Eur J Oral Sci 2015; 123: 61–64 Ó 2015 Eur J Oral Sci

DOI: 10.1111/eos.12175 European Journal of

Printed in Singapore. All rights reserved
Oral Sciences

Focus
Animal research as a basis for clinical Clovis M. Faggion Jr
Department of Periodontology, Faculty of
€nster, Mu
€nster,
trials
Dentistry, University of Mu
Germany

Faggion CM Jr. Animal research as a basis for clinical trials.

Eur J Oral Sci 2015; 123: 61–64. © 2015 Eur J Oral Sci
Animal experiments are critical for the development of new human therapeutics
because they provide mechanistic information, as well as important information on
efficacy and safety. Some evidence suggests that authors of animal research in den-
tistry do not observe important methodological issues when planning animal experi-
ments, for example sample-size calculation. Low-quality animal research directly
interferes with development of the research process in which multiple levels of
research are interconnected. For example, high-quality animal experiments generate
sound information for the further planning and development of randomized con-
Clovis M. Faggion Jr., Faculty of Dentistry,
trolled trials in humans. These randomized controlled trials are the main source for €nster, Waldeyerstraße 30,
University of Mu
the development of systematic reviews and meta-analyses, which will generate the 48149 Mu €nster, Germany
best evidence for the development of clinical guidelines. Therefore, adequate plan-
ning of animal research is a sine qua non condition for increasing efficacy and effi- E-mail: [Link]@[Link]
ciency in research. Ethical concerns arise when animal research is not performed
with high standards. This Focus article presents the latest information on the stan- Key words: animal experimentation; clinical
dards of animal research in dentistry, more precisely in the field of implant den- trials; systematic review; meta-analysis;
tistry. Issues on precision and risk of bias are discussed, and strategies to reduce systematic review
risk of bias in animal research are reported. Accepted for publication January 2015

Animal experiments are critical for the development of a randomized controlled trial in which it is technically
new human therapeutics because they provide mecha- impossible to blind personnel (surgeon and patient)
nistic information, as well as important information on owing to the evaluated therapies being heterogeneous,
efficacy and safety. Although the number of animals as low quality. Nevertheless, this type of study may still
used in research seems to decrease over the years, there present ROB if the inability to blind interferes with the
are currently millions of animals being used for experi- final outcome. Therefore, ROB levels, rather than qual-
mental purposes (1). In dental specialties, such as ity scores, should be used to determine the magnitude
implant dentistry, new materials and techniques are of effectiveness of treatments.
regularly tested on animals, and the number of publica-
tions describing animal experiments in implant den-
tistry has increased considerably in recent years
Hierarchy in the research process
(Fig. 1). Therefore, it is necessary to evaluate whether
these experiments are performed rigorously. Several types of research methods are used to evaluate
The aims of this Focus article were to describe the con- different outcomes in dental research within a hierarchi-
cept of research hierarchy (a process in which multiple cal research process (Fig. 2). Ideally, the findings at
levels of research are conducted to increase efficacy and each level guide subsequent phases of research; for
efficiency) and to report and discuss the current animal example, research output from animal experiments
research standards in the field of implant dentistry. guides clinical research. Therefore, research performed
at each level should be of the highest quality possible
and have a low level of bias to ensure high accuracy.
As shown in Fig. 2, the final step of the research pro-
Quality and risk of bias
cess is the development of clinical guidelines to improve
It is important to differentiate between quality and risk the quality of clinical treatments. By improving all
of bias (ROB). Quality refers to the level of rigor in phases of research, including animal research, the most
research studies, whereas ROB refers to the extent to accurate information will be used to develop these
which the results should be believed (2). Even carefully guidelines and consequently improve treatment quality.
planned studies performed at a high level can still In the hierarchical research process, evidence-based
present ROB. For example, it would be unfair to judge guidelines, which help clinicians make decisions, are
62 Faggion

