicon 55 (2010) 1195–1212

Contents lists ava

Toxicon

journal homepage: www.elsevier .com/locate/ toxicon

Review

Methodology of clinical studies dealing with the treatmentof envenomation

Jean-Philippe Chippaux a,*, Roberto P. Stock b, Achille Massougbodji c

a IRD UMR 216, ‘‘Mothers and children facing the tropical infections’’ Institut de Recherche pour le Developpement, Paris, Franceb Instituto de Biotecnologıa, Universidad Nacional Autonoma de Mexico, Av. Universidad 2001, Cuernavaca, Morelos 62210, Mexicoc Faculte des Sciences de la Sante de Cotonou, Universite d’Abomey-Calavi, Benin

a r t i c l e i n f o

Article history:Received 20 March 2009Received in revised form 5 February 2010Accepted 18 February 2010Available online 26 February 2010

Keywords:EnvenomationsTreatmentClinical studiesClinical trialMethodology

* Corresponding author. IRD, UMR 216, 08 B.P. 841+229 97 39 73 66.

E-mail address: jean-philippe.chippaux@ird.fr (J.

0041-0101/$ – see front matter � 2010 Elsevier Ltddoi:10.1016/j.toxicon.2010.02.022

a b s t r a c t

A total of 142 clinical studies have been devoted to the treatment of envenomations, ofwhich 115 address snake bites, 20 scorpion stings, and 8 other animals (one addresses bothsnake and spider envenomation). Antivenom use was studied in 118, of which 82addressed efficacy, 43 evaluated safety, 23 studied dosage and 8 explored other issues.Besides anecdotal clinical reports, three classes of clinical studies are distinguished: (a)observational clinical studies (55 of the total) which analyze series of cases, (b) compar-ative clinical studies (36) which compare therapeutic products or treatment regimenswithout a gold standard for comparison and (c) randomized clinical trials (RCT, 51). Thegoals, methods and constraints of design of RCT are determined by whether explanatory(analytical) or pragmatic considerations are prioritized. Explanation-oriented RCT rely onstrict group comparability before and during treatment, in order to ensure the internalvalidity of the study. The pragmatically-oriented RCT aims at establishing the superiority ofa treatment over another, the goal being to maximize the external validity of the trial (thatis, its application in current practice). We found that all clinical studies of treatment ofenvenomation lean markedly toward the explanatory end and suggest that, given someparticularities of envenomation as a medical condition, a more pragmatic approach may beof value, particularly under the conditions prevalent for clinical studies in developingnations.

� 2010 Elsevier Ltd. All rights reserved.

1. Introduction

Envenomations by dangerous species of scorpions andsnakes constitute a frequent medical emergency and animportant public health problem in numerous tropicalnations. In developing countries their management is largelyinadequate, resulting in high levels of morbidity andmortality. Furthermore, incidence and mortality are notprecisely known: scorpion stings are estimated in 1.2million, resulting in 3,250 yearly fatalities (Chippaux andGoyffon, 2008), and snake bites are thought to fall within

, Cotonou, Benin. Tel.:

-P. Chippaux).

. All rights reserved.

a range of 1.2–6 million a year, resulting in 20,000–132,000fatalities (Chippaux, 1998a ; Kasturiratne et al., 2008). Thewide intervals are a result of, on one hand, the scarcity ofepidemiological information in countries with elevatedprevalence and mortality and, on the other, the poor repre-sentativeness of the available information, often obtainedfrom very localized surveys. If more extensive, this infor-mation would certainly permit improvement in patient care(Chippaux, 2008).

Treatment of envenomations rests on symptomaticdrugs, blood transfusion and manipulations such asmechanical ventilation, renal dialysis, fasciotomy or surgicalinterventions (SDM) and antivenoms (AV). The use of theformer remains mostly empirical, while the principles of thelatter were discovered over a century ago (reviewed in

J.-P. Chippaux et al. / Toxicon 55 (2010) 1195–12121196

Chippaux and Goyffon, 1998). After several decades ofsuccess, antivenoms underwent a period of rejection bynumerous prescribers, mainly owing to controversy ontheir efficacy and their frequent, and often severe, adverseeffects. Up to the sixties, the efficacy and safety (or tolera-bility) of antivenoms were established by animal experi-mentation (Grasset, 1957; WHO, 1981) and clinical practiceon the basis of isolated cases or variably extensive series ofpatients according to the methods accepted at the time.

The development of clinical trials in the last fifty yearshas allowed for progress in therapeutic protocols of enve-nomation without, however, eliminating controversy byway of contested and contradictory results (Chippaux,2006; Fisher, 1999). To illustrate this point, it suffices tomention the debate surrounding antivenom therapy ofscorpion stings (Abroug et al., 1999; Bawaskar andBawaskar, 2007; Boyer et al., 2009), or the use of heparinin the treatment of venom-induced coagulopathies irre-spectively of the species involved (see for instance, Myint-Lwin et al., 1989; Paul et al., 2003, 2007; Tin Nu Swe et al.,1992; Warrell et al., 1976). One consequence has been theappearance of often imprecise, and even unsound, recom-mendations; another has been a notable insufficiency ofuseful products, such as AV in Africa (Chippaux, 1998b,2002, 2008; Gutierrez et al., 2006; Stock et al., 2007;Theakston et al., 2003; Theakston and Warrell, 2000) andother underdeveloped regions, whereas doubtful – andeven dangerous – treatments remain a widespreadrecourse (Warrell, 2008). In general, drugs and vaccines areclinically tested in four successive phases (Table 1). Beforethey are marketed, it is necessary to establish their safetyand describe their pharmacokinetics (Phase I), as well astheir efficacy and tolerability (Phases II and III). After theyare commercialized, further clinical studies allow a confir-mation of their efficacy and safety profiles as well as thestudy other properties (Phase IV or post-marketing phase).

The World Health Organization recommends clinicalvalidation of marketed AV after rigorous preclinical control(WHO, in press). However, the methodology of clinicalstudies to be used for AV assessment – and eventually thoseSDM of use in the treatment of envenomation – is notestablished and should be the result of a consensus. Theobjective of this article is to compile and analyze theprocedures of envenomation-related clinical studies.Neither the results of the trials nor the use to which theproducts studied were put to, whether AV, SDM or prophy-lactic treatments, are evaluated, but rather the validity oftheir conclusions as a function of the methods employed.Discussion of the results of the studies would requireconsideration of a number of factors, including the circum-stances and conditions of each trial, which exceeds the scope

Table 1Goals and characteristics of the four phases of clinical assessment during drug d

Main goal

Phase I Safety – pharmacology – dose findingPhase II Efficacy – dose finding – tolerancePhase III Overall validation of previous trials

RegistrationPhase IV Post-marketing studies – improvement

of this review. The properties, constraints and demands ofthe different methods used to assess the treatment ofenvenomation, as well as the criteria necessary for validationof results will be discussed. Recommendations aimed at animprovement of the practice of clinical studies of the treat-ment of envenomations and the presentation of their resultswill also be made. These may be of use to researcherspreparing a study, or organizing their results for publication,as well as to reviewers or editors of such articles.

2. Material and methods

2.1. Investigation of the published record

The search for clinical studies of envenomation wasdone by a systematic interrogation of PubMed usingthe NCBI interface (http://www.ncbi.nlm.nih.gov/sites/entrez?db¼PubMed) using different combinations of thewords «envenom*» and «treatment» or «antiven*». Thissearch was supplemented by exploration of the Cochranedatabase. The last searches were made on December 27,2009. The bibliographical search was further pursued byfollowing the references of selected articles to find publi-cations in non-indexed journals, theses and proceedings, aswell as available clinical reports. Only articles in English,French, Spanish, Portuguese and German, expresselymentioning a goal of therapeutic evaluation in the title orabstract, were retained for this study. We discarded thosearticles not making explicit reference to validation orassessment of a treatment of envenomation by terrestrialanimals. We also excluded studies based on small numbersof subjects (less than 15 patients, with the exception ofprospective studies where the protocol required reducednumbers of patients for reasons logistical or ethical innature) or where the design was not precise or pertinent.We did not consider anecdotal studies, by far the mostabundant, although they may be very informative(MacMahon and Collins, 2001; Vandenbroucke, 2004).

2.2. Analysis of clinical studies

A clinical study can be descriptive or comparative inorder to analyze or demonstrate a pathological mechanismor to validate a new treatment. Therefore, it may addressproducts (AV or SDM), therapeutical procedures (mode ofadministration, indications, posology, delays, surveillance,drug associations, etc.) or prophylactic measures. We keptthe denominations used by the authors of the clinicalstudies to divide them into four groups: (a) anecdotalClinical Studies (ACS) describing the treatment of one ora few envenomation cases without a standardized

evelopment.

# Subjects involved Clinical study type

w10 Healthy subjects ERCT60–100 Patients ERCT100–500 Patients ERCT, PRCT, CCS, OCS

100–1000 Patients All types

Table 2Geographic distribution of clinical studies on the treatment ofenvenomation.

Region Type of study

OCS CCS RCT Total

Europe 2 (4%) 3 (8%) 0 5 (4%)North Africa,

Middle East5 (9%) 7 (19%) 2 (4%) 14 (10%)

Sub-Saharan Africa 8a (14%) 3 (8%) 5 (10%) 16a (11%)North America 12 (21%) 2 (6%) 3 (6%) 17 (12%)Latin America 6 (11%) 7 (19%) 13 (25%) 26 (18%)Asia 13a (23%) 13 (37%) 24 (47%) 50a (35%)Australasia 10 (18%) 1 (3%) 4 (8%) 15 (10%)

a One study regards both Sub-Saharan Africa and Asia (Malasit et al.,1986).

J.-P. Chippaux et al. / Toxicon 55 (2010) 1195–1212 1197

protocol; (b) observational Clinical Studies (OCS) thatmethodically analyze the therapeutic effects or the safety(tolerability) of a treatment in a series of patients; (c) non-randomized Comparative Clinical Studies (CCS), includingcase-control studies, where the groups to compare aredetermined by circumstance (or ad hoc, that is, selection ofpatients is based on availability of treatments, route ofadministration, delay in treatment, symptomatology orother non-random causes); and (d) Randomized ClinicalTrials (RCT) where the protocol is determined by a veryprecise methodological standard (Fisher, 1999; Harrington,2000; Schwartz and Lellouch, 1967; Thorpe et al., 2009).Studies were assessed in terms of geographic region,zoological group, the stated goals of the study, products orprocedures used, the arms of the study and the numbers ofpatients involved, modality of selection of subjects andcriteria for their assignment to the study arms, decisioncriteria and statistical methods (randomization procedures,confidence interval and statistical power). Assessment ofthe validity of randomization was based on the [often brief]description of the methodology provided by the authors;a cursory mention to randomization was not sufficient toconsider that the study was correctly randomized. Whenthe article did not mention the statistical power, we madea retrospective estimation using the formula of Casagrandeet al. (1978) for proportions and that of Machin et al. (1997)for the means. For comparison between more than twogroups, we used One-way Anova Power Analysis describedby Cohen (1988) by way of a free program (Faul et al., 2007).

Table 3Distribution of clinical studies of envenomation treatment in terms of animals a

OCS CCS

Animal Snake 46a (82%) 26 (72Scorpion 7 (13%) 10 (28Spider 3a (5%) 0Ant 0 0Caterpillar 0 0

Goals of the studyb Efficacy 35 (64%) 29 (81Tolerance 26 (47%) 6 (17Dosage 6 (11%) 8 (22Procedure 2 (4%) 3 (8%Other 0 1 (3%

Total 55 36

a One study explores the treatment of envenomation by both snakes and spidb Studies may have several goals.

In addition, we took into account that methodologicalnorms have evolved since the beginning of immuno-therapy. Ethical approval and informed consent have beenimplemented gradually, particularly in developing coun-tries, since the appearance of Ethics Committees in 1990(Chippaux, 2006). We have considered these issues, then,only for those studies appearing after 1990. The principle ofanalyses in «intention to treat» or «per protocol» wasdefined in 1976 (Peto et al., 1976) We have thereforeconsidered this aspect only for studies since 1979. Finally,standardized presentation of results according to theCONSORT Statement was described in 1996 (Altman, 1996;Begg et al., 1996) so we have considered it a criterion forthose studies carried out after 1997.

3. Results

The database search allowed identification of 3534articles corresponding to the criteria. After selection ofpertinent articles and identification of non-indexed sources,we kept 142 clinical studies identifiable as OCS, CCS or RCTthat, either in the title or abstract, expressely made refer-ence to a goal of assessment of a specific or symptomatictreatment of envenomation by snakes (115), scorpions (20),spiders (5), ants (2) or caterpillars (1); one article reporteda simultaneous study of snake and spider envenomations.We excluded studies on treatment of envenomations by seaanimals, reviewed by Atkinson et al. (2006).

Of the total of 142 reports, 55 were OCS (39%), 36 wereCCS (25%) and 51 (36%) were RCT. An ancillary pharmaco-logical study (Isbister et al., 2008c) derived from an RCTdescribed elsewhere (Isbister et al., 2008b) was not includedin this analysis. Table 2 summarizes the geographic distri-bution of the different clinical studies. North Africa and theNear and Middle East were grouped together by virtue oftheir characteristic endemicity of scorpion envenomation(Chippaux and Goyffon, 2008). The distribution of thesestudies in terms of the animals responsible for theenvenomations and their goals are presented in Table 3.

Of 55 OCS, 26 were retrospective studies (47%) and 29were prospective (53%). Their main characteristics are lis-ted in Tables 3 and 4.

Of 36 CCS, 1 (3%) was a case-control study, 22 (61%) wereretrospective studies and 13 (36%) were prospective

nd goals of the study.

RCT (AVþ SDM) RCT (AV) Total

%) 43 (84%) 28 (82%) 115a (80%)%) 3 (6%) 2 (6%) 20 (14%)

2 (4%) 2 (6%) 5a (4%)2 (4%) 1 (3%) 2 (1%)1 (2%) 1 (3%) 1 (1%)

%) 41 (80%) 24 (71%) 104 (73%)%) 11 (22%) 11 (32%) 43 (30%)%) 9 (18%) 9 (26%) 23 (16%)) 1 (2%) 1 (3%) 6 (4%)) 1 (2%) 1 (3%) 2 (1%)

51 34 142

ers.

J.-P. Chippaux et al. / Toxicon 55 (2010) 1195–12121198

studies, out of which 4 made use of either historical orgeographical controls (Tables 3 and 5). In total, 15 studiesmade one or several historical comparisons (42%) andanother one used a geographical control. Thirty one (82%)evaluated AV and 7 (18%) SDM, two studies evaluating bothAV and SDM. Statistical power was sufficient (1� b� 80%)in 10 (53%) of 19 analyzable studies, insufficient(25%> 1� b< 80%) in 5 (26%), and very insufficient(1� b� 25%) in 4 others (21%).

