Journal of Clinical Periodontology

Quality of reporting of experimental research in implant dentistry. Critical aspects in design, outcome assessment and model validation

Authors

  • Fabio Vignoletti,

    1. Department of Periodontology, Faculty of Odontology, Complutense University of Madrid, Madrid, Spain
    Search for more papers by this author
  • Ingemar Abrahamsson

    Corresponding author
    1. Department of Periodontology, Institute of Odontology, The Sahlgrenska Academy, University of Gothenburg, Gothenburg, Sweden
    • Department of Periodontology, Faculty of Odontology, Complutense University of Madrid, Madrid, Spain
    Search for more papers by this author

  • Conflict of interest and source of funding statement

  • The authors declare that they have no conflict of interests. This review was funded by the Authors' Institutions. This workshop was financially supported by the European Federation of Periodontology and by unrestricted grants from Astra, Nobel Biocare and Straumann.

Address:

I. Abrahamsson

Department of Periodontology

Institute of Odontology

The Sahlgrenska Academy

University of Gothenburg

Gothenburg

Sweden

E-mail: ingemar.abrahamsson@odontologi.gu.se

Abstract

Objective

The aim was to assess the quality of reporting of experimental research in implant dentistry by a critical evaluation of study design, outcome assessments and model validation.

Material & Methods

An online search was performed using the MEDLINE. Experimental studies performed in both animals and humans were included. A’stratified random sample of the included studies was extracted and used for quantitative and qualitative analyses. Modified versions of the ARRIVE guidelines were used for quality assessment.

Results

A total of 982 papers were eligible and used for quantitative analyses. A’stratified random sample of 193 publications was extracted. The dog model was the most used experimental model whereas experimental studies on humans were few. Intra-oral experimental sites dominated in human, monkey, dog and mini-pig studies. Extra oral sites dominated in rabbit, rodent and goat/sheep studies. Studies on the pathogenesis and treatment of peri-implant diseases were few.

Conclusion

Different animal models, experimental protocols and methods of analysis have been used to address different areas of experimental research in implant dentistry. Standardized designs for investigations within this type of experimental research seem to be lacking. Furthermore, in many of these studies there were limitations in reporting on methodology and statistical methods.

Introduction

Treatment with osseointegrated dental implants have shown to be a successful therapy with good predictability for incorporation (97–98%) (Berglundh et al. 2002, Pjetursson et al. 2007) and a high survival rate after 10 years of function (approximately 89%) (Pjetursson et al. 2007). Therapy with dental implants offers the possibility of a significant improvement in function and well being for edentulous patients (Albrektsson et al. 1987). The further development and improvement of dental implants and treatment techniques requires testing in experimental models before the introduction in general clinical practice.

As of today, experimental research in implant dentistry has a strong focus on providing solutions for a predictable implant therapy also for groups of patients that previously were excluded from this treatment form, i.e. patients with a compromised general health or patients with compromised implant sites. Much interest has also been paid to both the study of biological complications at dental implants, i.e. mainly peri-implant mucositis and peri-implantitis, and the treatment of these complications. Hence, three different areas within the literature of experimental research in implant dentistry may be distinguished. These are (i) studies with focus on osseointegration/healing of peri-implant tissues, (ii) studies on healing after placement of implants in compromised patients or compromised sites and (iii) studies with focus on periimplantitis/mucositis and their treatment. Depending on the research objective the researcher may choose between a variety of different surgical models, experimental techniques, outcome variables and animal models.

Data from human studies are often considered to provide the highest level of scientific evidence. Due to the heterogeneity of human samples in terms of age, general health and different site related factors, a very high number of patients are usually required to identify any differences and to obtain a sample with an adequate statistical power. In many cases such experiments are impossible to perform for ethical reasons. Ethical considerations must be made also when planning experiments using animal models. Therefore, experimental studies must be carefully designed to allow for valid outcome assessments. That was also the reason, why Kilkenny et al. (2010) presented the ARRIVE guidelines (Animal Research: Reporting In Vivo Experiments) intended as a guide when preparing or reviewing manuscripts on animal research. The objective of the present systematic review was to assess the quality of reporting of experimental research in implant dentistry by a critical evaluation of study design, outcome assessments and model validation. The term experimental research is used for experiments where the researcher manipulates one variable, and controls/randomizes the other variables. Experimental research has a control group, the subjects have been randomly assigned between the groups, and the researcher generally only tests one effect at a time. For experimental research different animal models, including the human model can be used.

Material and Methods

Focused question

What is the quality of reporting of experimental research in implant dentistry, assessed using a modification of the ARRIVE guidelines?

Search strategy

A systematic search strategy, described in Fig. 1, was used. In the initial phase of the review a literature search was performed using the PubMed database of the US National Library of Medicine (MEDLINE via PubMed) as source in the search for studies meeting the inclusion criteria. The database was searched for studies published until June 30, 2011.

Figure 1.

Flowchart of the study selection procedure.

The following search terms were used for animal research:

“dental implant” OR “endosseous implant” OR “titanium implant”.

