manuscript


doi: 10.5599/admet.3.4.257 363 

ADMET & DMPK 3(4) (2015) 363-365; doi: 10.5599/admet.3.4.257 

 
Open Access : ISSN : 1848-7718  

http://www.pub.iapchem.org/ojs/index.php/admet/index   

Opinion  

Ethical Considerations of Preclinical Testing 

Allen Goldenthal* 

Faculty of Health Sciences, University of Macau, Taipa, Macau  

*Corresponding Author:  E-mail: aegoldenthal@umac.mo; Tel.: +853-623-75280 

Received: November 17, 2015; Revised: December 02, 2015; Published: December 30, 2015  
 

Abstract 

The numbers of animal tests being conducted are on a sharp incline.  Much of this increase is directly due 
to our ability to generate transgenic models and knock-outs, thereby increasing the validity of the animal 
model but not necessarily correlating directly with any translational medical benefits to the human 
counterpart.  In spite of our best efforts, there still exist species differences that prevent the application 
directly from animal to human, and in some examples having a completely different and adverse effect 
from that seen in the animal model. There are several ways in which we can improve the opportunity for a 
positive test outcome and at the same time reduce the animal usage which is associated with our current 
animal testing practices. The benefit of the 3R’s is that they encourage us not only to avoid wastage of life 
but that they require us to provide considerable foresight and extrapolated thought before directly 
engaging in the preclinical testing phase. 

Keywords 

Animal models; in vivo; regulatory; pharmaceuticals; ethics 

 

Introduction 

It has been estimated that of every 5000 new drug development projects that make it through initial 

computer modelling and in-vitro testing, only 5 will ever successfully pass preclinical tests which are 

mandatory before clinical testing in humans.  Moreover, of those drug developments that successfully pass 

the preclinical testing phase, only 5 % ever make it to the market as a licensed treatment for humans. 

Essentially, that suggests that there is less than a 0.1 % chance of a successful drug product being 

developed from inception to market.  Although these statistics justify why drug companies must charge 

accordingly to recover their cost outputs which includes all the costs of the majority of failed projects, it 

cannot ethically justify the number of animals that are sacrificed in order to achieve this high failure rate.  

There is this vast chasm between preclinical testing in animals and the eventual successful outcome of 

clinical trials in humans into which the majority of these drug trials disappear.  In order to compensate or 

reduce this failure rate, additional safeguards must be placed on which drugs do make it to the animal 

testing phase to at least ensure that the percentage of drugs successfully passing preclinical trials is much 

higher than the current 0.1 %.  We cannot do much at our current level of testing and knowledge to 

increase the translational level of success between animals and humans, since that is still not fully 

understood why a drug is successful in one species and not another, but we certainly can narrow the gap 

http://www.pub.iapchem.org/ojs/index.php/admet/index
mailto:aegoldenthal@umac.mo


Allen Goldenthtal  ADMET & DMPK 3(4) (2015) 363-365 

364  

between drugs that appear to have great potential in the computer modelling and in-vitro stage, only to 

fail when they are introduced into animals. 

Major Challenges  

The issue with safety and toxicity testing in a preclinical trial centres on dosage. When examining the 

weight ratio of a mouse at 25 g versus the average human at 60 kg, then the ratio of 1:2400 would suggest 

that we must administer to the mouse 1/2400 of the anticipated human dose on a dose/kg basis. This 

would require working with extremely small quantities which may in turn be so small that they are 

ineffective even on an animal of such small size. That certain drugs may have a threshold level before they 

have any efficacy is a commonly found factor, but in terms of trying to identify what that threshold might 

be, often numerous pilot studies are required which not only involve time and expense on behalf of the 

Principal Investigator but may be exorbitant in the number of animals sacrificed in order to find that level. 

Even so, once established, will this threshold really correlate to the quantity that must be administered to 

humans? The dosages may be so high, that the equivalent dosage on a dose/kg basis may far exceed the 

safe levels for human administration. Sensitivity to particular drugs and treatments will vary between 

species, simply because we are different as even the best humanized models can confirm. But the danger 

in relying on the dose established in the animal model as a guideline for even administering a much lower 

dose in humans is a minefield as was clearly evident in the limited clinical trial of TGN 1412, using a mouse-

human hybrid antibody that resulted in a cytokine storm event [1].  