Table 4Main characteristics of Observational Clinical Studies (OCS).

Method Product Animal Goal Ethic Consent

Prospect AV Snake Dosage Yes YesRetrosp AV Snake Efficacy No NARetrosp AV Snake Tolþ dose No NARetrosp AV Scorpion Efficacy No NARetrosp AV Snake Tolþ effic NA NARetrosp SDM Snake Efficacy No NAProspect AV Snake Dosage No NoProspect AV Snake Efficacy Yes YesProspect AV Snake Tolþ effic No NoRetrosp AV Snake Tolerance Yes NARetrosp AV Snake Tolerance No NAProspect AV Snake Tolþ effic Yes YesProspect AV Snake Tolþ effic Yes YesProspect AV Snake Tolþ effic Yes YesProspect AV Snake Tolþ effic Yes YesProspect AV Snake Tolerance Yes YesProspect AV Snake Efficacy NA NAProspect AV Scorpion Efficacy Yes YesRetrosp AV Scorpion Tolþ effic No NARetrosp SDM Scorpion Efficacy Yes NARetrosp AV Scorpion Efficacy No NAProspect AV Snake Efficacy Yes YesProspect AV Snake Tolþ effic Yes YesProspect AV Spider Efficacy Yes YesRetrosp AV Spider Tolerance Yes NARetrosp AV Snake Efficþ dose Yes NAProspect AV Snake Tolerance Yes NAProspect AV Snake Tolerance Yes NoProspect AV Snake Efficacy No NoProspect AV Snake Efficacy No NoRetrosp AV Snake Efficacy Yes NARetrosp AV Snake Tolþ effic Yes NARetrosp AV Snake Tolerance Yes NAProspect AV Scorpion Tolerance Yes NAProspect AV Snake Tolerance Yes YesProspect AV Snake Efficacy NA NARetrosp AV Snake Efficacy NA NARetrosp AV Snake Tolerance No NARetrosp AV Snake Efficacy NA NARetrosp AV Snake Tolþ effic Yes NAProspect AV Snake Tolþ effic Yes NoProspect AV Snake Efficacy NA NARetrosp AV Snake Tolþ effic Yes NAProspect AV Scorpion Efficacy No YesRetrosp AV Snake Efficacy No NAProspect AV Snake Tolerance No NoRetrosp AV Snake Efficacy NA NAProspect AV Snake/Spider Tolerance NA NoRetrosp AV Snake Procedures NA NAProspect AV Snake Efficacy No NoRetrosp AV Snake Tolerance No NAProspect AV Snake Dosage No NoProspect AV Snake Efficacy NA NARetrosp AV Snake Procedures Yes NARetrosp AV Snake Dosage NA NA

AV, antivenom; SDM, symptomatic drug or manipulation.

Of 51 RCT, 32 (63%) were open trials, that is, not blinded(Tables 3, 6 and 7); for the remaining 19 (37%), administra-tion of treatment was double blind in 17 and single blind in 2.Three RCT made use of crossing-over, that is, treatment foreach subject was reversed during the protocol (Brown et al.,2003; Hile et al., 2006; Watt et al., 1986). This procedureallows each subject to be considered its own control, whichconsiderably reduces the number of necessary patients,minimizes confounding factors and increases the robustness

# Patients Region Reference

35 Asia Ariaratnam et al., 199910 ME/N. Afr. Bentur et al., 199754 ME/N. Afr. Bentur et al., 2004951 ME/N. Afr. Bouaziz et al., 200849 Australasia Brian and Vince, 198724 Lat. Am. Bucaretchi et al., 199440 Lat. Am. Bucher et al., 199715 N. Am. Bush et al., 2002310 Lat. Am. Caiaffa et al., 199493 N. Am. Cannon et al., 2008130 Asia Chen et al., 2000223 SubS. Afr. Chippaux et al., 199846 SubS. Afr. Chippaux et al., 1999289 SubS. Afr. Chippaux et al., 200711 N. Am. Dart et al., 199751 Asia Dassanayake et al., 200240 SubS. Afr. Daudu and Theakston, 1988275 ME/N. Afr. El Hafny and Ghalim, 2002151 Lat. Am. Gateau et al., 199433 N. Am. Gibly et al., 199940 ME/N. Afr. Hammoudi-Triki et al., 200430 Asia Hanvivatvong et al., 199746 Europe Haro et al., 199868 Australasia Isbister and Gray, 200395 Australasia Isbister, 200727 Australasia Isbister et al., 2007195 Australasia Isbister et al., 2008a36 Asia Kularatne et al., 2003166 Australasia Lalloo et al., 199532 Australasia Lalloo et al., 199632 N. Am. Lavonas et al., 200450 N. Am. LoVecchio and DeBus, 2001181 N. Am. LoVecchio et al., 2003116 N. Am. LoVecchio et al., 199935 Africa/Asia Malasit et al., 198628 Lat. Am. Mille et al., 197047 Asia Mitrakul et al., 198417 SubS. Afr. Moran et al., 1998123 Asia Myint-Lwin et al., 198565 N. Am. Offerman et al., 200112 N. Am. Offerman et al., 200223 Asia Phillips et al., 198824 N. Am. Pizon et al., 200718 Lat. Am. Rezende et al., 199528 N. Am. Ruha et al., 2002184 Asia Seneviratne et al., 2000113 Europa Stahel et al., 1985181 Australasia Sutherland and Lovering, 197923 SubS. Afr. Swinson, 19765 Asia Theakston et al., 1990254 Asia Thiansookon and Rojnuckarin, 200842 Asia Vijeth et al., 200010 SubS. Afr. Warrell et al., 1975136 Australasia Williams et al., 200735 Australasia Yeung et al., 2004

Table 5Main characteristics of envenomation Comparative Clinical Studies (CCS).

Method Product Control Animal Purpose Ethics Consent Consort # Groups # Sujets Distributionof criteria

1� b Region Reference

Prospect AV Nil Snake Dosage Yes Yes No 3 45 Symptoms NA ME/N. Afr. Açikalin et al., 2008Retrosp AV AV Snake Dosage No NA No 2 55 Sequential NA Asia Agarwal et al., 2005Prospect AV AV Snake Effþ Pharm No Yes NA 3 37 Delay NA Lat. Am. Amaral et al., 1997Retrosp AV AV Snake Efficþ Proc No NA NA 2 97 Procedure 6 Lat. Am. Amaral et al., 1998Retrosp SDM Nil Scorpion Efficacy No NA NA 2 35 Ad hoc 85 Asia Bawaskar and Bawaskar, 1992Retrosp SDM Nil Scorpion Efficacy No NA No 2 32 Historical 87 Asia Bawaskar and Bawaskar, 2000Prospect AV SDM Scorpion Efficacy No No No 2 53 Geographical 99 Asia Bawaskar and Bawaskar, 2007Prospect AV Nil Scorpion Efficacy Yes NA No 2 270 C/C 21 ME/N. Afr. Belghith et al., 1999Retrosp SDM Nil Scorpion Efficacy Yes Yes No 2 384 Historical 80 Asia Biswal et al., 2006Retrosp AV Nil Snake Efficacy No No No 3 288 Procedure NA SubS. Afr. Bregant et al., 2006Prospect SDM Nil Snake Tolerance Yes Yes No 2 129 Historical NA Lat. Am. Caron et al., 2009Retrosp AV AV/Nil Snake Tolþ effic No NA No 3 149 Historical NA N. Am. Corneille et al., 2006Prospect AV AV Snake Tolþ effic Yes No No 2 1619 Historical NA Asia Ghosh et al., 2008Prospect AV AV/Nil Scorpion Dosage No Yes No 3 275 Doses 75 ME/N. Afr. Ghalim et al., 2000Retrosp AV Nil Snake Efficacy Yes Yes No 2 119 Ad hoc 77 Europe Harry et al., 1999Retrosp AVþ SDM SDMþNil Snake Efficacy NA NA NA 2 83 Ad hoc 99 N. Am. Huang et al., 1978Prospect AV AV/Nil Snake Efficþ dos No NA No 3 27 Symptoms NA Asia Hung et al., 2003Retrosp AV AV Snake Efficacy No Yes No 2 31 Historical 25 Asia Hung et al., 2006Retrosp AV AV Snake Dosage No NA NA 2 730 Historical NA Lat. Am. Jorge and Ribeiro, 1994Retrosp AV Nil Snake Tolþ effic No NA NA 2 46 Historical 99 Europe Karlson-Stiber and Persson, 1994Retrosp AV Nil Snake Tolþ effic Yes NA NA 2 60 Historical 5 Europe Karlson-Stiber et al., 1997Prospect AV AV Snake Dosage No No NA 2 43 Symptoms NA Asia Kimoto et al., 1997Retrosp AV Nil Scorpion Efficþ proc No NA No 6 147 Procedures NA ME/N. Afr. Krifi et al., 1999Retrosp SDM Nil Scorpion Efficacy No NA No 5 38 Symptoms NA Asia Krishnan et al., 2007Retrosp AV Nil Snake Efficacy NA NA NA 2 18 Ad hoc NA ME/N. Afr. Mann, 1978Prospect AV AV Snake Efficacy No No No 4 50 Delay NA Asia Narvencar, 2006Prospect AVþ SDM AVþNil Scorpion Efficacy No Yes No 4 48 SymptþHist. NA Asia Natu et al., 2006Prospect AV AV/Nil Snake Dosage No No No 5 85 Symptoms. NA ME/N. Afr. Shemesh et al., 1998Retrosp AV Nil Scorpion Efficacy No NA NA 2 104 Historical 40 ME/N. Afr. Sofer et al., 1994Retrosp AV AV Snake Tolþ effic No NA No 2 79 Historical NA Lat. Am. Sotelo, 2008Retrosp AV SDM Snake Efficacy Yes NA NA 2 50 Sequential 68 Lat. Am. Thomas et al., 1995Retrosp AV Nil Snake Tolþ effic Yes NA No 2 132 Sequential 93 Lat. Am. Thomas et al., 1998Prospect AV AV Snake Efficacy No NA NA 2 156 Delay 99 Australasia Trevett et al., 1995aRetrosp AV AV Snake Efficacy NA NA No 2 344 Historical 90 SubS. Afr. Visser et al., 2008Retrosp AV Nil Snake Efficþ dos No NA No 2 310 Historical 91 SubS. Afr. Visser et al., 2004Prospect AV AV Snake Efficþ proc No No NA 2 118 Delay 54 Asia Win-Aung et al., 1996

AV, antivenom; SDM, symptomatic drug or manipulation; Nil, no control treatment; C/C, case control study.

J.-P.Chippauxet

al./Toxicon

55(2010)

1195–12121199

Table 6Main characteristics of Randomized Clinical Trial (RCT) regarding AV.

Method Control Animal Purpose Ethics Consent Randomprecision

CONSORTpresentation

# Groups # Patients # Criteria 1� b Analysis Region Reference

Open P Scorpion Efficacy Yes Yes Yes Yes 2 825 2 25 NP ME/N. Afr. Abroug et al., 1999Open AV Snake Tolþ dos Yes Yes No No 3 24 2 NP NP SubS. Afr Abubakar et al., 2010Open AV Snake Tolþ effic Yes Yes Yes No 2 43 3 13 NP Asia Ariaratnam et al., 2001Blind P Scorpion Tolþ effic Yes Yes Yes Yes 2 15 3 80 ITT N. Am. Boyer et al., 2009Blind P Ant Efficacy Yes Yes Yes Yes 2 52 1 99 CrossOþ PP Australasia Brown et al., 2003Open AV Caterpillar Dosage Yes Yes No Yes 2 44 1 53 NP Lat. Am. Caovilla and Guardao

Barros, 2004Open AV Snake Efficacy No Yes No NA 3 121 Multiple 18 NP Lat. Am. Cardoso et al., 1993Open AV Snake Efficacy Yes Yes Yes No 2 31 3 5 NP N. Am. Dart et al., 2001Blind AV Spider Efficacy Yes Yes Yes Yes 2 31 1 92 Dummyþ ITT Australasia Ellis et al., 2005Open AV Snake Pharmacol Yes Yes No NA 3 26 1 5 NP Asia Ho et al., 1990Open P Spider Efficþ proc Yes Yes Yes Yes 2 126 1 61 Dummyþ ITT Australasia Isbister et al., 2008bBlind AV Snake Dosage No No No NA 2 170 3 16 NP Lat. Am. Jorge et al., 1995Open AV Snake Tolþ effþ dos No No No NA 2 30 1 NP NP Asia Karnchanachetanee et al., 1994Open AV Snake Tolerance Yes Yes No NA 2 66 1 12 NP Asia Malasit et al., 1986Open AV Snake Efficacy No Yes Yes NA 2 39 3 5 NP SubS. Afr. Meyer et al., 1997Open AV Snake Tolþ effic No Yes No NA 2 39 3 12 NP Lat. Am. Otero et al., 1996Open AV Snake Tolþ effic Yes Yes No No 3 79 2 5 NP Lat. Am. Otero-Patino et al., 1998Open AV Snake Tolþ effic No Yes No No 2 53 3 5 NP Lat. Am. Otero et al., 1999Open AV Snake Tolþ effic Yes Yes No No 2 67 4 12 NP Lat. Am. Otero et al., 2006Open AV Snake Efficacy No Yes No No 2 74 4 99 NP Lat. Am. Pardal et al., 2004Open AV Snake Dosage Yes Yes No No 2 100 2 23 NP Asia Paul et al., 2004Open AVþNila Snake Dosage No No No NA 4 (H) 68 1 NA NP Asia Pochanugool et al., 1997Blind P/SDMa Snake Efficacy NA NA No NA 3 100 4 NP NP Asia Reid et al., 1963Blind P Snake Efficacy Yes Yes No No 2 28 2 81 ITT Asia Rojnuckarin et al., 2006Blind P Snake Tolþ effic No Yes No NA 2 63 2 29 NP Asia Sellahewa et al., 1995Blind AV Snake Tolþ effic Yes Yes Yes No 3 210 2 58 NP Lat. Am. Smalligan et al., 2004Blind AV Snake Dosage Yes Yes Yes No 9 (H) 90 Multiple NA NP Asia Srimannarayana et al., 2004Open AV Snake Dosage Yes Yes No No 2 60 4 99 NP Asia Tariang et al., 1999Open AV Snake Efficacy No No No NA 6 118 1 NA NP Lat. Am. Theakston et al., 1992Open AV Snake Dosage NA NA No NA 2 53 6 NP NP Asia Thomas and Jacob, 1985Open AV Snake Efficacy NA NA No NA 2 46 Multiple 66 NA SubS. Afr. Warrell et al., 1974Open AV Snake Efficacy NA NA No NA 2 14 7 49 NP SubS. Afr. Warrell et al., 1980Open AV Snake Efficacy Yes Yes No NA 3 46 Multiple 15 NP Asia Warrell et al., 1986Open SDM Snake Efficacy NA Yes No NA 3 8 1 NA NP Asia Watt et al., 1989

CONSORT presentation, complying with CONSORT presentation statement for all items but two or less; AV, antivenom; SDM, symptomatic drug or manipulation; P, placebo; NA, not applicable; NP, not performed;H, historical control.

a Studies with two controls; CrossO, cross over; ITT, intention to treat; PP, per protocol.

