Available online at www.sciencedirect.com

www.elsevier.com/locate/yrtph

Regulatory Toxicology and Pharmacology 50 (2008) 345–352

A European pharmaceutical company initiative challengingthe regulatory requirement for acute toxicity studies in

pharmaceutical drug development

Sally Robinson a,*, Jean-Luc Delongeas b, Elizabeth Donald c, David Dreher d,Matthias Festag e, Sophie Kervyn f,1, Ann Lampo g, Kamil Nahas h, Vicente Nogues i,

Deborah Ockert j,2, Kirsty Quinn k, Sally Old l,3, Nigel Pickersgill m, Kev Somers n,Claudia Stark o, Peter Stei p, Lynne Waterson q, Kathryn Chapman r,*

a AstraZeneca, Alderley Park, UKb Servier, Neuilly Sur Seine, France

c Charles River Laboratories, Edinburgh, UKd Covance Laboratories, Harrogate, UK

e Roche, Basel, Switzerlandf Eli Lilly and Company, UCB Pharma SA, Braine-l’Alleud, Belgium

g Johnson and Johnson, Beerse, Belgiumh Pfizer, Amboise, France

i Novartis, Basel, Switzerlandj Nycomed GmbH, Hamburg, Germany

k Aptuit, Edinburgh, UKl Sanofi-Aventis, Alnwick, UK

m MDS Pharma Services, Lyon, Francen GlaxoSmithKline, Ware, UK

o Bayer Schering Pharma AG, Berlin, Germanyp Boehringer Ingelheim, Biberach, Germany

q Huntingdon Life Sciences, Huntingdon, UKr UK National Centre for the Replacement, Refinement and Reduction of Animals in Research (UK NC3RS), London, UK

Received 24 October 2007Available online 5 December 2007

Abstract

Regulatory guidelines indicate acute toxicity studies in animals are considered necessary for pharmaceuticals intended for human use.This is the only study type where lethality is mentioned as an endpoint. The studies are carried out, usually in rodents, to support mar-keting of new drugs and to identify the minimum lethal dose. A European initiative including 18 companies has undertaken an evidence-based review of acute toxicity studies and assessed the value of the data generated. Preclinical and clinical information was shared on 74compounds. The analysis indicated acute toxicity data was not used to (i) terminate drugs from development (ii) support dose selectionfor repeat dose studies in animals or (iii) to set doses in the first clinical trials in humans. The conclusion of the working group is thatacute toxicity studies are not needed prior to first clinical trials in humans. Instead, information can be obtained from other studies,

0273-2300/$ - see front matter � 2007 Elsevier Inc. All rights reserved.

doi:10.1016/j.yrtph.2007.11.009

* Corresponding authors. Fax: +44 01625 516809 (S. Robinson), +44 0 207 670 5178 (K. Chapman).E-mail addresses: sally.robinson@astrazeneca.com (S. Robinson), kathryn.chapman@nc3rs.org.uk (K. Chapman).

1 Represented Eli Lilly until the end of 2006 prior moving to UCB Pharma SA.2 As of 1 July 2007. Previously ALTANA Pharma AG.3 Sanofi-Aventis were represented separately as their former legacy companies prior to the merger in 2005.

346 S. Robinson et al. / Regulatory Toxicology and Pharmacology 50 (2008) 345–352

which are performed at more relevant doses for humans and are already an integral part of drug development. The conclusions have beendiscussed and agreed with representatives of regulatory bodies from the US, Japan and Europe.� 2007 Elsevier Inc. All rights reserved.

Keywords: Acute toxicity; Pharmaceuticals; Regulatory guidelines

1. Introduction

1.1. Background

The pharmaceutical industry recognises the need to re-assess the design and conduct of toxicology studies asnew scientific practices and knowledge develop. The assess-ment often includes the consideration of the 3Rs principles(Russell and Burch, 1959) such that studies that are notproviding optimal information are challenged and oppor-tunities for replacing or reducing animal use are identified.