What that particular case brings to mind is that the animal model itself was incapable of manifesting the 

adverse response evident in humans and therefore left a huge gap of possible or unknown consequences 

that only became clear when administered into human subjects.  We can extrapolate this same concept 

into animal models of disease such as diabetes both Type I and Type II, which although similar to their 

counterparts in humans, are not identical, and as we have come to appreciate, will not respond in the 

same manner in both species to the same drug.  As can be heard in many institutions, “We have some 

excellent cures for diabetes in mice but unfortunately they have proven of little value in humans.” As some 

of these models are induced, rather than having an identical genetic predisposition or aetiology as in the 

human subject, even though we will classify them as Type I or Type II diabetes they are in fact mimics of 

the disease but not the identical human disease. 

Pre-test Requirements and Resolutions 

As such, it is important that prior to conducting a preclinical test for a drug, we must first demonstrate 

that the model we will use has been proven or shown to be a reflection of the human counterpart and 

therefore there is a higher value to any data generated from the study.  This requires that we have both in-

vitro and computational studies to verify the evidence and increase the probability of a successful 

outcome.  Without this initial testing, moving forward into animal testing should not be considered. 

Secondly, we must have all the information at hand concerning the drug being tested, or a similar drug 

from the same family of compounds, especially where it concerns previous animal models tested.  If similar 

or related class of drugs were reported in preclinical trials, having failed to move forward into human trials, 

or in fact failing in human trials, then it is ethically important to make the argument as to why the 

particular drug now being tested will have a greater chance to succeed than it did in the prior tests.  If such 

an argument cannot be made and subsequently approved, then once again animal testing should not be 

considered. 



ADMET & DMPK 3(4) (2015) 363-365 Ethical considerations of preclinical testing 

doi: 10.5599/admet.3.4.257 365 

In this vein, a third issue is that much of preclinical research outcomes are never published and that 

distorts the evidence base on which to make decisions. When we say that researchers should be 

interpreting preclinical outcomes in light of how similar drugs have performed in the past, it is very difficult 

to do that if one simply cannot consult the record of prior drugs because the results were not published as 

a corporate or institutional decision. We would want to see measures going forward that encourage and 

incentivize publication of all preclinical results, even if those results were negative, as this is vitally 

important to reduce redundancy of experiments with no chance of success, a situation which is frequently 

happening within our current framework. 

Fourthly, we should perform an up-down study using a minimum amount of animals in order to 

establish what we believe to be the NOEL (No Observable Effects Level), LOEL (Lowest Observable Effects 

Level) and MTD (Maximum Tolerated Dose).  By pairing animals at a particular dose then observing for 

effects and based on those results determining if the next pair of animals receive a higher or lower 

administrated dosage, then we can establish more effectively the groupings to be used for any preclinical 

study. 

Finally, we must give consideration to intercession time points when establishing the protocol or any 

preclinical studies. This practice will identify key critical decision points where the study should be 

terminated for reasons of either severe adverse effects or no indication of any desired effects. These 

require that a scale be predetermined of various levels of adverse effects, as well as one on determining 

data to quantify that there is absolutely no beneficial outcome before the preclinical study commences.   

Conclusions 

It must be remembered that Preclinical studies are all operated under the umbrella of Good Laboratory 

Practices (GLP) and therefore any Principal Investigator conducting such studies must be familiar with the 

GLP and that their assistants are also equally aware of the regulatory requirements for conducting a 

preclinical study before they participate in the trial.  Straying from the GLP not only places the study at risk 

but also any data generated from such a study since it may not accurately reflect the actual effects of the 

drug but may be unduly influenced and adversely affected by both poor husbandry and animal 

management.  

The obtained results from animal testing must undergo several checks for reliability before presuming 

there will be a correlating beneficial effect in humans. Certain adverse effects of the therapy may go 

completely undetected because we do not have the ability to recognize subclinical or monitor low 

incidence effects in animals such as muscle aches, low grade fever, and even psychological manifestations 

that might be occurring within the animal even when being observed [2]. Hence, it is imperative that we 

improve our skills in animal modelling, test design, and effect recognition, if we are to attain the benefits 

that animal testing can provide. 

References 

[1] H. Attarwala. J. Young Pharm. 2(3) (2010) 332-336. 

[2] V. Baumans. Gene Therapy 11 (2004) S64-S66. 

©2015 by the authors; licensee IAPC, Zagreb, Croatia. This article is an open-access article distributed under the terms and 
conditions of the Creative Commons Attribution license (http://creativecommons.org/licenses/by/3.0/)  

http://creativecommons.org/licenses/by/3.0/