J.-P.Chippauxet

al./Toxicon

55(2010)

1195–1212120

0

Tab

le7

Mai

nch

arac

teri

stic

sof

the

RC

Tre

gard

ing

SD.

Met

hod

Test

edp

rod

uct

Con

trol

An

imal

Goa

lEt

hic

sC

onse

nt

Ran

dom

CO

NSO

RT

pre

sen

tati

on#

Gro

up

s#

Pati

ents

#C

rite

ria

1�

bA

nal

ysis

Reg

ion

Ref

eren

ce

Bli

nd

SDM

PSc

orp

ion

Effi

cacy

Yes

Yes

No

NA

26

00

35

NP

ME/

N.A

fr.

Ab

rou

get

al.,

19

97

Bli

nd

SDM

PSn

ake

Effi

cacy

No

Yes

Yes

No

21

01

195

NP

Lat.

Am

.Fa

net

al.,

19

99

Bli

nd

SDM

PSn

ake

Effi

cacy

Yes

Yes

Yes

No

35

21

39N

PA

sia

Gaw

aram

man

aet

al.,

20

04

Bli

nd

SDM

PA

nt

Effi

cacy

Yes

Yes

No

No

1a

24

180

Cro

ssO

N.A

m.

Hil

eet

al.,

200

6B

lin

dA

Vþ

SDM

AVþ

PSn

ake

Effi

cacy

No

No

No

No

22

51

211

NP

Lat.

Am

.Jo

rge

etal

.,2

00

4O

pen

SDM

Nil

Snak

eEf

fica

cyN

oN

oN

oN

A2

11

41

15N

PLa

t.A

m.

Ker

riga

net

al.,

19

97

Bli

nd

SDM

PSn

ake

Effi

cacy

Yes

Yes

No

No

21

44

216

NP

Asi

aK

ula

ratn

eet

al.,

20

05

Op

enA

Vþ

SDM

AVþ

Nil

Snak

eEf

fica

cyN

AN

AN

oN

A2

28

3N

DN

PA

sia

Myi

nt-

Lwin

etal

.,1

98

9B

lin

dSD

MP

Snak

eEf

fica

cyY

esY

esY

esN

o2

43

141

NP

Asi

aN

uch

pra

yoon

etal

.,2

00

8O

pen

SDM

Nil

Snak

eEf

fica

cyN

oY

esN

oN

o2

12

27

23N

PA

sia

Pau

let

al.,

20

03

Op

enA

Vþ

SDM

AVþ

Nil

Snak

eEf

fica

cyN

oY

esN

oN

o2

80

727

NP

Asi

aPa

ul

etal

.,2

00

7B

lin

dSD

MP

Snak

eEf

fica

cyY

esY

esY

esY

es2

10

51

97N

PA

sia

Prem

awar

dh

ena

etal

.,1

99

9O

pen

AVþ

SDM

AVþ

Nil

Snak

eEf

fica

cyN

oN

oN

oN

A2

15

299

NP

Asi

aSe

llah

ewa

etal

.,1

99

4O

pen

AVþ

SDM

AVþ

Nil

Snak

eEf

fica

cyN

oN

oN

oN

A2

20

7N

DN

PA

sia

Tin

Nu

Swe

etal

.,1

99

2B

lin

dA

Vþ

SDM

AVþ

PSn

ake

Effi

cacy

Yes

Yes

No

NA

31

62

NA

NP

Au

stra

lasi

aTr

evet

tet

al.,

19

95

bO

pen

AVþ

SDM

AVþ

Nil

Snak

eEf

fica

cyN

AN

AN

oN

A2

14

670

NA

Sub

S.A

fr.

War

rell

etal

.,1

97

6B

lin

dSD

MP

Snak

eEf

fica

cyN

AY

esN

oN

A2

10

454

Cro

ssO

.A

sia

Wat

tet

al.,

19

86

CO

NSO

RT

pre

sen

tati

on,c

omp

lyin

gw

ith

CO

NSO

RT

pre

sen

tati

onst

atem

ent

for

alli

tem

sb

ut

2or

less

;A

V,a

nti

ven

om;

SDM

,sym

pto

mat

icd

rug

orm

anip

ula

tion

;P,

pla

ceb

o;N

A,n

otap

pli

cab

le;

NP,

not

per

form

ed.

aTe

sted

and

con

trol

pro

du

ctad

min

iste

red

sim

ult

aneo

usl

yto

each

ind

ivid

ual

;C

ross

O,c

ross

over

.

J.-P. Chippaux et al. / Toxicon 55 (2010) 1195–1212 1201

of the conclusions. It is, however, intrinsically limited tostable persisent symptoms. Two other RCT made use ofa placebo under the form of a double dummy (Ellis et al.,2005; Isbister et al., 2008b). This method consists in asso-ciating the administration of the experimental drug to theadministration of the corresponding placebo to all subjects:each patient receives simultaneously the drug under oneform and the placebo under another so as to mask, at thesame time, the product and its mode of administration.

The randomization procedure was detailed in only 14RCT (27%). A placebo was used in 17 studies (33%).Comparison to a reference treatment was made in 28 RCT(55%) although previous validation of the referenceproducts – which would have justified their use as such –was not mentioned in any of them. When the goal wasa comparison of two therapeutic procedures or two dosagesthe same product was used. In 8 RCT (16%) the control groupreceived neither treatment nor placebo. Thirty four RCT(67%) were done to validate an AV. The characteristics ofthese RCT are comparable to those of other studies of thetype, except for blinding and control procedures (Table 8).

In 37 RCT comparisons were made between 2 groups ofpatients (72%); in 10 others, 3 groups were compared (20%)and in 3 RCT, 4 groups or more were compared (8%)although in one case using 2 by 2 groups. In another RCTeach subject was its own control at the moment of inter-vention and the comparison was made within the samegroup (Hile et al., 2006). There was a single criterion ofevaluation in 16 RCT (31%), two in 11 RCT (22%), three in 9RCT (18%) and 4 or more in 15 RCT (29%).

All RCT indicated the limit of a-risk (or confidenceinterval, 95% in all of them). In contrast, only 11 RCT (22%)mentioned the calculation of statistical power (b). Thenumber of recruited subjects in terms of the confidenceinterval (a¼ 5%) and statistical power (1�b� 80%) wassufficient in 10 RCT of 40 (25%), insufficient in 11 (28%), andvery insufficient (1�b� 25%) in 19 others (47%). Whetherthe analysis was in intention to treat or per protocol wasspecified in only 5 RCT of 49 done after 1980 (10%).

The advice of the Ethics Committee appears in 26 RCT(60%) of 43 done after 1990, and informed consent ofpatients was requested in 37 studies (82%) of 45 thatrequired it. Of 26 RCT carried out after 1999, only 3 complyfully to the applicable requirements, and 3 others presenttrivial omissions. All the others show several importantdefects: insufficient number of subjects, non-comparabilityof procedures between groups, not blinded, unspecified orinappropriate randomization, inadequate statistical anal-ysis, too many decision criteria, controls not comparable toexperimental groups, non-standardized and ambiguousresults. It would be vain to search for venial defects: in anRCT, only rigorous implementation of a sound methodologyauthorizes conclusions (Table 9).

4. Discussion

Although the invention of AV dates back to 1894 and thefirst applications on human beings to 1896, clinical studiesevaluating efficacy or safety of AV go back no further than 50years. The oldest RCT we could find studied an AV incombination or not with prednisolone in the treatment of

Table 8Characteristics of RCT performed for AV comparison and validation.

Method AV RCT (%) SDM RCT (%) Total RCT (%)

Blind 9 (26)a 10 (59)a 19 (37)Placebo 7 (21)a 10 (59)a 17 (33)No control treatment (¼nil) 1 (3)a 7 (41)a 8 (16)Advice from ethical committee 19/29 (66) 7/14 (50) 26/43 (60)Informed consent 26/30 (87) 11/15 (73) 37/45 (82)Procedure of randomization 10 (29) 4 (24) 14 (27)CONSORT presentation 6/18 (33) 1/9 (11) 7/27 (26)# RCT comparing two groups 23 (68) 14 (82) 37 (73)# RCT with one decision criteria 10 (29) 6 (35) 16 (31)# RCT with statistical power 1� b� 80% 6/26 (23) 4/14 (29) 10/40 (25)# RCT with statistical power 1� b� 25% 14/26 (54) 5/14 (36) 19/40 (48)# RCT with ITT or PP analysis 5/33 (15) 0/16 5/49 (10)

Total RCT 34 17 51

a Significant difference between AV and SDM RCT (P< 0.05).

J.-P. Chippaux et al. / Toxicon 55 (2010) 1195–12121202

envenomation by Ancistrodon (¼Calloselasma) rhodostomain Malaysia (Reid et al., 1963). Although database mining ofpublications might have missed some studies we think,without pretending exhaustiveness, that the vast majority,

Table 9Checklist of CONSORT Statement for reporting RCT on treatment of envenomationumber of complying articles out of the 25 studies identified after 1997.

# items Heading andsubheading

Descriptor

1 Title Identify the study as ‘‘comparative2 Abstract Use a structured and formal forma

eventually background, discussion3 Introduction:

backgroundScientific background and rational

4 Introduction:objectives

Hypothesis, general (no more than

5 Methods: datacollection

Settings and location of the study;

6 Methods: intervention Detail the interventions intended7 Methods: control

treatmentValidity of control treatment: placproperties.

8 Methods: participants Criteria of eligibility.9 Methods: decision Description and justification of de10 Methods: sample size How sample size was determined

interim analysis.11 Methods:

randomizationMethods used to generate the randstratification); methods use for imthe allocation sequence, enrolled t

12 Methods: blinding How blinding was obtained (similsuccess of blinding was evaluated.

13 Methods: statistics Statistical tests used for comparisoanalysis if any).

14 Methods: outcomes Defining primary and secondary e15 Ethics Ethical committee approval and in16 Results: participant

flowUsing a diagram (Fig. 3) that detaistudy (from eligibility to completi

17 Results: recruitment Date and circumstances of recruitm18 Results: baseline data Comparison of demographic and c19 Results: follow-up Description of the progress of the

in each group for analysis (e.g. ITT o(and not percentages).

20 Results: outcomes Summary of results for each expecestimation of effect of size with co

21 Results: ancillaryanalysis

Subgroup and adjusted analysis, inAnalysis of data not included in thestudy).

22 Discussion:interpretation

Interpretation of the results accord

23 Discussion: validity Generalizability of the results to sp24 Conclusion General interpretation of the resul

a Well completed or addressed.

and without doubt the most important, were retrieved andexamined.

Phase I, II and III clinical trials are recognized as stages inthe measurement of efficacy, safety and dosage of products

ns (after Begg et al. 1996; Altman et al., 2001; Hopewell et al., 2008) and

# ComplyingRCT studies

’’ or ‘‘randomized’’ trial. 19/26t (at least goal, methods, results, conclusion,).

15/26

e; in PRCT, expected characteristics of the product. 26/26

1) and specific objectives (as few as possible). 23/26

who collected the data? 21/26

for each group. 25/26ebo or reference treatment with well known 9/19

25/26cision criteria and stopping rules. 11aþ 10/25(statistical power); justification and explanation of 10/25

om allocation sequence and details of restriction (e.g.plementation of allocation sequence; who generateshe participants and assigned them to their groups?

13/25

arity of control treatment) and applied. How the 11/25

ns and additional analysis (subgroup and adjusted 25/25

xpected outcomes. 11aþ 11/26formed consent 19/26ls the flow of participants through each stage of theon).

9/26

ent and follow-up. 21/26linical data before the intervention. 25/26intervention. Precision of the number of participantsr PP and why such a choice) giving absolute numbers

26/26

ted primary and secondary outcome. If necessary,nfidence interval.

26/26

dicating those prespecified and those exploratory.main outcomes (and not planned in the design of the

26/26

ing to hypothesis, bias and confusing factors. 26/26

ecific and/or general populations. 25/26ts in the context of current evidence. 26/26

J.-P. Chippaux et al. / Toxicon 55 (2010) 1195–1212 1203

before they are marketed. Phase I trials are carried out inhealthy volunteers, payed to participate in the study. Thegoal is to establish the safety of a new drug or, sometimes,its pharmacokinetics. This phase is, in theory, indispensablefor any new drug but it is often considered unnecessarilydangerous in the case of antivenoms due to the unaccept-ably high risk of adverse effects (Gutierrez et al., 2010). Assafety is an essential requirement for registration of anantivenom which will be used in isolated environmentswith limited therapeutic resources, Phase I should becarried out in order to establish safety in appropriateconditions (e.g. laboratory and clinical surveillance),eventually in envenomated patients. This ongoing contro-versy does not, however, fall within the scope of thisreview. Phases II and III concern the evaluation of efficacy,safety and other factors (such as dosage, drug interactions,etc.) in patients. Criteria of inclusion depend on the goals ofthe trial and the methodology to be employed. Phase IVtrials, the protocols of which are equally rigorous butfollowing somewhat different standards, are carried outafter commercialization in order to ascertain utility inroutine practice, refine details of administration or improveuse (Table 1). They may have wider objectives and a lessrestrictive design aimed at an evaluation of the beneficialcontributions of a drug in current practice, notably a moreprecise measure of tolerability (Chippaux, 2006). Theproperties and methodological constraints of the differenttypes of clinical studies are summarized in Table 10.

The methodology of clinical studies, with a particularbenefit for clinical trials, has evolved during the period

Table 10Comparison of the characteristics of the various types of clinical studies.