Limits: “Animals”, “English Word”

AND

“implant placement” OR “immediate implant placement” OR “early implant placement” OR “delayed implant placement” OR “integration” OR “osseointegration” OR “osteointegration” OR “healing” OR “bone healing” OR “hard tissue healing” OR “soft tissue healing” OR “alveolar ridge” OR “bone crest” OR “socket” OR “tooth socket” OR “extraction socket” OR “fresh extraction socket” OR “post-extraction socket” OR “bone regeneration” OR “guided bone regeneration” OR “bone defect” OR “bone dehiscence” OR “ridge alteration” OR “bone resorption” OR “bone modeling” OR “bone remodeling” OR “membrane” OR “graft” OR “xenogenic graft” OR “bone graft” OR “autograft” OR “allograft” OR “allogenic graft” OR “peri-implantitis” OR “mucositis” OR “treatment”

The following search terms were used for human experimental research:

“dental implant” OR “endosseous implant” OR “titanium implant”.

Limits: “Humans”, “English Word”

AND

“implant placement” OR “immediate implant placement” OR “early implant placement” OR “delayed implant placement” OR “integration” OR “osseointegration” OR “osteointegration” OR “healing” OR “bone healing” OR “hard tissue healing” OR “soft tissue healing” OR “alveolar ridge” OR “bone crest” OR “socket” OR “tooth socket” OR “extraction socket” OR “fresh extraction socket” OR “post-extraction socket” OR “bone regeneration” OR “guided bone regeneration” OR “bone defect” OR “bone dehiscence” OR “ridge alteration” OR “bone desorption” OR “bone modeling” OR “bone remodeling” OR “membrane” OR “graft” OR “xenogenic graft” OR “bone graft” OR “autograft” OR “allograft” OR “allogenic graft” OR “peri-implantitis” OR “mucositis” OR “treatment”

AND

“histology”

Limits: Humans, English, Dental journals.

Only English written human studies published into dental journals were searched.

Inclusion criteria

Publications in peer-reviewed journal written in the English language

Experimental studies on animals and humans

A research focus on oral implant therapy

Exclusion criteria

  • In vitro studies
  • Less than four animals or subjects
  • Studies with a research focus on orthodontic or endodontic implants
  • Studies with a research focus on medical implants
  • Case reports
  • Human experimental studies published in non-dental journals

The two reviewers (IA and FV), independently, screened all titles and abstracts of the publications retrieved from the electronic search against the inclusion and exclusion criteria. A Kappa analysis for the agreement between the two reviewers was performed. In a second step, for studies appearing to meet the inclusion criteria, but for which there were insufficient data in the title and abstract to make a clear decision, the full report was obtained and was again, independently, screened by the two reviewers to establish whether the study met the inclusion criteria or not. Any disagreement was resolved by discussion among the reviewers.

The papers accepted for inclusion in the review were, subsequently, divided into seven groups depending on animal (experimental) model, i.e.:

  • experiments in dogs
  • experiments in rabbits
  • experiments in monkeys
  • experiments in goats (or sheep)
  • experiments in pigs
  • experiments in rodents
  • experiments in humans

To have a reasonable number of papers to analyse, it was decided to extract a stratified random sample of 199 publications. Due to the limited number of human experimental studies accepted in the review, the sampling fraction was one in the human data stratum, i.e. all experimental studies in humans were included. For the studies using different animal models a proportional allocation of the sample was decided. The random selection was generated in SAS (Statistical Analysis Software, SAS Institute Inc., Cary, NC, USA).

Review methods

Using specially designed data extraction forms, the two reviewers (IA and FV) independently, extracted data regarding the quality of each paper in the final sample. Any disagreement was discussed and resolved.

The ARRIVE (Animal Research: Reporting In Vivo Experiments) guidelines (Kilkenny et al. 2010) are based on the CONSORT check list (Schultz et al. 2010) and comprise 20 items which have to be considered when reporting of experiments on animals. The guidelines were developed in consultation with scientists, statisticians, journal editors and research funders to enable comprehensive and transparent reporting of animal studies in any area of bioscience research (Kilkenny et al. 2010). To tailor the ARRIVE guidelines to the objectives of this systematic review, a modified version of the ARRIVE guidelines (Table 1) was used allowing assessments of the quality of reporting in animal studies. For the human experimental studies, a similar modification of the ARRIVE guidelines (Table 2) was used for the data extraction. In addition, using a standard protocol, the following data were collected from the selected human and animal studies:

  1. tissue integration at different types of implant surfaces/materials,
  2. integration in compromised patients/sites (including early or immediate loading protocols),
  3. pathogenesis or treatment of mucositis/periimplantitis.
Table 1. Check list for quality assessment of reporting of animal research – according to a modification of the ARRIVE guidelines (Kilkenny et al. 2010)
ItemScoreRecommended information
Title1

0. poor

1. adequate

2. good

Provides as accurate and concise a description of the content of the article as possible.
Abstract2

0. poor

1. adequate

2. good

Provides an accurate summary of the background, research objectives, including details of the species or strain of animal used, key methods, principal findings and conclusions of the study.
Introduction
Background3

0. poor

1. adequate

2. good

Includes sufficient scientific background to understand the motivation and context for the study, and explains the experimental approach and rationale.

Explains how and why the animal species and model being used can address the scientific objectives and, where appropriate, the study's relevance to human biology.

Objectives4

0. poor

1. adequate

Describes clearly the objectives of the study, or specific hypotheses being tested.
Methods
Ethical statement5

0. no

1. yes

Contains information about the ethical review permissions, relevant licenses and national or institutional guidelines for the care and use of animals, that cover the research.
Study design

6

a,b

0. no

1. yes

Provides brief details of the study design including:

a. The number of experimental and control groups.

b. Any steps taken to minimize the effects of subjective bias when allocating animals to treatment (e.g. randomization procedure) and when assessing results (e.g. if done, describe who was blinded and when).