Acute toxicity studies in animals are considered neces-sary for any pharmaceutical intended for human use. His-torically, the first toxicity test performed has been the acutetoxicity study (Casarett and Doull, 1999). The main objec-tive of these studies is to identify a single dose causingmajor adverse effects/life threatening toxicity, which ofteninvolves an estimation of the minimum dose causing lethal-ity. The studies are usually carried out in rodents, and con-sist of a single dose up to a limit of 2000 mg/kg, or themaximum technically achievable. In pharmaceutical drugdevelopment this is the only study type where lethality/life-threatening toxicity is an endpoint as documented incurrent regulatory guidelines (Commission Directive,2003; Center for Drug Evaluation and Research, 1996;ICH Japan, 1999). Scientists and animal protection groupshave criticized the conduct of acute toxicity studies (Lang-ley, 2005). The predictivity of rodent acute toxic effects forhuman acute toxic effects has long been disputed but neverproven. In addition, in recent years there has been a changein strategy in the pharmaceutical industry such that earlytoxicology information is used to identify the most promis-ing compounds to progress as potential human medicines.Acute toxicity tests are often no longer the first toxicologystudies conducted therefore prompting this reassessment ofacute toxicity tests in pharmaceutical drug development. Inthis paper we present evidence-based arguments to demon-strate that these studies have limited value in terms of pre-clinical and human safety assessment compared to thesubstantial adverse effects experienced by some of theanimals.

1.2. Current regulatory climate

In recent years, progress has been made in refining acutetoxicity studies. Alternatives to the traditional LD50 test,designed in 1927 (Trevan, 1927) have been developed thatsignificantly improve animal welfare. These are the fixeddose procedure, the acute toxic class method, and the upand down procedure (Botham, 2002) and these are partic-

ularly used in the hazard classification of chemicals. In2002, the Organisation for Economic Co-Operation andDevelopment (OECD) eliminated the oral LD50 (lethaldose in 50% of animals) test from its guidelines for the test-ing of chemicals. The International Conference on Harmo-nisation of the Technical Requirements for theRegistration of Pharmaceuticals for Human Use (ICH)has also provided the opportunity to refine acute toxicitytests, with the acceptance of data from dose escalationstudies (ICH M3, 1997). However, ICH refers to nationalguidelines where they exist and these still indicate therequirement for acute toxicity data for pharmaceuticaldrugs. The requirements in terms of the species, exposureroute and observation period vary between geographicalregions as shown in Table 1. The European and US guide-lines specify a second route that ensures exposure. It isworth noting that acute toxicity is the only study typewhere a route other than the clinical route is routinelyrequired. Evidence of exposure in other study types is pro-vided by a toxicokinetic assessment.

1.3. Working group objectives

A European industry-working group was established in2003 and currently represents 13 pharmaceutical compa-nies and 5 contract research organisations (see author affil-iations). Aptuit, HLS and UCB Pharma SA joined in 2007and their data has been added to the survey retrospec-tively. The aim of the working group is to facilitate co-operation and data sharing on conventional acute toxicitystudies in rodents. The objectives were (a) to review howacute toxicity data are gathered and used across the phar-maceutical industry and (b) to develop a strategy for chal-lenging the guidelines on the requirement for acute toxicitywhere lethality/life threatening toxicity is an endpoint. Theultimate goal of the working group is to assess the rele-vance of acute toxicity data in the drug development pro-cess and to provide recommendations for alternativetesting strategies to replace acute toxicity studies by usinginformation derived from other studies. Acute toxicitytesting of chemicals and intermediates by the pharmaceu-tical and chemical industries was not addressed as this iscarried out for a different purpose and under different reg-ulatory guidance. In addition, specifically designedextended single dose studies (e.g., microdosing, CHMP,2004) and single dose studies for biotechnology products(ICH S6, 1997) are not part of the scope of this initiativeas again the purpose and regulatory guidance for thesestudies are different to those for standard therapeuticagents and regimens.

Table 1Regulatory framework

EEC US Japan

2 Speciesa 2 Species b 2 Speciesb

2 Routes, clinical route plus a routeensuring exposurec

2 Routes (asEEC)

Clinical

7–14-Day observation 14-Dayobservation

14-Dayobservation

a Usually rat and mouse.b US/Japan: specify a non-rodent. Dose-escalation is acceptable alter-

native for non-rodents; no region mentions dose-escalation as alternativefor rodents.

c Only study type where a second route (parenteral, usually intravenous)is routinely required.