Items Type of study

Observational study (OCS) Non-randomizedcomparative stu

Objective Descriptive – analyticalRetrospective/prospective

Descriptive – anProspective/retr

Type of comparison Not applicable Case/control – had hoc

Hypothesis Not applicable Not applicable o

Protocol Definition of criteria;Definition of proceduresif exploratory study

External validity

Experimental group Not applicable Variable represe

Control group Not applicable Free

Group comparabilitybefore treatment

Not applicable Exceptionally

Treatment procedures Design not decided by theinvestigator

Procedures indethe investigator

Group comparabilityduring treatment

Not applicable Exceptionally

Decision criteria Multiple: depends on thetype of investigation

Multiple: functioinvestigation

Statistical analysis Not applicable Intention to trea

Statistical risks Not applicable a¼ 5%; possibly

CONSORT presentation Not applicable Preferable

a In the case of equivalence RCT.

covered by the review (1963–2009), most notably in thestatistical and ethical domains. Improvements in statisticalmethods have allowed consideration of variables judgedquite intractable until recently. For example, averagelethality by a venomous species within a limited regionvaries little in time. If conditions remain constant, itbecomes possible to evaluate the impact of the applicationof a particular treatment during one period by comparinglethality before and after its use. Twelve CCS resort tohistorical comparison and one to a geographical control. Itis important to stress that these comparisons must beregarded with caution because comparability is conditionaland sometimes debatable. However, the use of existentdata or results from routine procedures generally releasesCCS from many ethical and logistical constraints, makingthem attractive under some circumstances.

Observational Clinical Studies (OCS) are the mostnumerous (Table 4). Restrospective OCS are simple andinexpensive. They contribute essential clinical and thera-peutic information, all the more so if the number ofpatients is high, but are limited by non-standardized datacollection (Concato and Horwitz, 2004). Nonetheless, OCSallow preparation of more appropriate and informative RCTto answer specific therapeutic questions (Isbister, 2006).Prospective OCS are preferred for a more precise analysisbased on particular criteria clearly determined andincluded in the protocol. Methodological constraints areminimal without preventing OCS from validating certainresults, particularly in the assessment of safety. Half of theOCS address the frequency and severity of undesirable

dies (CCS)Randomized clinicaltrial (ERCT)

Pragmatic clinicaltrial (PRCT)

alyticalospective

Explanatory – causality Decision-making

istory –

r variable H0¼H1 (superiority) orH0 s H1

a (equivalence)H0 s H1

Internal validity External validity

ntativity Selected Representative ofcurrent practice

Characteristics identicalto experimental group

Representative ofcurrent practice

Randomization Randomization

pendent of Identical procedures forall groups

Procedures optimizedfor each group

Blind Blind not required

n of the Single – simple Multiple – intrincate

t Intention to treat(superiority)Per protocol(equivalence)

Intention to treat

b¼ 20% a¼ 5%; b¼ 20% g (risk to reverse theresult)

Essential Preferable

Fig. 1. Difference of activity between the experimental treatment andplacebo (equivalent to absence of treatment) or reference treatment. Thegreater the difference, the lower the necessary number of patients in eachgroup.

J.-P. Chippaux et al. / Toxicon 55 (2010) 1195–12121204

effects, while these issues concern less than 20% of theother types of studies. A retrospective, and better yetprospective, OCS can measure the incidence of side effectsfollowing a treatment administered routinely.

The least common type of trial is the ComparativeClinical Study, probably because its conclusions suffer frominevitably ambiguous protocols. The goal of the CCS isgenerally descriptive and can be occasionally analytical.However, the design of CCS does not allow validation ofa treatment, which needs explanatory studies in theterminology of evidence-based medicine. Their maindefect, except in case-control studies, is the lack ofa comparable and representative control. However, CCS areof interest in that they can contribute sometimes solidarguments at minimal cost since some criteria may besufficiently robust to permit some historical or geograph-ical comparisons using adequate statistical methods,notably correct confidence intervals and power. Criteria canbe considered robust if they result from a widespreadconsensus or if the comparison makes use of appropriatestatistical methods and takes into account the importantexpected variance of the samples analyzed.

Randomized clinical trials (RCT) constitute the goldstandard in clinical studies (Fisher, 1999; Harrington, 2000).Their objective is to test a therapeutic hypothesis and permita causal imputation by way of a strict methodology.However, and contrary to widespread opinion, randomiza-tion – even if correctly implemented (Altman and Schulz,2001; Altman et al., 2001; Berger and Bears, 2003) – doesnot by itself define a RCT. Other equally important condi-tions must be met to ensure the validity of the result.

In their now classical article, Schwartz and Lellouch(1967) made the distinction between explanatory (oranalytical) attitudes in clinical studies – corresponding to‘‘classical’’ randomized clinical trials (ERCT) aiming at theacquisition and validation of knowledge (i.e. test a causalresearch hypothesis)– and pragmatic randomized clinicaltrials (PRCT) which have the goal of guiding decision-making. The adjective ‘‘pragmatic’’ should be understoodin terms of the philosophical concept of pragmatism whichasserts that ‘‘the importance of an idea or action lies inwhether it makes a difference in everyday life’’ (Maclure,2009). The protocols of PRCT, while equally rigorous, areusually applied only during Phases III and IV studies. Thecharacteristics defining each approach correspond to twoopposing but complementary visions of medical science:explanation versus evaluation, of medical interventions(Vandenbroucke, 2008) or, in the words of Maclure (2009)‘‘a pragmatic trial is a real-world test in a real-world pop-ulation whereas an explanatory trial is a specializedexperiment in a specialized population.’’

The original explanatory versus pragmatic formulationhas been further refined with the recognition that there isa continuum between pragmatic and explanatory trialswith no clear-cut border separating them, only differencesin emphasis which tilt the study preferentially one way orthe other. Recently, Thorpe et al. (2009) defined a toolcalled PRECIS (PRagmatic-Explanatory Continuum Indi-cator Summary) with 10 key domains to help researchersdetermine the best design for their trial, and a representa-tion under the form of a wheel graph to summarize the

different states of indicators giving a general view ofthe trial’s orientation (Fig. 4). For reasons of simplicity inthe following discussion of RCT we will, however, mostlyrefer to the conceptually ‘‘pure’’ categories of ERCT andPRCT of Schwartz and Lellouch while keeping in mind themore nuanced current understanding of the concepts.None of the studies uncovered in our search could beclassified as approaching a PRCT as originally defined. Wehave found that RCT of the treatment of envenomation haveleaned mostly toward the explanatory type in theirconceptual basis, the stated goals and in the methodologiesemployed. We think that a comparison of the explanatoryand pragmatic approaches to RCT as they apply to thetreatment of envenomation, particularly in the assessmentof AV in developing nations, is on order.

The efficacy of an AV is equal to the difference (D)between the number of patients cured by it and a placebo(Fig. 1). It follows that it becomes possible to measure theefficacy of another AV by comparing it to the first accordingto the formula D¼ d1þ d2 (d1 and d2 being the differencebetween the placebo and the first AV, and that between thefirst AV, now the reference, and the second, now theexperimental AV, respectively). A superiority ERCT aims atproving that a treatment is better than another, whethera placebo or a reference treatment. The statistical hypoth-esis is the similarity between the two treatments and, inpractical terms, the objective is to obtain a significant resultto disprove the hypothesis and deduce a differencebetween treatments. However, contrary to what most ofstudies report, a non-significant result does not mean thatthe two treatments are equivalent. To conclude this,a sufficient number of subjects must be involved (Altmanand Bland, 1995) and, eventually, an equivalence ERCT –or the conceptually similar non-inferiority ERCT – must beperformed. In these, a non-significant comparison confirmsthe similarity between the two treatments within thearbitrary limits defined by the researcher (Fig. 2). Equiva-lence ERCT are useful when comparing two treatmentswith effects that are estimated to be very close and,

J.-P. Chippaux et al. / Toxicon 55 (2010) 1195–1212 1205

consequently, neither a placebo will be used since its effectis considered nil and therefore very far, a priori, from theexperimental treatment, nor a reference treatmentconsidered to be very insufficient. Equivalence ERCTrequire a highly constrained methodology and, generally,a larger number of subjects (D’Agostino et al., 2003; Joneset al., 1996). The validity of the results will depend on thestrict comparability of the groups before treatment (whichis determined inclusion criteria and randomization), duringtreatment (which depends on blinding and identicalprocedures for both groups excepting the treatment itself)and after treatment, which is guaranteed by a rigorousstatistical analysis.

The inevitable constraint of patient heterogeneity incases of envenomation, and therefore their response to thetreatments being compared, may be reduced to manage-able proportions by two strategies: the selection of patientsusing restrictive criteria to minimize differences betweenthem (cause of envenomation, symptomatology, delaybetween accident and treatment, age and body volume,pathological background, etc.), and randomization toequitably distribute residual differences among the groupsto be compared. Patient homogeneity and group compa-rability in envenomation cases have been extensively dis-cussed by Isbister (2005), who underscored the inadequacyof the definition of a case in several RCT. Randomizationensures group comparability before treatment only if it iscorrectly conceived and applied, as all patients must havean identical chance of getting one or another treatment(Altman and Bland, 1999). It is therefore recommended thatrandomization be carried out by third parties, not theinvestigators, and that it be masked (Altman and Schulz,2001). Reporting of results must include an explicitmention to the randomization method and how it wasapplied to the patients (Altman et al., 2001).

Comparability of procedures during treatment is alsoessential because it guarantees control over the study andthe validity of the comparison. Blinding is surely thedeterminant factor but all other procedures applied topatients of both groups must be identical and rigorouslyrespected for the duration of the study. Blinding is rela-tively easy to perform when an experimental treatment iscompared to a placebo which perfectly resembles itsaspect, or when the reference treatment is identical in itspresentation and mode of administration. It is more diffi-cult, however, to organize when there are differences

Fig. 2. Limits of decision (�D) in equivalence RCT including confidenceintervals of experimental results (after Jones et al., 1996).

between the experimental and reference treatments interms of formulation (e.g. liquid versus lyophilized AV) oradministration (differing dosages, intramuscular versusintravenous injection, etc.) Most (63%) of the RCT of enve-nomation were open label design. ‘‘Dummies’’ (tablets orsolutions) are a recourse consisting in the simultaneousadministration of both formulations to the patients, onecontaining the experimental or reference treatment andthe other a placebo. This ensures identical procedures forall patients, irrespective of their group (Ellis et al., 2005;Isbister et al., 2008b, c). Comparability also requires thatancillary procedures be identical; in several RCT, symp-tomatic treatments were administered to some patients inresponse to the evolution of envenomation: anti-inflammatories, anti-histaminics, hydrocortisone or bloodtransfusion. Such treatments have a variable impact on thestate of the patients which can alter in a non-trivial mannerthe comparison between groups. Therefore, essentialsymptomatic treatments must be either administered to allpatients in all groups to be compared (which may presentethical and logistical problems) or explicitly treated aspossible confounders in the statistical analysis. Ethical,scientific, logistical or financial considerations must guidethe choice of solutions best adapted to the interest of thepatients and the context of the study.

The use of a placebo is often considered unethical,particularly when the condition is severe – that is, poten-tially lethal or impairing – and there is a treatmentconsidered adequate (Streiner, 2007). However, somemethodological and ethical arguments may lend support toa comparison to a placebo in cases of envenomation(D’Agostino et al., 2003). First, if a placebo is excluded, forwhatever reason, the existence of a reference treatment ofknown efficacy becomes indispensable, particularly inequivalence RCT (Jones et al., 1996). Second, the differentialbetween two levels of efficacy bears directly on the powerof the comparison and therefore on the number of subjectsrequired to arrive at a statistically valid result: the greaterthe differential, the lower the number of patients needed(Fig. 1). By convention, placebo means absence of activityand this favors large differentials between groups, allowinga rapid and statistically valid assessment of an experi-mental product with reduced numbers of patients. Incontrast, comparison to a reference product with anunknown level of efficacy does not: at most a hierarchy canbe proposed without being able to precise a level of efficacybecause comparison to a reference AV with imprecisequalities will only lead, in the best case, to demonstrateimprecise properties. This is why it is not recommendableto use a reference AV for which efficacy has not been reli-ably measured. Envenomations constitute acute patholo-gies for which an immediate therapeutical response issought, and this can be exploited in the experimentalprotocol. For ethical reasons and to limit the risk ofaggravation in the placebo group, the time of observationmay be short (for example, 2–6 h, which is acceptable inmoderate envenomations) such that an alternative treat-ment can be administered.

A statistical treatment that will not conflict with respectof the procedures used during the trial is indispensable.Three elements are essential: (a) the statistical power that

J.-P. Chippaux et al. / Toxicon 55 (2010) 1195–12121206

determines the number of subjects per group (Schulz andGrimes, 2005), (b) the decision criterion, which must besingle due to the impact that it may have on the analysisand statistical power of the comparison (Schwartz andLellouch, 1967) and (c) the modality of analysis of thedata. We will examine each in turn. Insufficient statisticalpower may lead to a conclusion of unacceptably limitedvalidity, e.g. a power of 1� b� 25% means that outcomescould be erroneous in 75% of cases. When the authors didnot mention statistical power we calculated it according tothe number of recruited subjects, and found that power isinadequate in most RCT. These post-hoc calculations do notallow interpretation of the results (Goodman and Berlin,1994) but may reveal generalizations to be unjustified ifbased on results obtained with insufficient numbers ofpatients. The accumulation of decision criteria may poseconceptual, ethical and logistical problems (Schwartz andLellouch, 1967). First, the number of subjects necessaryfor each criterion is, by definition, different because theexpected differential will not be equal, leading to recruit-ment problems with simultaneous logistical and ethicalconsequences: it is not acceptable to expose one group toa risk of insufficient or useless treatment or severe sideeffects if it can be avoided. Second, results may be contra-dictory and preclude valid conclusions. An importantnumber of RCT (69%) made use of two or more criteria.Finally, analysis of the data should proceed in «intention totreat» (ITT), that is, using data from all patients included, inRCT aiming at demonstrating the superiority of one treat-ment (Fisher, 1999), and in «per protocol» (PP) – consid-ering only those patients for which the protocol has beenrigorously followed – in equivalence or non-inferiority RCT(Jones et al., 1996). Concordance of results between ITT and

Fig. 3. Model of diagram used to illustrate the flow of patients during CCS, ERCT

PP analyses further strengthens the conclusion. Only 5(10%) studies done after 1980 specified the modality of thestatistical analysis.

Presentation of results must summarize the ensemble ofinformation. Since 1996, the Consort Statement (Altman,1996; Begg et al., 1996) and its revisions (Altman et al.,2001; Hopewell et al., 2008) recommend a standardizedand usable presentation of the conclusions of an ERCT ora PRCT (Table 9 and Fig. 3).