Experimental procedures

7

a,b,c

0. no

1. yes

Provides precise details of all procedures carried out, for each experiment and each experimental group, including controls.

a. Surgical/clinical procedures

b. Histological processing and analysis

c. Test device/material description

Experimental animals8

0. no

1. yes

Provides details of the animals used, including species, strain, sex, developmental stage and weight.
Housing and husbandry9

0. no

1. yes

Provides details of housing and husbandry conditions
Sample size

10

a,b

0. no

1. yes

a. Specifies the total number of animals used in each experiment, and the number of animals in each experimental group.

b. Explains how the number of animals was arrived at. Provides details of any sample size calculation used.

Experimental outcomes11

0. no

1. yes

Defines clearly the primary and secondary experimental outcomes assessed
Statistical methods

12

a,b,c

0. no

1. yes

a. Provides details of the statistical methods used for each analysis.

b. Specifies the unit of analysis for each dataset

c. Describes methods used to assess whether the data met the assumptions of the statistical approach.

Results
Numbers analysed13

0. no

1. yes

The number of animals in each group included in each analysis is reported.

If any animals or data were not included in the analysis, an explanation is provided.

Outcomes and estimation14

0. no

1. yes

Reports the results for each analysis carried out, with a measure of precision (e.g. standard error or confidence interval).
Adverse events15

0. no

1. yes

Adverse events in each experimental group are reported. Describes any modifications to the experimental protocols made to reduce adverse events.
Discussion
Interpretation/scientific implications16

0. poor

1. adequate

2. good

Interprets the results, taking into account the study objectives and hypotheses, current theory and other relevant studies in the literature.

Comments on the study limitations including potential sources of bias, limitations of the animal model, and the imprecision associated with the results.

Generalisability/translation17

0. no

1. yes

Comment on whether, and how, the findings of this study are likely to translate to other species or systems, including any relevance to human biology.
Funding18

0. no

1. yes

Provides information of funding sources and any conflict of interest.
Table 2. Check list for quality assessment of reporting of experimental research on humans – according to a modification of the ARRIVE guidelines (Kilkenny et al. 2010)
ItemScoreRecommended information
Title1

0. poor

1. adequate

2. good

Provides as accurate and concise a description of the content of the article as possible.
Abstract2

0. poor

1. adequate

2. good

Provides an accurate summary of the background, research objectives, including details of the species or strain of animal used, key methods, principal findings and conclusions of the study.
Introduction
Background3

0. poor

1. adequate

2. good

Includes sufficient scientific background to understand the motivation and context for the study, and explains the experimental approach and rationale.

Explains how and why the model being used can address the scientific objectives.

Objectives4

0. poor

1. adequate

Describes clearly the objectives of the study, or specific hypotheses being tested.
Methods
Ethical statement5

0. no

1. yes

Contains information about the ethical review permissions, relevant licenses and national or institutional guidelines, that cover the research.
Study design

6

a,b

0. no

1. yes

Provides brief details of the study design including:

a. The number of experimental and control groups.

b. Any steps taken to minimize the effects of subjective bias when allocating to treatment (e.g. randomization procedure) and when assessing results (e.g. if done, describe who was blinded and when).

Experimental procedures

7

a,b,c

0. no

1. yes

Provides precise details of all procedures carried out, for each experiment and each experimental group, including controls.

a. Surgical/clinical procedures

b. Histological processing and analysis

c. Test device/material description

Subjects8

0. no

1. yes

Provides information about medical condition and (mean) age
Dental history9

0. no

1. yes

Provides details of dental history
Sample size

10

a,b

0. no

1. yes

a. Specifies the total number of subjects used in each experiment, and the number of subjects in each experimental group.

b. Explains how the number of subjects was arrived at. Provides details of any sample size calculation used.

Experimental outcomes11

0. no

1. yes

Defines clearly the primary and secondary experimental outcomes assessed
Statistical methods

12

a,b,c

0. no

1. yes

a. Provides details of the statistical methods used for each analysis.

b. Specifies the unit of analysis for each dataset

c. Describes methods used to assess whether the data met the assumptions of the statistical approach.

Results
Number analysed13

0. no

1. yes

The number of subjects/units in each group included in each analysis is reported.

If any subjects or data were not included in the analysis, an explanation is provided.

Outcomes and estimation14

0. no

1. yes

Reports the results for each analysis carried out, with a measure of precision (e.g. standard error or confidence interval).
Adverse events15

0. no

1. yes

Adverse events in each experimental group are reported. Describes any modifications to the experimental protocols made to reduce adverse events.
Discussion
Interpretation/scientific implications16

0. poor

1. adequate

2. good

Interprets the results, taking into account the study objectives and hypotheses, current theory and other relevant studies in the literature.

Comments on the study limitations including potential sources of bias, limitations of the model, and the imprecision associated with the results.

Generalisability/translation17

0. no

1. yes

Comment on whether, and how, the findings of this study are likely to translate to other subject or treatments, including the relevance to human biology.
Funding18

0. no

1. yes

Provides information of funding sources and any conflict of interest.

When making the qualitative assessments of the selected papers a score system of 0-1 (absent-present) or 0-1-2 (poor-adequate-good) (Tables 1, 2) was used to assess the quality of each item included in the modified ARRIVE checklist. The total sum of scores was calculated for each of the selected papers. The total scores could range between 0 (minimum) and 28 (maximum). Mean, (SD), Median, Minimum and Maximum scores were calculated for all selected papers. In addition, the results from the qualitative assessments were further stratified into different groups, i.e. type of experimental model, type of research focus and time-point of publication (until 1990, 1991–2000 and 2001–2011).