S. Robinson et al. / Regulatory Toxicology and Pharmacology 50 (2008) 345–352 347

This initiative has the full support of the European Fed-eration of Pharmaceutical Industries and Associations(EFPIA), in particular the preclinical safety group.

2. Obtaining information on current practice

2.1. Initial discussions and data sharing

Historically, there were preclinical (Sections 2.1.1 &2.1.2) and clinical (Section 2.1.3) objectives proposed forconducting acute toxicity studies in the development ofnew medicines.

2.1.1. Dose selection

Preclinically, it was often assumed that the data fromacute toxicity studies might be used to aid dose selectionfor other studies and provide preliminary target organ tox-icity. However, initial discussions within the working groupindicated that for many companies the acute study is nolonger the first toxicology study performed and dose-selec-tion information is gained from many other study types,including non-GLP (Good Laboratory Practice) explor-atory studies.

2.1.2. Identification of target organs of toxicityIt is also often stated that such studies support identifica-

tion of target organs of toxicity. However, since microscopicpathology is not routinely performed on acute toxicologystudies they of very limited value to identify target organs.

2.1.3. Assessment of acute overdose situation

Clinically, it is assumed that the data obtained from acutetoxicity studies is used to give information on the likelyeffects of acute overdose in humans. Initial discussions indi-cated that the studies do not include clinical pathology,microscopic pathology or toxicokinetic evaluation and thatthe clinical observations seen in rodents at high doses areoften non-specific and may help in assessing the lethaldose-range, but they do not add information that wouldsupport measures to be taken in overdose situations inhumans. Therefore, studies evaluating vital organ functionsin more detail, such as safety pharmacology studies, mightbe appropriate to support assessment of acute overdose.

An initial data sharing exercise was undertaken by theworking group to compare the design of acute toxicity stud-ies and to review how internal regulatory and clinical col-leagues use the data generated. All of the companiesinvolved in the working group completed a questionnaireon study designs for pharmaceutical drugs at the time theyjoined the working group. This included questions on stan-dard acute toxicity package and the timing and purpose ofthe studies. Detail was also gathered on specific study designsand numbers of animals used for acute toxicity studies.

Data were collected from all companies able to supplythe information. One contract research organisation hasno set policy on acute toxicity studies and no informationwas included for this company. Therefore, in some of thefigures presented in this paper the total number of compa-nies may not equal 18.

2.2. Results

At the start of the initiative there was variability in theacute toxicity packages conducted by the companies repre-sented on the working group (Fig. 1a). For the majority ofthe companies the acute studies conducted were consistentwith the European guideline, using two species (rat andmouse) and two routes of administration i.e. the intendedhuman route of administration and a parenteral route (fourrodent studies in total). The current European Guidelinesdo not specifically request a non-rodent species where (a)lethality is not an acceptable endpoint and (b) dose-escala-tion studies are an acceptable alternative.

Several companies were using minimised study packagessuccessfully within the confines of the current regulatoryguidelines.

This variability in acute toxicity packages is not uniqueto Europe. Six US companies completed the questionnaireand the results showed that there was a similar variabilityin acute toxicity packages from one rodent study to fourrodent studies. The data from the US companies is notincluded in the figures in this paper.

From current study designs, in general only very limiteddata, other than maximum non-lethal dose and minimumlethal doses, are provided (Fig. 2). The microscopic evalu-ation of selected tissues is extremely rare and done on acase-by-case basis. It is clear from the information collectedthat these studies are not used to identify target organs.

There was large inter-company variability in the num-bers of rodents used at the start of the initiative (Fig. 3a/b). This is particularly evident in the use of mice, with fourcompanies not using any mice and six companies using 60to 100 or more mice per compound. The companies notusing mice were those that had already employed a mini-mised approach to acute toxicity testing.

The data also illustrated that the majority of the compa-nies (13/17) conducted acute toxicity studies prior to thefirst clinical trial in humans. This timing appeared to bedriven by the ICH M3, 1997 guideline that specifies therequirement for data prior to first clinical trial in humans,

Fig. 1. (a) Survey of standard acute toxicity packages (2003–2005). (b) Survey of standard acute toxicity packages (2006). (c) Survey of standard acutetoxicity packages (2007).