The main strength of a correctly conceived and appliedERCT is its internal validity which ensures that its conclu-sions are applicable with high confidence. The mainlimitation is that this confidence is high only under thecircumstances in which the ERCT was conducted; if inclu-sion criteria are more restrictive (e.g. exclusion ofvery severely envenomated patients, or patients withconcomitant conditions) the comparison will be morerobust but the extrapolation to other types of cases moredifficult. If the goal is, however, to establish treatmentdecisions in routine practice, the external validity of a trialmust be privileged, maximizing the generalizability of itsresults to the greatest possible number of cases. It isnecessary then to do away with certain rules of procedure,particularly statistical, and replace them with thorough butmore flexible methodological strategies. The pragmaticapproach to RCT is adapted to this end (Schwartz andLellouch, 1967; Armitage, 1998; Macpherson, 2004;Maclure, 2009; Thorpe et al., 2009). In this case, thecentral issue is no longer to establish and accuratelymeasure a statistical difference in an outcome of twotreatments, but rather to accumulate arguments in favor ofeach and formulate a preference in the context of day-to-dayuse. It is assumed, a priori, that each treatment will present

or PRCT (after Begg et al., 1996; Altman et al., 2001; Hopewell et al., 2008).

Fig. 4. The blank pragmatic-explanatory continuum indicators summary(PRECIS) wheel (from Thorpe et al., 2009). The broader the surface, the morepragmatic trial design should be.

J.-P. Chippaux et al. / Toxicon 55 (2010) 1195–1212 1207

advantages and disadvantages and that these will beassessed to determine which is of greater interest to thepatients. However, and despite an obvious theoretical andpractical interest, PRCT remain very much under-utilized ingeneral (Armitage,1998) and in envenomation in particular.

Since, ideally, extrapolation of results to the totality ofpatients is the goal of the PRCT, an effort will be made torecruit subjects representative of current practice irre-spective of the differences between them, which is alwayseasier than selecting patients in terms of rigidly pre-defined criteria. Randomization will permit a comparisonat later stages by assuring comparability before treatment.It is, therefore, as important PRCT as in ERCT. In contrast,comparability during and after treatment is of limitedvalue (Table 10). The choice between a placebo andreference AV is not necessary in PRCT, and even the effi-cacy of a reference AV is of marginal interest. Blindingprocedures which are necessary to maintain comparabilitybetween groups receiving treatments have no raison d’etrein PRCT since the goal is not to ensure the internal validityof the result but rather to choose the best of two treat-ments when each is administered under optimal condi-tions. Generally, a PRCT aims at maximizing the benefit ofeach treatment and, therefore, procedures will be neces-sarily and deliberately different, a fact which excludesblinding by definition and authorizes certain liberties ofuse. For PRCT, analysis must be done in intention to treatsince it is desirable to approach the conditions of normalpractice: it is necessary then to include all patients in theanalysis regardless of adhesion to treatment and respect ofthe protocol.

Finally for a PRCT, the number of patients to be includedis calculated according to a formula different from thatused for the ERCT (Schwartz and Lellouch, 1967).

As Armitage (1998) emphasizes, the interest of prag-matically oriented RCT is simultaneously operational, sinceit answers a practical question and ethical because besidesthe individual benefit of receiving an optimized treatment,there is a collective advantage in reaching a mandatorydecision on the superiority of one of the two treatments.

5. Conclusion

The choice of the type of clinical study depends on thegoals, conditions and available resources: patient number,severity of envenomations, alternative treatments, medicalinfrastructure and financing. The utility of OCS in theassessment of the safety of an AV, or the expected benefit ofa SDM is recognized. CCS are equally informative butcomparisons made with this type of clinical study areconditional. Therefore, it is necessary to use randomizedclinical trials to effectively compare two or more treatments.

Randomized clinical trials aim at ascertaining if productA is better than or equivalent to product B, and not atdetermining the overall or intrinsic therapeutic value of oneor both products. The validity of RCT tilted toward theexplanatory type is founded on the strict comparability ofgroups before and during treatment to ensure internalconsistency, i.e. the results are categorical but only underthe same conditions. Constraints that must be rigorouslyrespected are: (a) a comparison to a control group, eitherplacebo or validated reference product, (b) blind proce-dures, (c) strict inclusion criteria, (d) randomization, (e)identical procedures for all groups, (e) number of patientsbased on statistical rigor (a-risk: confidence interval; b-risk:statistical power), and (f) standardization of analysis andpresentation of results.

A more pragmatic approach to RCT represents a veryvaluable and, to this day, mostly unexplored alternative inour field. PRCT aim at establishing the superiority of a givenproduct over another, the explicit goal being to ensure theexternal validity of the trial (i.e. generalizability to as manytypes of patients as possible) and applicability underconditions of current practice (Maclure, 2009). They arefounded on the comparability of groups before, but notduring, treatment. Consequently, patient randomizationmust be performed but there is no requirement for controlgroups or blind procedures, and calculation of the number ofrequired patients follows somewhat different statisticalassumptions.

Randomized clinical trials aiming at comparing orimproving the treatment of envenomations are few andgenerally incomplete or poorly exposed, and are mostlyconceived as explanatory. As a consequence, the results ofmany of these trials may be considered as debatable oreven speculative. The burden of responsibility falls as muchon the journals and their reviewers as on the authorsthemselves when all the relevant information does notfigure explicitly in the articles (Altman, 2002).

Acknowledgements

We thank Dr. Jacques Gardon for advice during thewriting of the manuscript and help with the calculations ofstatistical power.

Conflict of interest

JPC was principal investigator in clinical studies partlysupported by Pasteur Merieux Serums & Vaccins (nowSanofi Pasteur). JPC, AM and RPS were investigators in

J.-P. Chippaux et al. / Toxicon 55 (2010) 1195–12121208

clinical studies partly supported by the Instituto Bioclon(a Silanes company).

References

Abroug, F., Elatrous, S., Nouira, S., Haguiga, H., Touzi, N., Bouchoucha, S.,1999. Serotherapy in scorpion envenomation: a randomisedcontrolled trial. Lancet 354, 906–909.

Abroug, F., Nouira, S., Haguiga, H., Elatrous, S., Belghith, M., Boujdaria, R.,Touzi, N., Bouchoucha, S., 1997. High-dose hydrocortisone hemi-succinate in scorpion envenomation. Ann. Emerg. Med. 30, 23–27.

Abubakar, S.B., Abubakar, I.S., Habib, A.G., Nasidi, A., Durfa, N., Yusuf, P.O.,Larnyang, S., Garnvwa, J., Sokomba, E., Salako, L., Laing, G.D.,Theakston, R.D., Juszczak, E., Alder, N., Warrell, D.A.Nigeria–UKEchiTab Study Group, 2010. Pre-clinical and preliminary dose-findingand safety studies to identify candidate antivenoms for treatment ofenvenoming by saw-scaled or carpet vipers (Echis ocellatus) innorthern Nigeria. Toxicon 55, 719–723.

Açikalin, A., Gokel, Y., Kuvandik, G., Duru, M., Koseoglu, Z., Satar, S., 2008.The efficacy of low-dose antivenom therapy on morbidity andmortality in snakebite cases. Am. J. Emerg. Med 26, 402–407.

Agarwal, R., Aggarwal, A.N., Gupta, D., Behera, D., Jindal, S.K., 2005. Lowdose of snake antivenom is as effective as high dose in patients withsevere neurotoxic snake envenoming. Emerg. Med J. 22, 397–399.

Altman, D.G., 1996. Better reporting of randomised controlled trials: theCONSORT statement. Br. Med. J. 313, 570–571.

Altman, D.G., 2002. Poor-quality medical research. What can journals do?JAMA 287, 2765–2767.

Altman, D.G., Bland, J.M., 1995. Absence of evidence is not evidence ofabsence? Br. Med. J. 311, 485.

Altman, D.G., Bland, J.M., 1999. Treatment allocation in controlled trials:why randomize? Br. Med. J. 318, 1209.

Altman, D.G., Schulz, K.F., 2001. Statiscal notes. Concealling treatmentallocation in randomized trials. Br. Med. J. 323, 446–447.

Altman, D.G., Schulz, K.F., Moher, D., Egger, M., Davidoff, F., Elbourne, D.,Gøtzsche, P.C., Lang, T.CONSORT GROUP (Consolidated Standards ofReporting Trials), 2001. The revised CONSORT statement for reportingrandomized trials: explanation and elaboration. Ann. Intern. Med.134, 663–694.

Amaral, C.F., Campolina, D., Dias, M.B., Bueno, C.M., Chavez-Olortegui, C.,Penaforte, C.L., Diniz, C.R., Rezende, N.A., 1997. Time factor in thedetection of circulating whole venom and crotoxin and efficacy ofantivenom therapy in patients envenomed by Crotalus durissus. Tox-icon 35, 699–704.

Amaral, C.F., Campolina, D., Dias, M.B., Bueno, C.M., Rezende, N.A., 1998.Tourniquet ineffectiveness to reduce the severity of envenoming afterCrotalus durissus snake bite in Belo Horizonte, Minas Gerais, Brazil.Toxicon 36, 805–808.

Ariaratnam, C.A., Meyer, W.P., Perera, G., Eddleston, M., Kuleratne, S.A.,Attapattu, W., Sheriff, R., Richards, A.M., Theakston, R.D., Warrell, D.A.,1999. A new monospecific ovine Fab fragment antivenom for treat-ment of envenoming by the Sri Lankan Russell’s viper (Daboia russeliirusselii): a preliminary dose-finding and pharmacokinetic study. Am.J. Trop. Med. Hyg. 61, 259–265.

Ariaratnam, C.A., Sjostrom, L., Raziek, Z., Kularatne, S.A., Arachchi, R.W.,Sheriff, M.H., Theakston, R.D., Warrell, D.A., 2001. An open, random-ized comparative trial of two antivenoms for the treatment of enve-noming by Sri Lankan Russell’s viper (Daboia russelii russelii). Trans. R.Soc. Trop. Med. Hyg. 95, 74–80.

Armitage, P., 1998. Attitudes in clinical trials. Stat. Med. 17, 2675–2683.Atkinson, P.R., Boyle, A., Hartin, D., McAuley, D., 2006. Is hot water

immersion an effective treatment for marine envenomation? Emerg.Med. J. 23, 503–508.

Bawaskar, H.S., Bawaskar, P.H., 1992. Role of atropine in management ofcardiovascular manifestations of scorpion envenoming in humans. J.Trop. Med. Hyg. 95, 30–35.

Bawaskar, H.S., Bawaskar, P.H., 2000. Prazosin therapy and scorpionenvenomation. J. Assoc. Physicians India 48, 1175–1180.

Bawaskar, H.S., Bawaskar, P.H., 2007. Utility of scorpion antivenin vs prazosinin the management of severe Mesobuthus tamulus (Indian red scorpion)envenoming at rural setting. J. Assoc. Physicians India 55, 14–21.

Begg, C., Cho, M., Eastwood, S., Horton, R., Moher, D., Olkin, I., Pitkin, R.,Rennie, D., Schulz, K.F., Simel, D., Stroup, D.F., 1996. Improving thequality of reporting of randomized controlled trials. The CONSORTstatement. JAMA 276, 637–639.

Belghith, M., Boussarsar, M., Haguiga, H., Besbes, L., Elatrous, S., Touzi, N.,Boujdaria, R., Bchir, A., Nouira, S., Bouchoucha, S., Abroug, F., 1999.

Efficacy of serotherapy in scorpion sting: a matched-pair study. J.Toxicol. Clin. Toxicol. 37, 51–57.

Bentur, Y., Raikhlin-Eisenkraft, B., Galperin, M., 2004. Evaluation of anti-venom therapy in Vipera palaestinae bites. Toxicon 44, 53–57.

Bentur, Y., Zveibel, F., Adler, M., Raikhlin, B., 1997. Delayed administrationof Vipera xanthina palaestinae antivenin. J. Toxicol. Clin. Toxicol. 35,257–261.

Berger, V.W., Bears, J.D., 2003. When can a clinical trial be called‘randomized’? Vaccine 21, 468–472.

Biswal, N., Bashir, R.A., Murmu, U.C., Mathai, B., Balachander, J.,Srinivasan, S., 2006. Outcome of scorpion sting envenomation aftera protocol guided therapy. Indian J. Pediatr. 73, 577–582.

Bregant, E.R., Cantoni, F., Tantardini, F., 2006. Snake bites in South Chad.Comparison between three different polyvalent anti-snake venomimmunotherapies. Giorn. Ital. Med. Trop. 11, 25–28.

Brian, M.J., Vince, J.D., 1987. Treatment and outcome of venomous snakebite in children Port Moresby General Hospital, Papua New Guinea.Trans. R. Soc. Trop. Med. Hyg. 81, 850–852.

Brown, S.G., Wiese, M.D., Blackman, K.E., Heddle, R.J., 2003. Ant venomimmunotherapy: a double-blind, placebo-controlled, crossover trial.Lancet 361, 1001–1006.

Bouaziz, M., Bahloul, M., Kallel, H., Samet, M., Ksibi, H., Dammak, H.,Ahmed, M.N., Chtara, K., Chelly, H., Hamida, C.B., Rekik, N., 2008.Epidemiological, clinical characteristics and outcome of severe scor-pion envenomation in South Tunisia: multivariate analysis of 951cases. Toxicon 52, 918–926.

Arizona Envenomation Investigators Boyer, L.V., Theodorou, A.A., Berg, R.A., Mallie, J., Chavez-Mendez, A., Garcıa-Ubbelohde, W., Hardiman, S.,Alagon, A., 2009. Antivenom for critically ill children with neurotox-icity from scorpion stings. N. Engl. J. Med. 360, 2090–2098.

Bucaretchi, F., Douglas, J.L., Fonseca, M.R., Zambrone, F.A., Vieira, R.J., 1994.Envenenamento ofıdico em crianças: frequencia de reaçoes precocesao antiveneno em pacientes que receberam pre-tratamento comantagonistas H1 e H2 da histamina e hidrocortisona. Rev. Inst. Med.Trop. Sao Paulo 36, 451–457.

Bucher, B., Canonge, D., Thomas, L., Tyburn, B., Robbe-Vincent, A.,Choumet, V., Bon, C., Ketterle, J., Lang, J.Research Group on Snake Bitesin Martinique, 1997. Clinical indicators of envenoming and serumlevels of venom antigens in patients bitten by Bothrops lanceolatus inMartinique. Trans. R. Soc. Trop. Med. Hyg. 91, 186–190.

Bush, S.P., Green, S.M., Moynihan, J.A., Hayes, W.K., Cardwell, M.D., 2002.Crotalidae polyvalent immune Fab (ovine) antivenom is efficaciousfor envenomations by Southern Pacific rattlesnakes (Crotalus helleri).Ann. Emerg. Med. 40, 619–624.

Caiaffa, W.T., Vlahov, D., Antunes, C.M., de Oliveira, H.R., Diniz, C.R., 1994.Snake bite and antivenom complications in Belo Horizonte, Brazil.Trans. R. Soc. Trop. Med. Hyg. 88, 81–85.