Results

Screening

The result of the screening procedure is described in Fig. 1. The electronic search on the PubMed database of the US National Library of Medicine resulted in 2132 retrieved papers. The manual screening of titles and abstracts of the retrieved papers resulted in 982 papers eligible for inclusion. Twenty studies were regarded as experimental studies on humans, whereas the remaining 962 studies were performed on animals. The Kappa analysis for the agreement between the two reviewers, in the manual screening, resulted in a Kappa value of 0.8345 (ASE 0.0809) and a 95% CI between 0.68 and 0.99.

Quantitative analyses

All papers were screened and grouped according to type of animal, research focus and time-point of publication. The results from the quantitative analyses are presented in Tables 3 and 4.

Table 3. The number of experimental studies on implant dentistry, meeting the inclusion criteria, grouped according to research model, research focus and intra- or extra- oral model
TopicDogsRabbitsMonkeysGoats/sheepMini-pigsRodentsHumansTotal
Healing/tissue integration199157414152669565
Healing/compro-mized sites1905016816649353
Study/treatment of mucositis/peri-implantitis49013000264
Total43820770496813020982
Intra-oral model41746914521920595
Extra oral model21203135161110387
Table 4. The number of experimental studies on implant dentistry, meeting the inclusion criteria, grouped according to research focus and publication period
Research focusHealing/tissue integrationHealing/compro-mized sitesStudy/treatment of mucositis/peri-implantitisTotal
–1990255030
1991–200014610029275
2001–201139424835677
Total56535364982

Animal models

0f all included publications, 438 papers used the canine model as experimental model of research, 207 papers used the rabbit model, 70 papers used the monkey model, 49 papers used the goat model, 68 papers used the mini-pig model and 130 papers used rodents as experimental animals (Table 3).

Type of research focus (Table 3)

Dog studies

A total of 199 papers investigated on the healing and tissue integration of implants. Thirteen (6.5%) of these papers used an extra-oral experimental model. A total of 190 papers investigated on the healing of implants into compromised sites/subjects; 8 (4.2%) of these papers utilized an extra-oral experimental model. Forty-nine papers evaluated the study/treatment of mucositis/periimplantitis. None of these studies utilized the extra-oral model.

Rabbit studies

A total of 157 papers investigated on the healing and tissue integration of implants. A total of 156 (99.4%) of these papers used an extra-oral experimental model. Fifty papers investigated on the healing of implants into compromised sites/subjects; 47 (94%) of these papers utilized an extra-oral experimental model. No papers evaluating the study/treatment of mucositis/periimplantitis were found.

Monkey studies

Forty-one papers investigated on the healing and tissue integration of implants. Sixteen papers investigated on the healing of implants into compromised sites/subjects and 13 papers evaluated the study/treatment of mucositis/periimplantitis. All studies but one, performed in the monkey model utilized intraoral sites.

Goat/sheep studies

Forty-one papers investigated on the healing and tissue integration of implants. 33 (80.5%) of these papers used an extra-oral experimental model. Eight papers investigated on the healing of implants into compromised sites/subjects; 2 (25%) of these papers utilized an extra-oral experimental model. No papers evaluating the study/treatment of mucositis/periimplantitis were found.

Mini-pig studies

Fifty-two papers investigated on the healing and tissue integration of implants. Fourteen (26.9%) of these papers used an extra-oral experimental model. Sixteen papers investigated on the healing of implants into compromised sites/subjects; 2 (12.5%) of these papers utilized an extra-oral experimental model. No papers evaluating the study/treatment of mucositis/periimplantitis were found.

Rodent studies

Sixty-six papers investigated on the healing and tissue integration of implants. Fifty-seven (86.4%) of these papers used an extra-oral experimental model. Sixty-four papers investigated on the healing of implants into compromised sites/subjects; 54 (84.4%) of these papers utilized an extra-oral experimental model. No papers evaluating the study/treatment of mucositis/periimplantitis were found.

Experimental research on humans

The search resulted in 187 retrieved papers. After an initial phase of title and abstract screening 20 potentially relevant articles were identified.

Type of research focus (Table 3)

Nine papers investigated the healing and tissue integration of implants. Nine papers investigated on the healing of implants into compromised sites/subjects; two papers evaluated the study/treatment of mucositis/periimplantitis. All of the papers with experimental research on humans utilized an intra-oral experimental model.

Time-point for publication

Thirty studies of all studies meeting the inclusion criteria (982) were published before 1991. A total of 275 studies were published during the years 1991 until 2000, while the majority of studies, i.e. 677, were published after the year 2000. The distribution between the three different research foci were similar in the two later age cohorts, whereas in the cohort before 1991 studies with a focus on healing/integration dominated and no studies on pathogenesis or treatment of peri-implant diseases were found. (Table 4)

Stratified random sample

The stratified random sample of all studies consisted of 199 publications. After reviewing the full text articles another six papers were excluded. Two papers were case reports, two papers were short communications, one paper was a review and one paper was excluded because of too low number of experimental animals (Fig. 1). The final sample is presented in Table 5a–g and described according to animal model and research focus in Table 6.