900 approaches. Because, in implant dentistry, animal

experiments are primarily focused on testing the effi-
600 cacy of new therapeutic approaches (4), studies with
low ROB are essential to guide the development of
clinical trials. Nevertheless, most ROB can be catego-
300
rized as unclear (4). An unclear ROB means that the
level of bias may be high risk, low risk, or in between.
0 This inability to determine ROB translates into a lack
of confidence in the research findings.
4

4
-7

-7

-8

-8

-9

-9

-0

-0

-1
70

75

80

85

90

95

00

05

10
Another important parameter to consider when
19

19

19

19

19

19

20

20

20
designing a research study is the calculation of sample
Fig. 1. Number of animal experiments in implant dentistry
published from 1970 to 2014. Publications were retrieved from size and power (5). Sample size contributes to the reli-
the PubMed database using the keyword ‘dental implants’ ability of precision measurements rather than to bias
and the filter ‘other animals’. The vertical axis represents the (2), and evidence suggests that authors of animal exper-
total number of publications, and the horizontal axis repre- iments in implant dentistry often do not calculate the
sents the year of publication. This information was retrieved sample size (4, 6). For experiments aiming to determine
on 11 December 2014. quantitative differences between groups or therapies, an
a priori calculation of sample size is recommended to
avoid type II errors. Type II errors occur when the
experiment has insufficient power to detect a difference
between different interventions. In other words, the
f) Clinical guidelines experiment failed to reject the null hypothesis (7).
e) Systematic reviews with meta- Improper sample sizes raise doubts about the validity
analysis of clinical trials
of results and may interfere with the development of
d) Clinical trials
further clinical research. Sample size is included in the
c) Systematic reviews with meta- ARRIVE (Animals in Research: Reporting of In Vivo
analysis of animal experiments Experiments) guidelines, a checklist for reporting ani-
mal experiments (8). Although this checklist was pub-
b) Animal experiments lished a few years ago, it does not appear to have
a) In vitro experiments improved the reporting and conducting of animal
research (9).
The lack of similarity between animal and human
research protocols is another methodological limitation
Fig. 2. The hierarchical research process in which multiple in implant dentistry. The data show that researchers
levels of research are interconnected. This process optimizes change their protocols significantly when they move
research output and generates reliable information for the from animal experiments to clinical trials. Conse-
development of evidence-based clinical guidelines.
quently, it becomes challenging (if not impossible) to
determine whether results were replicated (6). Replica-
tion is the process of confirming previous results in an
primarily based on systematic reviews (3). Ideally, sys- independent study and is one of the cornerstones of the
tematic reviews should generate a ‘summary of treat- scientific method (10).
ment effect’ meta-analysis to guide clinical treatments. Publication bias, a form of reporting bias, may also
The reliability of a meta-analysis depends upon the affect basic animal research. This type of bias occurs
degree of heterogeneity among the trials included in when studies presenting significant differences between
the systematic review. The higher the heterogeneity, the treatment arms are published more frequently than are
lower the confidence that the meta-analytic approach studies that report no significant differences (2). A
will generate reliable treatment-effect estimates. Simi- recently published survey evaluated the responses from
larly, the inclusion of trials with high or unclear ROB 454 researchers on questions regarding the magnitude,
in a meta-analysis increases deviations from the treat- drivers, consequences, and potential solutions for publi-
ment-effect estimates to more- or less-inflated meta-ana- cation bias (11). Their responses indicated that only
lytic estimates (2). Thus, trials with a low ROB approximately 50% of animal experiments performed
performed in animals or humans should be included in are published.
meta-analyses of animal studies and clinical trials,
respectively.
Strategies to reduce ROB in animal
research
Limitations in implant dentistry animal Researchers should adhere to procedures that consis-
research tently reduce the chances of ROB in animal research in
Animal research is used to develop models to study dis- order to improve efficiency in all phases of the research
ease and to test the safety and efficacy of therapeutic process. For example, animal experiments should be
 Animal research 63