Fig. 2. Data provided in toxicity studies.

348 S. Robinson et al. / Regulatory Toxicology and Pharmacology 50 (2008) 345–352

although the guideline indicates dose escalation data are anacceptable alternative. Interestingly, one company con-ducted no acute toxicity studies at all. Two companies con-ducted studies prior to registration rather than prior to the

first clinical trial in humans. This approach places the acutetoxicity studies later in the development programme result-ing in a reduction in the number of compounds tested dueto attrition during the development process.

From the information collected it is clear there is inconsis-tency in the approach to providing acute toxicity data withinthe pharmaceutical industry. By sharing information theworking group was able to demonstrate there was both thescope and willingness to agree upon a harmonised approachto reduce the number of studies and refine the design in theshort term. In addition, an assessment of the scientific valueof the data generated would form the basis to challenging therequirement for these studies in the longer term.

3. Obtaining information on how acute toxicity data are used

prior to the first clinical trial in humans

3.1. Compound specific data sharing

As the majority of companies conducted acute toxicitystudies prior to the first clinical trial in humans, a data

Fig. 3. a and b show the downward shift in number of rodents used perproject in each company between 2003–2005 and 2007.

S. Robinson et al. / Regulatory Toxicology and Pharmacology 50 (2008) 345–352 349

sharing exercise was conducted to assess how acute toxicitydata are used at this stage in drug development.

In this data sharing exercise two questions wereaddressed. (a) Are acute toxicity data used to stop develop-ment of compounds prior to the first clinical trial inhumans? (b) Are acute toxicity data used to help definedose levels or to identify additional clinical monitoringrequired for the first clinical trial in humans?

Table 2aData showing effects of concern for compound X and the studies where these

Therapeuticclass

Toxicologicaleffect(s) ofconcern

Specieseffect(s)wereidentifiedin

Primarystudy type(s)in whicheffect(s) wereidentified

Doses atwhich theeffects ofconcern wereseen (mg/kg)

Gastrointestinal CytochromeP4501A1induction

Rat One monthstudy

19, 88, 38. Nono-adverseeffect level(NOAEL)

Data were collected on 34 compounds for which devel-opment was terminated prior to administration to humansand 40 compounds that entered into clinical trials inhumans. An example of the types of data collected for acompound that was given in clinical trials is illustrated inTables 2a and 2b. The types of information collectedincluded therapeutic class, toxicological effects of concern,species and study type effects were identified in, dose levelsrequired for effects of concern, reasons for termination ofthe compound, or justification for dose levels in man.Available acute toxicity data were also summarised.

The following therapy areas were covered by the data oncompounds that entered clinical trials in humans: cardio-vascular, CNS/PNS, metabolic disorders, urology, inflam-matory disorders, endocrine, gastro-intestinal, respiratoryand anti-infective (Table 3).

The data sharing demonstrated the following key-points:

- Acute toxicity studies were never responsible for termi-nation of a compound. The most common reason fortermination of compounds prior to the first clinical trialin humans was the identification of dose limiting targetorgan toxicity in the 2 or 4 week repeat dose toxicitystudies (Table 4).

- No additional information was gained from studies in asecond rodent species. Data from 28 compounds givenorally in both the rat and the mouse showed similareffects (e.g., either in terms of clinical observations orlethality) for 22 compounds at comparable doses (lessthan 20% difference in dose between the two species).For the remaining six compounds there was no morethan a twofold difference in doses showing similar clini-cal observations or lethality between the two species.The picture was similar for 22 compounds given paren-terally in both the rat and the mouse.