Table 5. The scores assessed for each item of the modified ARRIVE guidelines (Table 1 and 2) and seven additional descriptive ítems of the random selection of papers (193). Items 1–3 and 16 scored 0, 1 or 2 (poor, adequate or good). All other items scored 0 or 1 (yes or no). The data are presented for experimental studies in the (a) dog model, (b) rabbit model., (c) monkey model, (d) goat/sheep model, (e) mini-pig model, (f) rodent model and (g) human model. RF, research focus
Authors and titleCountryYearAnimalsFollow-up timeType of analysisRFExtra oral123456a6b7a7b7c8910a10b1112a12b12c131415161718Score
(a) dog studies
Abrahamsson et al. (2009)Sweden2009201-2-4-6-8-12 weeksHistology-Bio-mechanical test1 22211111111010111011121125
Abushahba et al. (2008)Ireland200843 monthsHistology2 21211111111110111100121023
Albouy et al. (2009)Sweden200966 monthsHistology3 21211111111010111001121022
Araujo et al. (2006)Sweden/Brazil200664-12 weeksHistology2 11211001111010100001021016
Assenza et al. (2003)Italy200346-12 monthsHistology1 22210001101010110001111017
Becker et al. (1991)Sweden/US1991318 weeksClinical/surgical variables-Histology2 21110111111010101010011118
Berglundh et al. (2005)Sweden2005610 monthsHistology-Rx1 21211101111010110001121121
Bonfante et al. (2011)US/Brazil201162-4 weeksHistology-Bio-mechanical test1111211101111010110101120120
Braceras et al. (2009)Spain2009NA3-6 monthsHistology1 11011101110000110001010113
Breine & Branemark (1980)Sweden198045-7 monthsHistology-Rx2 00110001111010000010101010
Broggini et al. (2003)US/Mexixo/Switzwerland200356 monthsHistology1 01111100010010111101021116
Caneva et al. (2010)Brazil/Italy201064 monthsHistology2 21211101111010110011121122
Casati et al. (2000)Brazil/US200061-3 monthsHistology2 22211111101010111001121123
Caudill & Meffert (1991)US199139 weeksClinical/surgical variables-Histology2 10011011111010000000101112
Chang et al. (2009)South Korea2009412 weeksHistology2 11111111111110111011111021
Cochran et al. (1998)US/Switzerland199863-6-12 monthsHistology1 22110111111010110101121021
Coelho et al. (2010)US/Brazil2010123-5 weeksBio-mechanical test1121211111111010110111110021
de Oliveira et al. (2009)Brazil2009712 weeks healing + 8 weeks loadingHistology1 11111111111110111111101122
de Sanctis et al. (2010)Spain, Italy201086 weeksHistology2 22110111110010100011011118
Ericsson & Lindhe (1993)Sweden1993512 monthsHistology-Rx1 21110101111010100001121118
Ericsson et al. (1996)Sweden199656 monthsHistology-Rx1 21210101111010111001121121
Fiorellini et al. (1998)US/Germany199844 monthsHistology2 21211111111010111011121124
Foley et al. (2010)US2010612 weeksHistology1 21211111111010111001121123
Ghanavati et al. (2006)US/Iran2006153 monthsHistology1 22211101110010110001021120
Gineste et al. (1999)France199963-6-12 monthsHistology1 21110101110010110001021016
Gotfredsen et al. (1994)Denmark19941012 weeksHistology2 21211101111110100011021121
Hermann et al. (1997)US/Switzerland199756 monthsHistology1 22211111111010111011121125
Hong et al.(1994)the Netherlands1995106 monthsHistology1 10110101110010110011010013
Jovanovic et al. (1993)NA199332-4.5 monthsHistology3 1100010101001000000000017
Jovanovic et al. (1995)NA199556 monthsClinical/surgical variables-Histology2 111111111110100000011111#
Kim et al. (2006)Korea2008610 weeksMicro-CT1 111111010110101100011211#
Kim et al. (2008)Korea200948-16 weeksHistology2 212111111110101100111211#
Kim et al. (2009)Korea2006120-3-8 weeksHistology-Bio-mechanical test11112101111110101100111211#
Kohal et al. (1998)US199866 monthsHistology-Bio-mechanical test2 222111111110101100111211#
Lai et al. (2009)China/Switzerland200962-4-8 weeksHistology2 212111111110101101111211#
Lang et al. (1994)US/Switzerland199454 monthsHistology1 122101011110101110111211#

Lee et al.

(2010a)