Cannon, R., Ruha, A.M., Kashani, J., 2008. Acute hypersensitivity reactionsassociated with administration of crotalidae polyvalent immune Fabantivenom. Ann. Emerg. Med. 51, 407–411.

Caovilla, J.J., Guardao Barros, E.J., 2004. Efficacy of two different doses ofantilonomic serum in the resolution of hemorrhagic syndromeresulting from envenoming by Lonomia obliqua caterpillars:a randomized controlled trial. Toxicon 43, 811–818.

Cardoso, J.L., Fan, H.W., França, F.O., Jorge, M.T., Leite, R.P., Nishioka, S.A.,Avila, A., Sano-Martins, I.S., Tomy, S.C., Santoro, M.L., Chudzinski, A.M.,Castro, A.S., Kamiguti, A.S., Kelen, E.M.A., Hirata, M.H., Mirandola, R.M.S., Theakston, R.D.G., Warrell, D.A.,1993. Randomized comparative trialof three antivenoms in the treatment of envenoming by lance-headedvipers (Bothrops jararaca) in Sao Paulo, Brazil. Q. J. Med. 86, 315–325.

Caron, E.J., Manock, S.R., Maudlin, J., Koleski, J., Theakston, R.D., Warrell, D.A., Smalligan, R.D., 2009. Apparent marked reduction in early anti-venom reactions compared to historical controls: was it prophylaxisgold method off administration? Toxicon Jun 9 (Epub ahead off print).

Casagrande, J.T., Pike, M.C., Smith, P.G., 1978. An improved approximateformula for calculating sample sizes for comparing two binomialdistributions. Biometrics 34, 483–486.

Chen, J.C., Bullard, M.J., Chiu, T.F., Ng, C.J., Liaw, S.J., 2000. Risk of imme-diate effects from F(ab)2 bivalent antivenin in Taiwan. WildernessEnviron. Med. 11, 163–167.

Chippaux, J.P., 1998a. Snake-bites: appraisal of the global situation. Bull.World Health Org. 76, 515–524.

Chippaux, J.P., 1998b. The development and use of immunotherapy inAfrica. Toxicon 36, 1503–1506.

Chippaux, J.P., 2002. The treatment of snake bites: analysis of require-ments and assessment of therapeutic efficacy in tropical Africa. In:Menez, A. (Ed.), Perspectives in Molecular Toxicology. John Wiley &Sons, Ltd, Australia, pp. 457–472.

Chippaux, J.P., 2006. Pratique des essais cliniques en Afrique. IRD, Paris.

J.-P. Chippaux et al. / Toxicon 55 (2010) 1195–1212 1209

Chippaux, J.P., 2008. Estimating the global burden of snakebite can help toimprove management. PLoS Med. 5, e221.

Chippaux, J.P., Goyffon, M., 1998. Venoms, antivenoms and immuno-therapy. Toxicon 36, 823–846.

Chippaux, J.P., Goyffon, M., 2008. Epidemiology of scorpionism: a globalappraisal. Acta Trop. 107, 71–79.

Chippaux, J.P., Lang, J., Amadi-Eddine, S., Fagot, P., Le Mener, V., 1999.Short report: treatment of snake envenomations by a new polyvalentantivenom composed of highly purified F(ab)2: results of a clinicaltrial in northern Cameroon. Am. J. Trop. Med. Hyg. 61, 1017–1018.

Chippaux, J.P., Lang, J., Eddine, S.A., Fagot, P., Rage, V., Peyrieux, J.C., LeMener, V.VAO (Venin Afrique de l’Ouest) Investigators, 1998. Clinicalsafety of a polyvalent F(ab’)2 equine antivenom in 223 African snakeenvenomations: a field trial in Cameroon. Trans. R. Soc. Trop. Med.Hyg. 92, 657–662.

Chippaux, J.P., Massougbodji, A., Stock, R.P., Alagon, A.Invest. AfricanAntivipmyn in Benin, 2007. Clinical trial of an F(ab0)2 polyvalentequine antivenom for African snake bites in Benin. Am. J. Trop. Med.Hyg. 77, 538–546.

Cohen, J., 1988. In: Statistical Power Analysis for the Behavioral Sciences,second ed. Lawrence Erlbaum Associates, Mahwah, NJ.

Concato, J., Horwitz, R.I., 2004. Beyond randomised versus observationalstudies. Lancet 363, 1660–1661.

Corneille, M.G., Larson, S., Stewart, R.M., Dent, D., Myers, J.G., Lopez, P.P., McFarland, M.J., Cohn, S.M., 2006. A large single-center expe-rience with treatment of patients with crotalid envenomations:outcomes with and evolution of antivenin therapy. Am. J. Surg. 192,848–852.

D’Agostino Sr., R.B., Massaro, J.M., Sullivan, L.M., 2003. Non-inferioritytrials: design concepts and issues – the encounters of academicconsultants in statistics. Statist. Med. 22, 169–186.

Dart, R.C., Seifert, S.A., Boyer, L.V., Clark, R.F., Hall, E., McKinney, P.,McNally, J., Kitchens, C.S., Curry, S.C., Bogdan, G.M., Ward, S.B.,Porter, R.S., 2001. A randomized multicenter trial of crotalinaepolyvalent immune Fab (ovine) antivenom for the treatment forcrotaline snakebite in the United States. Arch. Intern. Med. 161,2030–2036.

Dart, R.C., Seifert, S.A., Carroll, L., Clark, R.F., Hall, E., Boyer-Hassen, L.V.,Curry, S.C., Kitchens, C.S., Garcia, R.A., 1997. Affinity-purified, mixedmonospecific crotalid antivenom ovine Fab for the treatment of cro-talid venom poisoning. Ann. Emerg. Med. 30, 33–39.

Dassanayake, A.S., Karunanayake, P., Kasturiratne, K.T., Fonseka, M.M.,Wijesiriwardena, B., Gunatilake, S.B., de Silva, H.J., 2002. Safety ofsubcutaneous adrenaline as prophylaxis against acute adverse reac-tions to anti-venom serum in snakebite. Ceylon Med. J. 47, 48–49.

Daudu, I., Theakston, R.D., 1988. Preliminary trial of a new polyspecificantivenom in Nigeria. Ann. Trop. Med. Parasitol. 82, 311–313.

El Hafny, B., Ghalim, N., 2002. Evolution clinique et taux circulants duvenin dans les envenimations scorpioniques au Maroc. Bull. Soc.Pathol. Exot. 95, 200–204.

Ellis, R.M., Sprivulis, P.C., Jelinek, G.A., Banham, N.D., Wood, S.V.,Wilkes, G.J., Siegmund, A., Roberts, B.L., 2005. A double-blind,randomized trial of intravenous versus intramuscular antivenom forred-back spider envenoming. Emerg. Med. Australas 17, 152–156.

Fan, H.W., Marcopito, L.F., Cardoso, J.L., França, F.O., Malaque, C.M.,Ferrari, R.A., Theakston, R.D., Warrell, D.A., 1999. Sequential rando-mised and double blind trial of promethazine prophylaxis againstearly anaphylactic reactions to antivenom for Bothrops snake bites.Br. Med. J. 318, 1451–1452.

Faul, F., Erdfelder, E., Lang, A.G., Buchner, A., 2007. G*Power 3: a flexiblestatistical power analysis program for the social, behavioral, andbiomedical sciences. Behav. Res. Meth. 39, 175–191.

Fisher, L.D., 1999. Advances in clinical trials in the twentieth century.Annu. Rev. Public Health 20, 109–124.

Gateau, T., Bloom, M., Clark, R., 1994. Response to specific Centruroidessculpturatus antivenom in 151 cases of scorpion stings. J. Toxicol. Clin.Toxicol. 32, 165–171.

Gawarammana, I.B., Kularatne, S.A., Dissanayake, W.P., Kumarasiri, R.P.,Senanayake, N., Ariyasena, H., 2004. Parallel infusion of hydro-cortisoneþ/� chlorpheniramine bolus injection to prevent acuteadverse reactions to antivenom for snakebites. Med. J. Aust.180, 20–23.

Gibly, R., Williams, M., Walter, F.G., McNally, J., Conroy, C., Berg, R.A., 1999.Continuous intravenous midazolam infusion for Centruroides exili-cauda scorpion envenomation. Ann. Emerg. Med. 34, 620–625.

Ghalim, N., El-Hafny, B., Sebti, F., Heikel, J., Lazar, N., Moustanir, R.,Benslimane, A., 2000. Scorpion envenomation and serotherapy inMorocco. Am. J. Trop. Med. Hyg. 62, 277–283.

Ghosh, S., Maisnam, I., Murmu, B.K., Mitra, P.K., Roy, A., Simpson, I.D.,2008. A locally developed snakebite management protocol

significantly reduces overall anti snake venom utilization in WestBengal, India. Wilderness Environ. Med. 19, 267–274.

Goodman, S.N., Berlin, J.A., 1 Aug 1994. The use of predicted confidenceintervals when planning experiments and the misuse of power wheninterpreting results. Ann. Intern. Med. 121 (3), 200–206.

Grasset, E., 1957. Survey of assay methods of antivenins; immunologicalfactors influencing antivenin standardization. Bull. World Health Org.16, 79–122.

Gutierrez, J.M., Theakston, R.D., Warrell, D.A., 2006. Confronting theneglected problem of snake bite envenoming: the need for a globalpartnership. PLoS Med. 3 (6), e150.

Gutierrez, J.M., Williams, D., Fan, H.W., Warrell, D.A., 2010. Snakebiteenvenoming from a global perspective: towards an integratedapproach. Toxicon [Epub ahead of print].

Hammoudi-Triki, D., Ferquel, E., Robbe-Vincent, A., Bon, C., Choumet, V.,Laraba-Djebari, F., 2004. Epidemiological data, clinical admissiongradation and biological quantification by ELISA of scorpion enve-nomations in Algeria: effect of immunotherapy. Trans. R. Soc. Trop.Med. Hyg. 98, 240–250.

Hanvivatvong, O., Mahasandana, S., Karnchanachetanee, C., 1997. Kineticstudy of Russell’s viper venom in envenomed patients. Am. J. Trop.Med. Hyg. 57, 605–609.

Haro, L. de, Lang, J., Bedry, R., Guelon, D., Harry, P., Marchal-Mazet, F.,Jouglard, J., 1998. Envenimations par vipers europeennes. Etudemulticentrique de tolerance du Viperfav�, nouvel antivenin par voieveineuse. Ann. Fr. Anesth. Reanim. 17, 681–687.

Harrington, D.P., 2000. The randomized clinical trial. J. Am. Stat. Assoc. 95,312–315.

Harry, P., de Haro, L., Asfar, P., David, J.M., 1999. Evaluation de l’immu-notherapie antiviperine par fragments F(ab0)2 purifies (Viperfav�)par voie veineuse. Presse Med. 28, 1929–1934.

Hile, D.C., Coon, T.P., Skinner, C.G., Hile, L.M., Levy, P., Patel, M.M.,Miller, M.A., 2006. Treatment of imported fire ant stings with miti-gator sting and bite treatment – a randomized control study.Wilderness Environ. Med. 17, 21–25.

Ho, M., Silamut, K., White, N.J., Karbwang, J., Looareesuwan, S., Phillips, R.E., Warrell, D.A., 1990. Pharmacokinetics of three commercial anti-venoms in patients envenomed by the Malayan pit viper, Calloselasmarhodostoma, in Thailand. Am. J. Trop. Med. Hyg. 42, 260–266.

Hopewell, S., Clarke, M., Moher, D., Wager, E., Middleton, P., Altman, D.G.,Schulz, K.F.CONSORT Group, 2008. CONSORT for reporting random-ized controlled trials in journal and conference abstracts: explanationand elaboration. PLoS Med. 5, e20.

Huang, T.T., Blackwell, S.J., Lewis, S.R., 1978. Hand deformities in patientswith snakebite. Plast. Reconstr. Surg. 62, 32–36.

Hung, D.Z., Liau, M.Y., Lin-Shiau, S.Y., 2003. The clinical significance ofvenom detection in patients of cobra snakebite. Toxicon 41, 409–415.

Hung, D.Z., Yu, Y.J., Hsu, C.L., Lin, T.J., 2006. Antivenom treatment andrenal dysfunction in Russell’s viper snakebite in Taiwan: a case series.Trans. R. Soc. Trop. Med. Hyg. 100, 489–494.

Isbister, G.K., 2005. Snake antivenom research: the importance of casedefinition. Emerg. Med. J. 22, 399–400.

Isbister, G.K., 2006. Clinical trials in toxinology: more than a discussionbetween friends? Emerg. Med. Australas 18, 4–6.

Isbister, G.K., 2007. Safety of i.v. administration of redback spider anti-venom. Intern. Med. J. 37, 820–822.

Isbister, G.K., Brown, S.G., MacDonald, E., White, J., Currie, B.J.AustralianSnakebite Project Investigators, 2008a. Current use of Australiansnake antivenoms and frequency of immediate-type hypersensitivityreactions and anaphylaxis. Med. J. Aust. 188, 473–476.

Isbister, G.K., Brown, S.G., Miller, M., Tankel, A., Macdonald, E., Stokes, B.,Ellis, R., Nagree, Y., Wilkes, G.J., James, R., Short, A., Holdgate, A.,2008b. A randomised controlled trial of intramuscular vs. intravenousantivenom for latrodectismdthe RAVE study. Q. J. Med. 101, 557–565.

Isbister, G.K., Gray, M.R., 2003. Latrodectism: a prospective cohort study ofbites by formally identified redback spiders. Med. J. Aust. 179, 88–91.

Isbister, G.K., O’Leary, M., Miller, M., Brown, S.G., Ramasamy, S., James, R.,Schneider, J.S., 2008c. A comparison of serum antivenom concentra-tions after intravenous and intramuscular administration of redback(widow) spider antivenom. Br. J. Clin. Pharmacol. 65, 139–143.

Isbister, G.K., O’Leary, M.A., Schneider, J.J., Brown, S.G., Currie, B.J., Invest., A.S.P., 2007. Efficacy of antivenom against the procoagulant effect ofAustralian brown snake (Pseudonaja sp.) venom: in vivo and in vitrostudies. Toxicon 49, 57–67.

Jones, B., Jarvis, P., Lewis, J.A., Ebbutt, A.F., 1996. Trials to assess equiva-lence: the importance of rigorous methods. Br. Med. J. 313, 36–39.