- For approximately 25% of the compounds assessed, thedata indicated non-rodents were more sensitive indetecting effects that determined dose levels in the firsthuman clinical trials and therefore data from short termnon-rodent studies already conducted as part of drugdevelopment may provide more useful information thanacute studies in rodents.

were identified

Justification for starting dose inman

Justificationformaximumdoseadministeredin man

Reasonforstoppingdoseescalationin man

Comments

Cytochrome P4501A1induction not consideredrelevant to man. Effects in thedog were used to limit doses/exposure in man

Exposure atthe NOAELin the dog

— —

Table 2bSummary of acute toxicity data for compound X

Therapeuticclass

Effects of concern observed inacute toxicity (clinical route,oral)

Dose range in mg/kg(clinical route)

Effects of concernobserved in acutetoxicity(parenteral route)

Dose rangein mg/kg(parenteralroute)

Comments Was the acutedata used to setor justify humandose levels

Gastrointestinal Mouse: not performedRat acute: increased salivation,ploughing, pilo-erection, ataxiaand hunched posture on theday of dosingDog dose-escalation study:Emesis

Mouse: not performedRat: minimum lethaldose >2000 mg/kgDog: 25–300 mg/kg.Emesis with noincrease in systemicexposure with dose

Mouse: notperformedRat: notperformed

Mouse: notperformedRat: notperformed

No mouse data.No parenteralrouteDog dose escala-tion data used asan alternative toacute data

No

Table 4The reasons for termination of the 34 compounds prior to the first clinicaltrial in humans

Reason for termination Number ofcompounds

Insufficient bioavailability 1Chemical stability issues 1Safety pharmacology issues 2Genetic toxicity issues 2Dose-limiting target organ effects in 2 and 4 week

repeat dose studies with no margin of safety, and/orno biomarker

22

Business reasons (e.g. class discontinued, switch tomolecule with better safety or efficacy profile)

5

Potential for developmental toxicity 1

Table 3Numbers of compounds for which data was shared within each therapyarea

Therapy area Compoundsterminatedprior to the firstclinicaltrial in humans

Compoundsthatentered clinicaltrials inhumans

Cardiovascular 5 3Central/peripheral nervous

system9 9

Metabolic 4 7Inflammatory 5 5Respiratory 7 4Endocrine 1 3Gastrointestinal 0 5Anti-infective 2 1Urology 0 3Oncology 1 —

Total = 34 Total = 40

350 S. Robinson et al. / Regulatory Toxicology and Pharmacology 50 (2008) 345–352

- Acute toxicity studies were never used to set the dose lev-els for the first clinical trials in humans, identify targetorgans or to indicate markers for monitoring in the clin-ical trials. This information was obtained from the twoto four week repeat dose toxicity studies where full bio-

chemical and histopathological evaluations were per-formed along with the measurement of plasmaconcentrations of drug.

- No additional information was obtained from the paren-teral route. The most common parenteral route wasbolus intravenous injection giving a pattern of exposurecompletely different from that anticipated for the clinicalroute and often leading to a rapid onset of both clinicalobservations and substantial adverse effects. In addition,19 of the 40 compounds that were given in clinical trialsto humans did not have acute studies conducted by theparenteral route, demonstrating that this informationwas not essential.

3.2. Working group recommendations based on data-sharing

Based upon the review of the 40 compounds that pro-ceeded to human clinical trials, the following recommenda-tions are made by the working group:

• Acute toxicity studies should not be required prior tothe first clinical trial in humans.

• Any short term or dose-escalation data (including fromnon-GLP studies) should be considered acceptable toallow assessment of acute toxicity and this should beby the clinical route only.

• In situations where overdosing may be a cause for con-cern (e.g., where there may be uncontrolled access todrug as in Phase 3 clinical trials or at registration) estab-lish, by reviewing both the pre-clinical and clinical data-set, whether additional high dose information isnecessary. In the majority of cases a stand-alone acutetoxicity study should not be required.

• Lethality should not be the primary endpoint.• Rather than using a second rodent species, data should be

provided from a rodent and a non-rodent species prefer-ably from short duration studies that are already an inte-gral part of drug development and that includeparameters in the study design that assist risk assessment.

In November 2006 a workshop that included representa-tives of the European, Japanese and United States regula-

S. Robinson et al. / Regulatory Toxicology and Pharmacology 50 (2008) 345–352 351

tory bodies was held to discuss this topic. The participants,including the regulatory representatives, fully supportedthe recommendations of the working group (Chapmanand Robinson, 2007).