US/Sweden/Italy2010128 weeksHistology-Rx2 222111111111101111111211#
Lekholm et al. (1993)US/Sweden1993516 weeksClinical/surgical variables-Histology2 211101011100101110111110#
Li et al. (2007)China200781-3 monthsHistology-Bio-mechanical test1 011001101110100000011011#
Lima et al. (2003)Brazil/Switzerland2003416-24 weeksHistology2 112101011110101000011211#
Marin et al. (2008)Brazil/Italy/US200862-4 weeksHistology-Bio-mechanical test11222111011110101100111211#
Meraw et al. (2000)USA200053 monthsHistology2 111111111101101101111111#
Mizuno et al. (2008)Japan200843 monthsHistology2 111111111101101100111111#
Moon et al. (2008)South Korea2008516 weeksHistology1 121101111101101110111111#
Novaes et al. (2003)Brazil2003512 weeksHistology2 121111111111101110111111#
Pansegrau et al. (1998)US1998151-2-3 monthsHistology11112100011100101100010210#
Pilliar et al. (1991)Canada1991472 weeksHistology-Rx1 111101011111101111111111#
Polimeni et al. (2010)USA2010128 weeksHistology2 111111111111101111111111#
Polyzois et al. (2007)Ireland & Switzerland200744monthsHistology2 111111111101101111111111#
Pontes et al. (2008)Brazil2008690 daysClinical/surgical variables-Rx1 111111011111101111111111#
Ran et al. (2009)Republic of China200994-8-12 weeksHistology-Bio-mechanical test-Rx11 +  mandible111111111110101101111111#
Rasmusson et al. (2001)Sweden200164 monthsHistology-Bio-mechanical test1 111101111101101100111100#
Sagara et al. (1993)Japan199363 monthsHistology-Bio-mechanical test-Rx1 101101111111101100111100#
Schliephake et al. (2002)Germany2009101-3 monthsHistology1 111101011110101110111111#
Schliephake et al. (2003)Germany2002101-3 monthsHistology1 111101011110101110111111#
Schliephake et al. (2009)Germany200365 monthsHistology-Rx2 111101101100101100110110#
Schuler et al. (2010)USA2010816 weeksHistology2 111111111110101110111101#
Schultes & Gaggl (2001)Austria200188 monthsHistology2 101111011110100100110110#
Schwarz et al. (2008)Germany2008121-2-4-6-9-12 weeksHistology2 111111111110101100111111#
Shabahang et al. (2003)USA200357.5 monthsHistology2 111101111111101101111111#
Shigino et al. (2000)Japan20001214-21-30-90 daysHistology-Micro CT1 111111011110101101111111#
Sigurdsson et al. (1997)USA1997516 weeksHistology-Rx2 112111111111101110111111#
Simion et al. (2007)Italy/Denmark USA,Sweden200736 monthsHistology-Rx2 111111111100101000111000#
Steflik et al. (1989)USA1999363-6-12-18-24 monthsHistology-Rx1 111111001101100000100101#
Steflik et al. (1993)USA198981-2-3-5 monthsHistology1 011111011101101000010111#
Steflik et al. (1999)USA1993301 yearClinical variables1 101101011100101111011110#
Strnad et al. (2008)Czech republic200832- 5-12 weeksHistology-Bio-mechanical test11111111111110101111110101#
Sykaras et al. (2004)Greece, USA2004122-4-8-12 weeksHistology-Rx1 112111111111101111111211#
Tenenbaum et al. (2003)France200366 monthsHistology-Rx1 111111011100101110111101#
Vignoletti et al. (2009a)Spain, Italy, Sweden2009162h-1-2-4-8 weeksHistology2 112111111111101111111111#
Vignoletti et al. (2009b)Spain, Italy, Sweden2009161-2-4-8 weeksHistology2 111111011110101111111111#
Weiner et al. (2008)US200863-6 monthsHistology1 111101111100101010111110#
Weng et al. (2008)Germany, Brazil200886 monthsHistology1 111111110101111110111111#
Weng et al. (2010)Germany, Brazil201063 monthsHistology1 111111110101111110111111#
Wetzel et al. (1999)Switzerland, USA199976 monthsHistology3 111101011111101110111111#
Wikesjo et al. (2002)USA, Japan2002616 weeksHistology-Rx2 111111111111101110111101#
Yoo et al. (2008)South Korea200868 weeksmicro-CT1 111111011110101100111111#
Zanetta-Barbosa et al. (1993)Brazil/Sweden199350-24-48-72Laser doppler1 111101010110101010101110#
(b) Rabbit studies
Allegrini et al. (2006)Brazil/Germany200688 weeksHistology1101110101111110100000000112
Alvarez et al. (2010)Japan201094-12-23 weeksBio-mechanical test1112111101111010110001110017
Blanco et al. (2011)Spain2011182-4-8 weeksHistology-Bio-mechanical test1121111111111110111011121124
Carvalho et al. (2010)Brazil2010307-14 weeksBio-mechanical2112111111111110111101021123
Cho & Jung (2003)South Korea2003108 weeksBio--mechanical1121010101111010111001011117
Elias et al. (2008)Brazil/Us2008NA12 weeksBio-mechanical1111211100011000111001021117
Franke Stenport & Johansson (2003)Sweden200366 weeksBio-mechanical1111111101110010111011021119
Han et al. (1998)Sweden1998193 monthsHistology-Bio-mechanical1100010101111010111001010012
Hatley et al. (2001)US20012090 daysRx1 11011101111010110001000013
Ivanoff et al. (1996)Sweden1996186-12 weeksBio-mechanical1121110101111110110011111019
Jansen et al. (1991)the Netherlands199162-4-6-8-10-12 weeksHistology1110110101111100000000110011
Johansson et al. (1991)Sweden1998301-6-12 monthsHistology-Bio-mechanical1121110101111010111001010016
Johansson et al. (1998)Sweden199173 monthsHistology-Bio-mechanical1110010101111010111001000012
Johnsson et al. (2000)Sweden2000438 weeksHistology-Bio-mechanical2111011101110010111001011015
Jung et al. (2001)Korea20013612 weeksHistology-Bio-mechanical2110110101111110110001011116