Jorge, M.T., Cardoso, J.L., Castro, S.C., Ribeiro, L., França, F.O., deAlmeida, M.E., Kamiguti, A.S., Santo-Martins, I.S., Santoro, M.L.,Mancau, J.E., Warrell, D.A., Theakston, R.D., 1995. A randomized

J.-P. Chippaux et al. / Toxicon 55 (2010) 1195–12121210

‘blinded’ comparison of two doses of antivenom in the treatment ofBothrops envenoming in Sao Paulo, Brazil. Trans. R. Soc. Trop. Med.Hyg. 89, 111–114.

Jorge, M.T., Malaque, C., Ribeiro, L.A., Fan, H.W., Cardoso, J.L., Nishioka, S.A., Sano-Martins, I.S., França, F.O., Kamiguti, A.S., Theakston, R.D.,Warrell, D.A., 2004. Failure of chloramphenicol prophylaxis to reducethe frequency of abscess formation as a complication of envenomingby Bothrops snakes in Brazil: a double-blind randomized controlledtrial. Trans. R. Soc. Trop. Med. Hyg. 98, 529–534.

Jorge, M.T., Ribeiro, L.A., 1994. Efeito na reduçao dose de soro anti-botropico administrada a pacientes picados por serpentes do generoBothrops. Rev. Assoc. Med. Bras. 40, 59–62.

Karlson-Stiber, C., Persson, H., 1994. Antivenom treatment in Vipera berusenvenoming–report of 30 cases. J. Intern. Med. 235, 57–61.

Karlson-Stiber, C., Persson, H., Heath, A., Smith, D., Al-Abdulla, I.H.,Sjostrom, L., 1997. First clinical experiences with specific sheep Fabfragments in snake bite. Report of a multicentre study of Vipera berusenvenoming. J. Intern. Med. 241, 53–58.

Karnchanachetanee, C., Hanvivatvong, O., Mahasandana, S., 1994. Mono-specific antivenin therapy in Russell’s viper bite. J. Med. Assoc. Thai.77, 293–297.

Kasturiratne, A., Wickremasinghe, A.R., de Silva, N., Gunawardena, N.K.,Pathmeswaran, A., Premaratna, R., Savioli, L., Lalloo, D.G., de Silva, H.J.,2008. The global burden of snakebite: a literature analysis andmodelling based on regional estimates of envenoming and deaths.PLoS Med. 5, e218.

Kerrigan, K.R., Mertz, B.L., Nelson, S.J., Dye, J.D., 1997. Antibiotic prophy-laxis for pit viper envenomation: prospective, controlled trial. WorldJ. Surg. 21, 369–372.

Kimoto, T., Suemitsu, K., Nakayama, H., Komori, E., Ohtani, M., Ando, S.,1997. Therapeutic experience of venomous snakebites by the Japa-nese viper (Agkistrodon halys Blomhoffii) with low dose of antivenin:report of 43 consecutive cases. Nippon Geka Hokan [Archiv furjapanische Chirurgie] 66, 71–77.

Krifi, M.N., Amri, F., Kharrat, H., El Ayeb, M., 1999. Evaluation of antivenomtherapy in children severely envenomed by Androctonus australisgarzonii (Aag) and Buthus occitanus tunetanus (Bot) scorpions. Toxicon37, 1627–1634.

Krishnan, A., Sonawane, R.V., Karnad, D.R., 2007. Captopril in the treat-ment of cardiovascular manifestations of indian red scorpion (Meso-buthus tamulus concanesis Pocock) envenomation. J. Assoc. PhysiciansIndia 55, 22–26.

Kularatne, S.A., Gawarammana, I.B., Kumarasiri, P.V., Senanayake, N.,Dissanayake, W.P., Ariyasena, H., 2003. Safety and efficacy of subcu-taneous adrenaline as a treatment for anaphylactic reactions topolyvalent antivenom. Ceylon Med. J. 48, 148–149.

Kularatne, S.A., Kumarasiri, P.V., Pushpakumara, S.K., Dissanayaka, W.P.,Ariyasena, H., Gawarammana, I.B., Senanayake, N., 2005. Routineantibiotic therapy in the management of the local inflammatoryswelling in venomous snakebites: results of a placebo-controlledstudy. Ceylon Med. J. 50, 151–155.

Lalloo, D.G., Trevett, A.J., Black, J., Mapao, J., Saweri, A., Naraqi, S.,Owens, D., Kamiguti, A.S., Hutton, R.A., Theakston, R.D., Warrell, D.A.,1996. Neurotoxicity, anticoagulant activity and evidence of rhabdo-myolysis in patients bitten by death adders (Acanthophis sp.) insouthern Papua New Guinea. Q. J. Med. 89, 25–35.

Lalloo, D.G., Trevett, A.J., Korinhona, A., Nwokolo, N., Laurenson, I.F.,Paul, M., Black, J., Naraqi, S., Mavo, B., Saweri, A., Hutton, R.A.,Theakston, R.D.G., Warrell, D.A., 1995. Snake bites by the Papuantaipan (Oxyuranus scutellatus canni): paralysis, hemostatic and elec-trocardiographic abnormalities, and effects of antivenom. Am. J. Trop.Med. Hyg. 52, 525–531.

Lavonas, E.J., Gerardo, C.J., O’Malley, G., Arnold, T.C., Bush, S.P., Banner Jr., W., Steffens, M., Kerns. II, W.P., 2004. Initial experience with Crota-lidae polyvalent immune Fab (ovine) antivenom in the treatment ofcopperhead snakebite. Ann. Emerg. Med. 43, 200–206.

LoVecchio, F., DeBus, D.M., 2001. Snakebite envenomation in children:a 10-year retrospective review. Wilderness Environ. Med. 12, 184–189.

LoVecchio, F., Klemens, J., Roundy, E.B., Klemens, A., 2003. Serum sicknessfollowing administration of Antivenin (Crotalidae) Polyvalent in 181cases of presumed rattlesnake envenomation. Wilderness Environ.Med. 14, 220–221.

LoVecchio, F., Welch, S., Klemens, J., Curry, S.C., Thomas, R., 1999. Inci-dence of immediate and delayed hypersensitivity to Centruroidesantivenom. Ann. Emerg. Med. 34, 615–619.

Machin, D., Campbell, M., Fayers, P., Pinol, A., 1997. In: Sample Size Tablesfor Clinical Studies, second ed. Blackwell Science, Oxford, pp. 24–25.

Maclure, M., 2009. Explaining pragmatic trials to pragmatic policymakers.J. Clin. Epidemiol. 62, 476–478.

MacMahon, S., Collins, R., 2001. Reliable assessment of the effects oftreatment on mortality and major morbidity, II: observational studies.Lancet 357, 455–462.

Macpherson, H., 2004. Pragmatic clinical trials. Complement Ther. Med.12, 136–140.

Malasit, P., Warrell, D.A., Chanthavanich, P., Viravan, C., Mongkolsapaya, J.,Singhthong, B., Supich, C., 1986. Prediction, prevention, and mecha-nism of early (anaphylactic) antivenom reactions in victims of snakebites. Br. Med. J. 292, 17–20.

Mann, G., 1978. Echis colorata bites in Israel: an evaluation of specificantiserum use on the base of 21 cases of snake bite. Toxicol. Eur. Res.1, 365–369.

Meyer, W.P., Habib, A.G., Onayade, A.A., Yakubu, A., Smith, D.C., Nasidi, A.,Daudu, I.J., Warrell, D.A., Theakston, R.D., 1997. First clinical experi-ences with a new ovine Fab Echis ocellatus snake bite antivenom inNigeria: randomized comparative trial with Institute Pasteur Serum(Ipser) Africa antivenom. Am. J. Trop. Med. Hyg. 56, 291–300.

Mille, R., Saint-Cyr, C., Petit, P., 1970. La serotherapie specifique desenvenomations par « Bothrops lanceolatus ». Premier essai therapeu-tique d’un serum monovalent prepare par l’Institut Pasteur. Bord.Med. 3, 1023–1030.

Mitrakul, C., Dhamkrong-At, A., Futrakul, P., Thisyakorn, C., Vongsrisart, K.,Varavithya, C., Phancharoen, S., 1984. Clinical features of neurotoxicsnake bite and response to antivenom in 47 children. Am. J. Trop.Med. Hyg. 33, 1258–1266.

Moran, N.F., Newman, W.J., Theakston, R.D., Warrell, D.A., Wilkinson, D.,1998. High incidence of early anaphylactoid reaction to SAIMR poly-valent snake antivenom. Trans. R. Soc. Trop. Med. Hyg. 92, 69–70.

Myint-Lwin, Tin-Nu-Swe, Myint-Aye-Mu, Than-Than, Thein-Than, Tun-Pe.,1989. Heparin therapy in Russell’s viper bite victims with impendingdic (a controlled trial). Southeast Asian J. Trop. Med. Public Health 20,271–277.

Myint-Lwin, D.A., Phillips, R.E., Tin-Nu-Swe, Tun-Pe, Maung-Maung-Lay, Warrell, 1985. Bites by Russell’s viper (Vipera russelli siamensis) inBurma: haemostatic, vascular, and renal disturbances and response totreatment. Lancet 2, 1259–1264.

Narvencar, K., 2006. Correlation between timing of ASV administration andcomplications in snake bites. J. Assoc. Physicians India 54, 717–719.

Natu, V.S., Murthy, R.K., Deodhar, K.P., 2006. Efficacy of species specificanti-scorpion venom serum (AScVS) against severe, serious scorpionstings (Mesobuthus tamulus concanesis Pocock) – an experience fromrural hospital in western Maharashtra. J. Assoc. Physicians India 54,283–287.

Nuchprayoon, I., Pongpan, C., Sripaiboonkij, N., 2008. The role of prednis-olone in reducing limb oedema in children bitten by green pit vipers:a randomized, controlled trial. Ann. Trop. Med. Parasitol.102, 643–649.

Offerman, S.R., Bush, S.P., Moynihan, J.A., Clark, R.F., 2002. Crotaline Fabantivenom for the treatment of children with rattlesnake enveno-mation. Pediatrics 110, 968–971.

Offerman, S.R., Smith, T.S., Derlet, R.W., 2001. Does the aggressive use ofpolyvalent antivenin for rattlesnake bites result in serious acute sideeffects? West. J. Med. 175, 88–91.

Otero, R., Gutierrez, J.M., Nunez, V., Robles, A., Estrada, R., Segura, E.,Toro, M.F., Garcıa, M.E., Dıaz, A., Ramırez, E.C., Gomez, G., Castaneda, J.,Moreno, M.E., 1996. A randomized double-blind clinical trial of twoantivenoms in patients bitten by Bothrops atrox in Colombia. TheRegional Group on Antivenom Therapy Research (REGATHER). Trans.R. Soc. Trop. Med. Hyg. 90, 696–700.

Otero, R., Gutierrez, J.M., Rojas, G., Nunez, V., Dıaz, A., Miranda, E., Uribe, A.F., Silva, J.F., Ospina, J.G., Medina, Y., Toro, M.F., Garcıa, M.E., Leon, G.,Garcıa, M., Lizano, S., De La Torre, J., Marquez, J., Mena, Y., Gonzalez, N.,Arenas, L.C., Puzon, A., Blanco, N., Sierra, A., Espinal, M.E., Arboleda, M.,Jimenez, J.C., Ramırez, P., Dıaz, M., Guzman, M.C., Barros, J., Henao, S.,Ramırez, A., Macea, U., Lozano, R., 1999. A randomized blinded clinicaltrial of two antivenoms, prepared by caprylic acid or ammoniumsulphate fractionation of IgG, in Bothrops and Porthidium snake bites inColombia: correlation between safety and biochemical characteristicsof antivenoms. Toxicon 37, 895–908.

Otero, R., Leon, G., Gutierrez, J.M., Rojas, G., Toro, M.F., Barona, J.,Rodrıguez, V., Dıaz, A., Nunez, V., Quintana, J.C., Ayala, S., Mosquera, D.,Conrado, L.L., Fernandez, D., Arroyo, Y., Paniagua, C.A., Lopez, M.,Ospina, C.E., Alzate, C., Fernandez, J., Meza, J.J., Silva, J.F., Ramırez, P.,Fabra, P.E., Ramırez, E., Cordoba, E., Arrieta, A.B., Warrell, D.A.,Theakston, R.D., 2006. Efficacy and safety of two whole IgG polyvalentantivenoms, refined by caprylic acid fractionation with or withoutbeta-propiolactone, in the treatment of Bothrops asper bites inColombia. Trans. R. Soc. Trop. Med. Hyg. 100, 1173–1182.

Otero-Patino, R., Cardoso, J.L., Higashi, H.G., Nunez, V., Diaz, A., Toro, M.F.,Garcıa, M.E., Sierra, A., Garcıa, L.F., Moreno, A.M., Medina, M.C.,

J.-P. Chippaux et al. / Toxicon 55 (2010) 1195–1212 1211

Castaneda, N., Silva-Diaz, J.F., Murcia, M., Cardenas, S.Y., Dias daSilva, W.D., 1998. A randomized, blinded, comparative trial of onepepsin-digested and two whole IgG antivenoms for Bothrops snakebites in Uraba, Colombia. The Regional Group on Antivenom TherapyResearch (REGATHER). Am. J. Trop. Med. Hyg. 58, 183–189.

Pardal, P.P., Souza, S.M., Monteiro, M.R., Fan, H.W., Cardoso, J.L., França, F.O.,Tomy, S.C., Sano-Martins, I.S., de Sousa-e-Silva, M.C., Colombini, M.,Kodera, N.F., Moura-da-Silva, A.M., Cardoso, D.F., Velarde, D.T.,Kamiguti, A.S., Theakston, R.D., Warrell, D.A., 2004. Clinical trial of twoantivenoms for the treatment of Bothrops and Lachesis bites in the northeastern Amazon region of Brazil. Trans. R. Soc. Trop. Med. Hyg. 98, 28–42.

Paul, V., Prahlad, K.A., Earali, J., Francis, S., Lewis, F., 2003. Trial of heparinin viper bites. J. Assoc. Physicians India 51, 163–166.

Paul, V., Pratibha, S., Prahlad, K.A., Earali, J., Francis, S., Lewis, F., 2004.High-dose anti-snake venom versus low-dose anti-snake venom inthe treatment of poisonous snake bites–a critical study. J. Assoc.Physicians India 52, 14–17.

Paul, V., Pudoor, A., Earali, J., John, B., Anil Kumar, C.S., Anthony, T., 2007.Trial of low molecular weight heparin in the treatment of viper bites.J. Assoc. Physicians India 55, 338–342.

Peto, R., Pike, M.C., Armitage, P., Breslow, N.E., Cox, D.R., Howard, S.V.,Mantel, N., McPherson, K., Peto, J., Smith, P.G., 1976. Design andanalysis of randomized clinical trials requiring prolonged observa-tion of each patient. I. Introduction and design. Br. J. Cancer 34,585–612.