planned and performed taking into account domains of The evidence presented in this article demonstrates
bias that could affect the internal validity of the trial. A that standards in animal research can be improved.
good reference would be the Cochrane approach for Three ethical concerns arise when animal research is not
evaluating ROB in randomized controlled trials (RCTs) performed with high standards: (i) the misuse of ani-
(2), but it should be adjusted to different research envi- mals; (ii) the waste of resources; and (iii) the possibility
ronments. For example, the SYRCLE0 s (SYstematic of testing dangerous or ineffective therapies in humans.
Review Centre for Laboratory animal Experimentation)
ROB tool (12) comprises 10 items based on the Coch-
rane ROB tool (2). The tool has been adjusted for par-
ticular bias that may play an important role on the
Conclusions
internal validity of an animal experiment (12). Further- The quality of animal research in implant dentistry is in
more, to improve the quality of reporting, researchers urgent need of improvement. Inaccurate findings from
should adhere to guidelines, such as ARRIVE (8). The basic animal research may directly interfere with clinical
development of this guideline was based on the CON- trials. Tools are available that can assist researchers to
SORT (Consolidated Standards of Reporting Trials) perform animal research with higher standards; research-
checklist for reporting RCTs (13); it comprises 20 items ers and journal editors should avail themselves of these
describing essential information that should be included tools and should adhere to them. Conducting animal
in any publication based on animal research (8). research with limited methodologies may raise ethical
To improve precision in animal experiments, appro- concerns regarding the usefulness of animal experiments.
priate sample sizes should be used. The challenges of Because animal research data were evaluated in only
increasing the number of animals in experiments are one specialty in dentistry, implantology, the applicabil-
higher costs and the logistics of maintaining large colo- ity of these conclusions to other fields of dentistry can-
nies of animals. One possible solution would be the not be taken for granted. However, because there tends
development of multicenter animal studies, involving to be homogeneity among research methods and
the collaboration of two or more centers with animal reporting in dentistry as a whole, similar findings are
facilities in order to obtain adequate sample sizes of likely to be obtained in other fields of dentistry. Not-
animals. Alternatively, small laboratories with limited withstanding, the performance and reporting of animal
vivarium space may collaborate with larger research experiments should be evaluated in other fields.
institutions that can accommodate large groups of ani-
Conflicts of interest – The author declares no conflicts of interest.
mals, or researchers may perform a large study through
a series of replications with small cohorts.
To minimize publication bias, editors of scientific jour-
nals could insist on submission of original animal References
research protocols along with manuscripts. Furthermore,
1. VON AULOCH S. Number of animals used for experimental
a public database for animal research protocols should purposes lower in the European Union. ALTEX 2014; 31: 1.
be developed, like those already created for human clini- 2. HIGGINS JPT, ALTMAN DG, STERNE JAC (editors). Chapter 8:
cal trials (14). The registration of animal experiment pro- assessing risk of bias in included studies. In: HIGGINS JPT,
tocols would probably reduce the incidence of outcome GREEN S, eds. Cochrane Handbook for Systematic Reviews of
reporting bias (when the report of outcomes is based on Interventions. Version 5.1.0 [updated March 2011]. The Coch-
rane Collaboration 2011. Available from [Link]-hand-
the direction and nature of results) (15). [Link].
3. GUYATT G, OXMAN AD, AKL EA, KUNZ R, VIST G, BROZEK
J, NORRIS S, FALCK-YTTER Y, GLASZIOU P, DEBEER H, JA-
ESCHKE R, RIND D, MEERPOHL J, DAHM P, SCHUNEMANN € HJ.
Animal research and ethics GRADE guidelines: 1. Introduction-GRADE evidence pro-
files and summary of findings tables. J Clin Epidemiol 2011;
There has been discussion over the years regarding the 64: 383–394.
necessity of animal research before clinical trials (16). 4. FAGGION CM Jr, GIANNAKOPOULOS NN, LISTL S. Risk of bias
In specialties such as implant dentistry, animal experi- of animal studies on regenerative procedures for periodontal
ments play an important role in determining the behav- and peri-implant bone defects – a systematic review. J Clin
Periodontol 2011; 38: 1154–1160.
ior of new therapeutic approaches. For example, 5. LACHIN JM. Introduction to sample size determination and
BRANEMARK’s discovery of osseointegration (i.e. the power analysis for clinical trials. Control Clin Trials 1981; 2:
direct contact between living bone and the dental 93–113.
implant surface at the light microscopic level) (17) was 6. FAGGION CM Jr, SCHMITER M, TU YK. Assessment of replica-
tion of research evidence from animals to humans in studies
made in the 1950s by the insertion of titanium implants
on peri-implantitis therapy. J Dent 2009; 37: 737–747.
into the bone of rabbits (18). It took several years of 7. GUYATT G, JAESCHKE R, HEDDLE N, COOK D, SHANNON H,
animal testing before dental implants started to be used WALTER S. Basic statistics for clinicians: 1. Hypothesis testing.
in humans (18). Accordingly, it would be unrealistic to CMAJ 1995; 152: 27–32.
think that the sound current knowledge on the clinical 8. KILKENNY C, BROWNE WJ, CUTHILL IC, EMERSON M, ALTMAN
DG. Improving bioscience research reporting: the ARRIVE
use of dental implants would exist without this guidelines for reporting animal research. PLoS Biol 2010; 8:
preclinical phase of animal research. It is nevertheless a e1000412.
sine qua non condition that these experiments are 9. BAKER D, LIDSTER K, SOTTOMAYOR A, AMOR S. Two years
adequately conducted. later: journals are not yet enforcing the ARRIVE guidelines
64 Faggion