Kerner et al. (2010)France2010244-12 weeksHistology-Micro CT1110110111111110110011111119
Larsson et al. (1996)Sweden1996241-3-6 weeksHistology1111110101111010110001010115
Lee et al. (2009)South Korea200961 monthHistology1111111111111110110001010017
Lee et al. (2010a,b)South Korea20101212 weeksHistology1111110111111110111001010118
Li et al. (2004)South Korea2004104 weeksHistology-Bio-mechanical1111110101011010000001010112
London et al. (2002)USA2002111-2-3-4-5-8 weeksHistology1111110111110010110101011117
Lugero et al. (2000)Brazil2000208 weeksHistology2111210101111010110101011017
Muller-Mai et al. (1989)Germany1989NA84-168 daysHistology-Bio-mechanical1110110101110000110001011113
Park et al. (2005)Korea2005106 weeksHistology-Bio-mechanical micro CT1111110111111110111111111122
Piattelli et al. (1995)Italy1995121-2-3-4-8-26 weeksHistology1111110101111010100011010115
Piattelli et al. (1996)Italy199682-3-4-8 weeksHistology1111110101111010100011111117
Pimenta et al. (1995)Portugal, USA1995NA8-15-30-60 daysHistology110001000111100000000101108
Rasmusson et al. (1997)Sweden1997128-24 weeksHistology-clinical variables2111111111111010110011110119
Ronold & Ellingsen (2002)Norway20024810 weeksHistology-Bio-mechanical1111111111111110110101011120
Rupprecht et al. (2002)Germany20024042-84-168 daysHistology-Bio-mechanical1111111101111010110011010016
Scarano et al. (2003)Italy2007204-8 weeksHistology-Bio-mechanical2110011101111010110101011015
Scarano et al. (2007)Italy200354 weeksHistology1110010001110010100001101111
Sennerby et al. (2005)Sweden2005126 weeksHistology-Bio-mechanical1111111111111110110011111020
Slaets et al. (2009)Belgium2009123-7-14-28-42 daysHistology1111111111111010110111111121
Sul et al. (2010)Sweden20101010 weeksBio-mechanical1111111101111010111011111120
Suzuki et al. (2010)Brazil, USA2010162-4 weeksHistology1112111111111010111111110122
Vidigal et al. (1999)Brazil1999512weeksHistology1110011101110000110101010012
Yildiz et al. (2010)Turkey2010368 weeksHistology-Bio-mechanical1111111111111110111011111122
(c) Monkey studies
Akagawa et al. (1993)Japan199399 monthsMicrobiology3 21010101011010100001001113
Boyne & Herford (1994)US1994512 weeksHistology111001010101001000001100008
Carr et al. (1997)US199746 monthsBio-mechanical1 22110101111010111001121121
Fritz et al. (1994)US1994183-6 monthsClinical/surgical variables-Rx1 11011101010110000001000111
Gotfredsen et al. (1990)Denmark/Switzerland1990621 weeksClinical/surgical variables-Microbiology1 20110101011110111001011117
Miyata et al. (2000)Japan200044 weeksHistology-Rx2 11110101111010100011101015
Miyata et al. (2002)Japan200248 weeksHistology-Rx2 11110101011010100011101014
Quaranta et al. (2008)Italy200869 monthsHistology-Rx1 11111101111110110011111120
Scarano et al. (2000)Italy/US200066 monthsHistology2 21110000111010000001001112
Schou et al. (2003)Denmark200386 monthsHistology-clinical variables-Rx3 11111111111010110111111121
Trejo et al. (2006)USA, Germany, Mexico, Switzerland200692monthsHistology-clinical variables3 11111111111010110111011120
Watzak et al. (2006)Austria, Germany2006918 monthsHistology3 11111111111110111111111022
(d) Goat/sheep studies
Buser et al. (1999)Switzerland/US199994-8-12 weeksBiomechanical test1 21211111111010111001021122
Franchi et al. (2005)Italy200531 h-14-90 daysHistology1101210101111010000000100112
Frisken et al. (2002)Australia/Canada2002121-4-8-12 weeksHistology1 22211101011010111011121123
Nikolidakis et al. (2006)the Netherlands200666 weeksHistology1111110111111110110111111020
Schouten et al. (2009)the Netherlands201046 weeksHistology-Biomechanical test1111111111111110110011110120
Schouten et al. (2010a)the Netherlands200993 monthsHistology1112111111111110111011111123
Schouten et al. (2010b)the Netherlands201046 weeksHistology1111110111101011100011111118
Vercaigne et al. (2000)the Netherlands, Belgium2000126-12 weeksHistology1111110110110010111111111119
(e) Mini-pig studies
Allegrini et al. (2008)Germany/Brazil2008870 daysHistology-Bio-mechanical Clinical variables1 12211101010010110000111117
Budd et al. (1991)NA199195 weeks-3-6 monthsHistology1110010111110010100010000010
Buser et al. (1991)US/Switzerland200462-4-8 weeksHistology1 11211101110011111001011119
Buser et al. (2004)Switzerland/US199163-6 weeksHistology1121210101110010110011121120
Carmagnola et al. (2009)Italy200943 days-1-2-3-4-5-6 weeksHistology1 21111101111010100000111117
Endres et al. (2008)Germany2008144-8 weeksHistology1111211111111110110001110019
Fenner et al. (2009)Greece/Germany200986 monthsHistology1 22111101101010110011121020
Fuerst et al. (2003)Austria, Spain200384-8 weeksHistology11 + mandible11111101111010111011111120
Merten et al. (2001)Germany200194-86 weeksHistology1111000101111010000010111012
Nkenke et al. (2005)Germany, Greece200596 monthsHistology1 11111101111010111011011119
Schlegel et al. (2003) 200392-4-8 weeksHistology1111111101111010111011011119
Schlegel et al. (2006)Germany200692-4-8 weeksHistology1111111101111010111011011119