4. Repeat survey of working group on current practice

The questionnaire on acute toxicity packages and studydesign was repeated in May 2006 and June 2007. From theresults of the repeat surveys, it was clear that the ongoingdiscussions within the working group and other workinggroup activities had already facilitated the review of acutetoxicity strategy within some of the working group compa-nies. The working group members have adopted aminimised approach within the current regulatory require-ments based on other companies experience in providinglimited or no data from acute toxicity studies to regulators.During the time the working group has been reviewing thevalue of acute toxicity studies, the number of companiesproviding the full four-study package has significantlydecreased. By 2007, the majority of companies now provideonly one acute study plus other data, or no acute studies atall (Fig. 1b and c). These minimised strategies are withinthe constraints of the current regulatory guidelines. Inaddition, companies conducting acute toxicity studies priorto first clinical trial in humans have reduced from 13/17 ini-tially to 6/17 in 2007 thus reducing the numbers of com-pounds for which acute toxicity studies are conducted.The number of rodents used in acute toxicity studies hasalso significantly decreased (Fig. 3). This is predominantlydue to the fact that the majority of companies are conduct-ing few if any acute toxicity studies. Where they are stillconducted there has also been a reduction in the numberof rodents used per study due to the sharing of best practicewithin the group.

ICH M3 is currently undergoing revision (Section 6) andit will take time to finalise the guideline. In the meantimethe working group is continuing to communicate its recom-mendations more widely to regulators, colleagues and com-panies not involved in the working group. Thisdissemination strategy has included leaflets, posters, pre-sentations at conferences and an international regulatoryworkshop. Through the CROs on the working group,smaller companies are receiving information packs con-taining the information on the initiative and its recommen-dations. The significant changes in acute toxicity packagesbetween 2006 and 2007 (Fig. 1b and c) indicate that theregulatory workshop and other dissemination activitiesduring 2006 have had an impact on the number of acutetoxicity studies carried out.

5. Discussion

It is clear that the strategy for toxicity testing duringpharmaceutical development has changed significantly overthe years such that early toxicology information is used toidentify the most promising compounds to progress as

potential human medicines. Acute toxicity tests are oftenno longer the first tests performed and more useful datafor dose setting in other animal studies are obtained fromstudies that are already an integral part of drug develop-ment and that do not have lethality or life threatening tox-icity as an endpoint. Additionally it is clear from the sharingof data and analysis of 40 compounds that went into humanclinical trials that acute toxicity data should not be requiredbefore first clinical trial in humans. Regulatory bodies in allregions support this recommendation. The only possibleclinical reason for conducting acute toxicity studies is tosupport risk assessment in humans at a time when patientswill have uncontrolled access to large quantities of drug,although the usefulness of the limited animal data set pro-vided from acute toxicity studies is still questionable.

The recommendations from this work are specificallyfocussed on the development of new medicines and regula-tory guidance for pharmaceuticals. However, in Zbindenand Flury-Roversi (1981) reviewed the significance of theLD50 test for the toxicological evaluation of chemical sub-stances. Some of the points made then have been clearlyre-emphasised by the recent work e.g. that the data is extre-mely limited without laboratory diagnostic and histopa-thological assessments and that the data may not be thatuseful in dose setting for other animal studies. Given thatthis work is nearly 30 years old and the fact that the origi-nal ICH M3 clearly tried to reduce acute toxicity testing forpharmaceuticals over 10 years ago it might be questionedwhy there has not been more impact in the pharmaceuticalindustry. The section on single dose/acute toxicity in thecurrent ICH M3 guideline is extremely brief and lackingin detail and therefore refers to national guidelines formore detail. It is possible that the lack of clarity in thenational guidelines underpinning ICH M3 has been aninadvertent barrier to the implementation of the 3Rs.

The approach taken by the working group has demon-strated the value of sharing and collating data in order toharmonise study designs and implement the 3Rs. Moreimportantly the group have analysed how toxicologicaldata on active pharmaceuticals are practically used to sup-port the safety of potential new medicines given in humanclinical trials. It has been the use of relevant data sets thathas provided the evidence needed to convince both thepharmaceutical companies involved and the regulatorybodies that acute toxicity studies are clearly of limitedvalue in pharmaceutical drug development. This hasallowed the final steps to be taken to implement the reduc-tion and replacement of acute toxicity testing in pharma-ceutical drug development.