on reporting standards for pre-clinical animal studies. PLoS 14. [Link]. [Link] Accessed on 18
Biol 2014; 12: e1001756. December 2014.
10. MOONESINGHE R, KHOURY MJ, JANSSENS AC. Most published 15. STERNE JAC, EGGER M, MOHER D (editors). Chapter 10:
research findings are false-but a little replication goes a long addressing reporting biases. In: HIGGINS JPT, GREEN S, eds.
way. PLoS Med 2007; 4: e28. Cochrane Handbook for Systematic Reviews of 10.23 Interven-
11. TER RIET G, KOREVAAR DA, LEENAARS M, STERK PJ, VAN tions. Version 5.1.0 [updated March 2011]. The Cochrane
NOORDEN CJ, BOUTER LM, LUTTER R, ELFERINK RP, HOOFT Collaboration 2011. Available from [Link]-hand-
L. Publication bias in laboratory animal research: a survey on [Link].
magnitude, drivers, consequences and potential solutions. 16. LOEB JM, HENDEE WR, SMITH SJ, SCHWARTZ MR. Human vs
PLoS One 2012; 7: e43404. animal rights. In defense of animal research. JAMA 1989;
12. HOOIJMANS CR, ROVERS MM, DE VRIES RB, LEENAARS M, 262: 2716–2720.
RITSKES-HOITINGA M, LANGENDAM MW. SYRCLE’s risk of 17. ALBREKTSSON T, BRANEMARK PI, HANSSON HA, LINDSTROM € J.
bias tool for animal studies. BMC Med Res Methodol 2014; Osseointegrated titanium implants. Requirements for ensuring
14: 43. a long-lasting, direct bone-to-implant anchorage in man. Acta
13. SCHULZ KF, ALTMAN DG, MOHER D; CONSORT Group. Orthop Scand 1981; 52: 155–170.
CONSORT 2010 statement: updated guideliness for 18. BRANEMARK R, BRANEMARK PI, RYDEVIK B, MYERS RR.
reporting parellel group randomised trials. BMJ 2010; 340: Osseointegration in skeletal reconstruction and rehabilitation:
c332. a review. J Rehabil Res Dev 2001; 38: 175–181.
This document is a scanned copy of a printed document. No warranty is given about the
accuracy of the copy. Users should refer to the original published version of the material.