Phillips, R.E., Theakston, R.D., Warrell, D.A., Galigedara, Y., Abeysekera, D.T., Dissanayaka, P., Hutton, R.A., Aloysius, D.J., 1988. Paralysis, rhab-domyolysis and haemolysis caused by bites of Russell’s viper (Viperarusselli pulchella) in Sri Lanka: failure of Indian (Haffkine) antivenom.Q. J. Med. 68, 691–715.

Pizon, A.F., Riley, B.D., LoVecchio, F., Gill, R., 2007. Safety and efficacy ofCrotalidae Polyvalent Immune Fab in pediatric crotaline envenoma-tions. Acad. Emerg. Med. 14, 373–376.

Pochanugool, C., Limthongkul, S., Wilde, H., 1997. Management of Thaicobra bites with a single bolus of antivenin. Wilderness Environ. Med.8, 20–23.

Premawardhena, A.P., de Silva, C.E., Fonseka, M.M., Gunatilake, S.B., deSilva, H.J., 1999. Low dose subcutaneous adrenaline to prevent acuteadverse reactions to antivenom serum in people bitten by snakes:randomised, placebo controlled trial. Br. Med. J. 318, 1041–1043.

Reid, H.A., Thean, P.C., Martin, W.J., 1963. Specific antivenene and predni-sone in viper-bite poisoning: controlled trial. Br. Med. J. 2, 1378–1380.

Rezende, N.A., Dias, M.B., Campolina, D., Chavez-Olortegui, C., Diniz, C.R.,Amaral, C.F., 1995. Efficacy of antivenom therapy for neutralizingcirculating venom antigens in patients stung by Tityus serrulatusscorpions. Am. J. Trop. Med. Hyg. 52, 277–280.

Rojnuckarin, P., Chanthawibun, W., Noiphrom, J., Pakmanee, N.,Intragumtornchai, T., 2006. A randomized, double-blind, placebo-controlled trial of antivenom for local effects of green pit viper bites.Trans. R. Soc. Trop. Med. Hyg. 100, 879–884.

Ruha, A.M., Curry, S.C., Beuhler, M., Katz, K., Brooks, D.E., Graeme, K.A.,Wallace, K., Gerkin, R., LoVecchio, F., Wax, P., Selden, B., 2002. Initialpostmarketing experience with crotalidae polyvalent immune Fab fortreatment of rattlesnake envenomation. Ann. Emerg. Med. 39, 609–615.

Schulz, K.F., Grimes, D.A., 2005. Sample size calculations in randomisedtrials: mandatory and mystical. Lancet 365, 1348–1353.

Schwartz, D., Lellouch, J., 1967. Explanatory and pragmatic attitudes intherapeutical trials. J. Chronic Dis. 20, 637–648 (Reprinted in 2009. J.Clin. Epidemiol. 62, 499–505).

Sellahewa, K.H., Gunawardena, G., Kumararatne, M.P., 1995. Efficacy ofantivenom in the treatment of severe local envenomation by the hump-nosed viper (Hypnale hypnale). Am. J. Trop. Med. Hyg. 53, 260–262.

Sellahewa, K.H., Kumararatne, M.P., Dassanayake, P.B., Wijesundera, A.,1994. Intravenous immunoglobulin in the treatment of snake biteenvenoming: a pilot study. Ceylon Med. J. 39, 173–175.

Seneviratne, S.L., Opanayaka, C.J., Ratnayake, N.S., Kumara, K.E.,Sugathadasa, A.M., Weerasuriya, N., Wickrama, W.A., Gunatilake, S.B.,de Silva, H.J., 2000. Use of antivenom serum in snake bite:a prospective study of hospital practice in the Gampaha district.Ceylon Med. J. 45, 65–68.

Shemesh, I.Y., Kristal, C., Langerman, L., Bourvin, A., 1998. Preliminaryevaluation of Vipera palaestinae snake bite treatment in accordance tothe severity of the clinical syndrome. Toxicon 36, 867–873.

Smalligan, R., Cole, J., Brito, N., Laing, G.D., Mertz, B.L., Manock, S.,Maudlin, J., Quist, B., Holland, G., Nelson, S., Lalloo, D.G.,Rivadeneira, G., Barragan, M.E., Dolley, D., Eddleston, M., Warrell, D.A.,Theakston, R.D., 2004. Crotaline snake bite in the EcuadorianAmazon: randomised double blind comparative trial of three SouthAmerican polyspecific antivenoms. Br. Med. J. 329, 1129–1135.

Sofer, S., Shahak, E., Gueron, M., 1994. Scorpion envenomation and anti-venom therapy. J. Pediatr. 124, 973–978.

Sotelo, N., 2008. Review of treatment and complications in 79 childrenwith rattlesnake bite. Clin. Pediatr. 47, 483–489.

Srimannarayana, J., Dutta, T.K., Sahai, A., Badrinath, S., 2004. Rational useof anti-snake venom (ASV): trial of various regimens in hemotoxicsnake envenomation. J. Assoc. Physicians India 52, 788–793.

Stahel, E., Wellauer, R., Freyvogel, T.A., 1985. Vergiftungen durch einhei-mische Vipern (Vipera berus und Vipera aspis). Eine retrospektiveStudie an 113 Patienten. Schweiz. Med. Wschr. 115, 890–896.

Stock, R.P., Massougbodji, A., Alagon, A., Chippaux, J.P., 2007. Bringingantivenom to Sub-Saharan Africa. Nat. Biotechnol. 25, 173–177.

Streiner, D.L., 2007. Alternatives to placebo-controlled trials. Can. J. Neu-rol. Sci. 34, S37–S41.

Sutherland, S.K., Lovering, K.E., 1979. Antivenoms: use and adversereactions over a 12-month period in Australia and Papua New Guinea.Med. J. Aust. 2, 671–674.

Swinson, C., 1976. Control of antivenom treatment in Echis carinatus(Carpet Viper) poisoning. Trans. R. Soc. Trop. Med. Hyg. 70, 85–87.

Tariang, D.D., Philip, P.J., Alexander, G., Macaden, S., Jeyaseelan, L., Peter, J.V.,Cherian, A.M., 1999. Randomized controlled trial on the effective doseof anti-snake venom in cases of snake bite with systemic envenoma-tion. J. Assoc. Physicians India 47, 369–371.

Theakston, R.D., Fan, H.W., Warrell, D.A., Da Silva, W.D., Ward, S.A.,Higashi, H.G.The Butantan Institute Antivenom Study Group (BIASG),1992. Use of enzyme immunoassays to compare the effect and assessthe dosage regimens of three Brazilian Bothrops antivenoms. Am. J.Trop. Med. Hyg. 47, 593–604.

Theakston, R.D., Phillips, R.E., Warrell, D.A., Galagedera, Y., Abeysekera, D.T., Dissanayaka, P., de Silva, A., Aloysius, D.J., 1990. Envenoming by thecommon krait (Bungarus caeruleus) and Sri Lankan cobra (Naja najanaja): efficacy and complications of therapy with Haffkine antivenom.Trans. R. Soc. Trop. Med. Hyg. 84, 301–308.

Theakston, R.D., Warrell, D.A., Griffiths, E., 2003. Report of a WHOworkshop on the standardization and control of antivenoms. Toxicon41, 541–557.

Theakston, R.D.G., Warrell, D.A., 2000. Crisis in snake antivenom supplyfor Africa. Lancet 356, 2104.

Thiansookon, A., Rojnuckarin, P., 2008. Low incidence of early reactions tohorse-derived F(ab0)2 antivenom for snakebites in Thailand. ActaTrop. 105, 203–205.

Thomas, L., Tyburn, B., Bucher, B., Pecout, F., Ketterle, J., Rieux, D.,Smadja, D., Garnier, D., Plumelle, Y.Research Group on Snake Bites inMartinique, 1995. Prevention of thromboses in human patients withBothrops lanceolatus envenoming in Martinique: failure of anticoag-ulants and efficacy of a monospecific antivenom. Am. J. Trop. Med.Hyg. 52, 419–426.

Thomas, L., Tyburn, B., Lang, J., Ketterle, J., Biao, T., Moravie, V., Rouvel, C.,Numeric, P.Research Group on Snake Bites in Martinique, 1998.Tolerance et efficacite d’un fragment F(ab0)2 antivenimeux (equin)specifique administre par voie intraveineuse dans le traitement desmorsures de serpent (Bothrops lanceolatus) en Martinique. Rean. Urg.7, 381–387.

Thomas, P.P., Jacob, J., 1985. Randomised trial of antivenom in snakeenvenomation with prolonged clotting time. Br. Med. J. 291, 177–178.

Thorpe, K.E., Zwarenstein, M., Oxman, A.D., Treweek, S., Furberg, C.D.,Altman, D.G., Tunis, S., Bergel, E., Harvey, I., Magid, D.J., Chalkidou, K.,2009. A pragmatic-explanatory continuum indicator summary (PRE-CIS): a tool to help trial designers. J. Clin. Epidemiol. 62, 464–475.

Tin Nu Swe, Myint Lwin, Khin, Ei Han, Tin Tun, Tun P., 1992. Heparintherapy in Russell’s viper bite victims with disseminated intravas-cular coagulation: a controlled trial. Southeast Asian J. Trop. Med.Public Health 23, 282–287.

Trevett, A.J., Lalloo, D.G., Nwokolo, N.C., Naraqi, S., Kevau, I.H.,Theakston, R.D., Warrell, D.A., 1995a. The efficacy of antivenom in thetreatment of bites by the Papuan taipan (Oxyuranus scutellatus canni).Trans. R. Soc. Trop. Med. Hyg. 89, 322–325.

Trevett, A.J., Lalloo, D.G., Nwokolo, N.C., Naraqi, S., Kevau, I.H.,Theakston, R.D., Warrell, D.A., 1995b. Failure of 3,4-diaminopyridineand edrophonium to produce significant clinical benefit in neuro-toxicity following the bite of Papuan taipan (Oxyuranus scutellatuscanni. Trans. R. Soc. Trop. Med. Hyg. 89, 444–446.

Vandenbroucke, J.P., 2004. When are observational studies as credible asrandomised trials? Lancet 363, 1728–1731.

Vandenbroucke, J.P., 2008. Observational research, randomised trials, andtwo views of medical science. PLoS Med. 5, e67.

Vijeth, S.R., Dutta, T.K., Shahapurkar, J., Sahai, A., 2000. Dose andfrequency of anti-snake venom injection in treatment of Echis car-inatus (saw-scaled viper) bite. J. Assoc. Physicians India 48, 187–191.

J.-P. Chippaux et al. / Toxicon 55 (2010) 1195–12121212

Visser, L.E., Kyei-Faried, S., Belcher, D.W., 2004. Protocol and monitoring toimprove snake bite outcomes in rural Ghana. Trans. R. Soc. Trop. Med.Hyg. 98, 278–283.

Visser, L.E., Kyei-Faried, S., Belcher, D.W., Geelhoed, D.W., van Leeuwen, J.S.,van Roosmalen, J., 2008. Failure of a new antivenom to treat Echisocellatus snake bite in rural Ghana: the importance of quality surveil-lance. Trans. R. Soc. Trop. Med. Hyg. 102, 445–450.

Warrell, D.A., 2008. Unscrupulous marketing of snake bite antivenoms inAfrica and Papua New Guinea: choosing the right product – ‘what’s ina name?’. Trans. R. Soc. Trop. Med. Hyg. 102, 397–399.

Warrell, D.A., Davidson, N.M., Omerod, L.D., Pope, H.M., Watkins, B.J.,Greenwood, B.M., Reid, H.A., 1974. Bites by the saw-scaled or carpetviper (Echis carinatus): trial of two specific antivenoms. Br. Med. J. 4,437–440.

Warrell, D.A., Looareesuwan, S., Theakston, R.D., Phillips, R.E.,Chanthavanich, P., Viravan, C., Supanaranond, W., Karbwang, J., Ho, M.,Hutton, R.A., Vejcho, S., 1986. Randomized comparative trial of threemonospecific antivenoms for bites by the Malayan pit viper (Callos-elasma rhodostoma) in southern Thailand: clinical and laboratorycorrelations. Am. J. Trop. Med. Hyg. 35, 1235–1247.

Warrell, D.A., Ormerod, L.D., Davidson, N.M., 1975. Bites by puff-adder(Bitis arietans) in Nigeria, and value of antivenom. Br. Med. J. 4,697–700.

Warrell, D.A., Pope, H.M., Prentice, C.R., 1976. Disseminated intravascularcoagulation caused by the carpet viper (Echis carinatus): trial ofheparin. Br. J. Haematol. 33, 335–342.

Warrell, D.A., Warrell, M.J., Edgar, W., Prentice, C.R., Mathison, J.,Mathison, J., 1980. Comparison of Pasteur and Behringwerke

antivenoms in envenoming by the carpet viper (Echis carinatus). Br.Med. J. 280, 607–609.

Watt, G., Meade, B.D., Theakston, R.D., Padre, L.P., Tuazon, M.L.,Calubaquib, C., Santiago, E., Ranoa, C.P., 1989. Comparison of Tensilonand antivenom for the treatment of cobra-bite paralysis. Trans. R. Soc.Trop. Med. Hyg. 83, 570–573.

Watt, G., Theakston, R.D., Hayes, C.G., Yambao, M.L., Sangalang, R.,Ranoa, C.P., Alquizalas, E., Warrell, D.A., 1986. Positive response toedrophonium in patients with neurotoxic envenoming by cobras(Naja naja philippinensis). A placebo-controlled study. New Engl. J.Med. 315, 1444–1448.

WHO, 1981. Progress in the Characterization of Venoms and Standariza-tion of Antivenoms. WHO Offset Publication, Geneva.

WHO. Guidelines on Production, Control and Regulation of Snake Anti-venom Immunoglobulins. WHO, Geneva, in press.

Williams, D.J., Jensen, S.D., Nimorakiotakis, B., Muller, R., Winkel, K.D.,2007. Antivenom use, premedication and early adverse reactions inthe management of snake bites in rural Papua New Guinea. Toxicon49, 780–792.

Win-Aung, Tin-Tun, Khin-Maung-Maung, Aye-Kyaw, Hla-Pe, Tin-Nu-Swe,Saw-Naing, 1996. Clinical trial of intramuscular anti-snake venomadministration as a first aid measure in the field in the managementof Russell’s viper bite patients. Southeast Asian J. Trop. Med. PublicHealth 27, 494–497.

Yeung, J.M., Little, M., Murray, L.M., Jelinek, G.A., Daly, F.F., 2004. Anti-venom dosing in 35 patients with severe brown snake (Pseudonaja)envenoming in Western Australia over 10 years. Med. J. Aust. 181,703–705.