Establishing whether acute toxicity studies are requiredfor predicting human overdose is the next stage and anongoing collaboration with the Lyon Poison Centre willbe pivotal in determining whether acute toxicity informa-tion is needed to provide information that assists in themanagement of overdose situations.

This initiative illustrates the benefits that can be achievedin terms of implementing the 3Rs by a coordinated

352 S. Robinson et al. / Regulatory Toxicology and Pharmacology 50 (2008) 345–352

approach and sharing data within the pharmaceuticalindustry and with regulatory bodies to reach a commonposition based on evidence and science. It is important toconsider where else a similar strategy could be applied toenhance the application of the 3Rs.

6. Conclusion

Acute toxicity data are:

• Extremely limited with regard to the parameters exam-ined, concentrating on minimum lethal and maximumnon-lethal doses.

• Of less use than other, less harmful animal tests that aresuperior for deciding appropriate doses for further stud-ies using animals.

• Do not provide information on the nature of toxiceffects, which are better evaluated in other routinestudies.

• Not, in practice, used to set doses in the first human clin-ical trials because other routine studies provide moreinformative data.

The topic of acute toxicity requirements is currentlybeing addressed via the revision of ICH M3 and this hasbeen an opportunity to incorporate the recommendationsmade by the working group into the new guidance. Therecommendations were supported by representatives ofthe international regulatory bodies at the workshop heldin November 2006 (Chapman and Robinson, 2007). At thisstage the draft ICH guidance has incorporated all the rec-ommendations of the working group. Acceptance of therecommendations will effectively lead to ‘replacement’ ofacute studies in time as non-GLP dose-range finding infor-mation from rodent and non-rodents should be considered

to be acceptable supporting information on single highdose toxicity. The results show that there should be norequirement for stand-alone acute toxicity studies as theseare not pivotal in human safety assessment.

References

Commission Directive 2003/63/EC of the 25th June 2003 amendingdirective 2001/83/EC of the European parliament and of the councilon the community code relating to medicinal products for human use,Annex I. Official Journal of the European Communities L159 27/6/2003, pp. 46–94, Brussels.

Botham, P.A., 2002. Acute Systemic Toxicity. ILAR J, 2002;43 Sup-pl:S27–30. Review.

Casarett and Doull’s Toxicology. The basic science of poisons. (FifthEdition). Companion Handbook. Edited by C.D. Klaassen and J.B.Watkins III. Published by McGraw-Hill. 1999.

Center for Drug Evaluation and Research (CDER) Guidance forIndustry, August 1996. Single dose acute toxicity testing for pharma-ceuticals. Pt 1. <http://www.fda.gov/CDER/GUIDANCE/pt1.pdf>.

Chapman, K., Robinson, S., 2007. Challenging the regulatory requirementfor acute toxicity studies in the development of new medicines. Aworkshop report. <www.nc3rs.org.uk>.

EMEACHMP/SWP/2599/02/Rev 1, 23 June 2004. Position paper on non-clinical safety studies to support clinical trials with a single microdose.London. (CHMP adopted June 2004).

ICH Japan, 1999. Guidelines for New Drug Registration with Harmo-nized Tripartite Guidelines. Yakuji Nippo Ltd, Japan.

ICH M3, 1997. Guidance on non-clinical safety studies for the conduct ofhuman clinical trials for pharmaceuticals. <http://www.ich.org>.

ICH S6, 1997. Preclinical safety assessment of biotechnology-derivedpharmaceuticals. <http://www.ich.org>.

Langley, G., 2005. European coalition to end animal experiments: acutetoxicity testing without animals. ECEAE, London.

Russell, W.M.S., Burch, R.L., 1959 (reprinted 1992). The principles ofhumane experimental technique. Universities Federation for AnimalWelfare, Wheathamsted, UK.

Trevan, J., 1927. The error of determination of Toxicity. Proc. R. Soc.101, 483–514.

Zbinden, G., Flury-Roversi, M., 1981. Significance of the LD50-test forthe toxicological evaluation of chemical substances. Arch. Toxicol. 47,77–99.