Schliephake et al. (2010)Germany, Switzerland2010124-13 weeksHistology-Bio-mechanical1 11110101111010111011011118
Stadlinger et al. (2008)Germany, UK2008206 monthsHistology1 11111111111010110111011120
(f) Rodent studies
Atsuta et al. (2005)Japan2005104 weeksHistology1 21211101101110100000010116
Chehroudi et al. (1992)Canada1992NA1-2-3 weeksHistology1112110101111100111011021120
Chen et al. (1999)Japan1999207-14-28-56-84 daysHistology2121210101111010100011021119
Clokie & Warshawsky (1995)NA1995246-7-8-9-10-11-12 weeksHistology-Rx1121110101111010000000021115
De Smet et al. (2007)Belgium/the Netherlands2007394 weeksHistology-Bio-mechanical1111110101110010110011120016
De Smet et al. (2008)Belgium/the Netherlands200830NAHistology-Bio-mechanical1121110111110010111001121120
el-Montaser et al. (1999)UK1999103weeks-3 monthsHistology1111111101111010000011110116
Futami et al. (2000)Japan2000341-3-5-7-14-28 daysHistology1 22111101111100000000110015
Gabet et al. (2006)Israel/Switzerland/US2006508 weeksHistology-Bio-mechanical1112111111111110110011121123
Gao et al. (2009)China2009403 monthsHistology-Bio-mechanical micro CT2122210111111010110001021121
Gorustovich & Guglielmotti (2001)Argentina20012030 daysHistology2 22110101111110110001111120
Haga et al. (2009)Japan2009801-1.5-2-2.5-3-3.5-4-5-6-7-8-9-12 monthsHistology1 22211001111110100000121019
Kohavi et al. (1992)Germany1992326-14 daysHistology1121110101111110110001010117
McCracken et al. (2006)USA20061522-7-14-24 daysHistology2111111111111110110101111121
Morberg et al. (1997)Sweden1997344 monthsHistology1111110101111010110011111118
Narai & Nagahata (2003)Japan2003301 monthHistology-Bio-mechanical2111110101111010110101011016
Nociti et al. (2002)Brazil2002582 monthsHistology2111111101111110010101011118
Ogawa et al. (2011)Belgium, Japan2011951-4 weeksHistology1111111111111010110111110120
Okamatsu et al. (2007)Japan2007204 weeksHistology1111111101111010100001010115
Sakakura et al. (2006)Brazil2006388-12 weeksRx2111111101111110111111011121
Shirakura et al. (2003)Japan2003361-3-5-7-14-28 daysHistology1 11111111111110100001010117
Vandamme et al. (2011)France, Belgium2011402-7 daysHistology-Micro CT2111111101111110111101111121
Yamada et al. (2010)Japan2010NA16 weeksHistology1 00101001110000110101010111
Yamazaki et al. (1999)Japan1999607-168 daysHistology2111110111111110110001011017
(g) Human studies
Aimetti et al. (2008)Italy200846 months (after 6 months sinus lift)Histology2 11110111111010100011121018
Artzi et al. (2002)Israel20021012 monthsHistology2 21110001111010100011121017
Bosshardt et al. (2011)UK/Switzerland/China2011287-14-28-42 daysHistology1 21111101111010110011121121
Bullon et al. (2004)Spain/italy200415NAHistology3 22101100101110110001011117
Fontana et al. (1994)Italy199456 monthsHistology2 21211111111110100011121022
Goene et al. (2007)the Netherlands/Italy200794-8 weeksHistology1 21110111110010111011111019
Hallman et al. (2002)Sweden2002216-9 monthsHistology-clinical variables2 22211111110010110011021122
Lang et al. (2011)UK/Switzerland/China2011287-14-28-42 daysHistology1 21211111111110110011121124
Lazzara et al. (1999)US-Italy1999116 monthsHistology1 21210101110010110011121019
Lindgren et al. (2009)Sweden2009118 monthsHistology2 11211111111110110011121123
Lorenzoni et al. (1998)Austria1998424 weeksHistology2 11110101111010100011111016
Orsini et al. (2007)Italy, USA2007158 ± 1 weeksHistology1 11111111111010111111111122
Paolantonio et al. (2001)Italy20014812 monthsHistology-clinical variables-Rx2 11111111111010111111111122
Parma-Benfenati et al. (1999)Italy, Sweden1999612 monthsHistology2 11110101111010110011111017
Rocci et al. (2003)Italy, Sweden200355-9 months loadingHistology-Bio-mechanical1 11110101110010111011111017
Shibli et al. (2010)Brazil, Italy, USA2010102 monthsHistology1 11111101111110111011011019
Simion et al. (1996)Italy1996166 monthsHistology2 11111111110010100011111017
Trisi et al. (2003)Italy, USA2003112 monthsHistology1 11110101110010110011011116
Zitzmann et al. (2001)Switzerland, Sweden20011221 daysHistology3 12111101111010011011121020
Table 6. The number of experimental studies on implant dentistry, in the stratified random sample, grouped according to research model, research focus and intra- or extra- oral model
TopicDogsRabbitsMonkeysGoats/sheepMini-pigsRodentsHumansTotal
Healing/tissue integration42325814157123
Healing/compro-mized sites3363009960
Study/treatment of mucositis/peri-implantitis304000310
Total7838128142419193
Intra-oral model7111127619117
Extra oral model73716718076

Study design, outcome assessments and model validation

The vast majority of the dog, monkey and human studies used intra-oral research models. The mini-pig studies were evenly distributed between studies using intra- or extra-oral experimental sites, whereas the majority of rabbit, goat/sheep and rodent studies used extra oral experimental sites.

A wide heterogeneity was found regarding (i) the number of animals included in each study, (ii) the number of test groups (different healing times or treatments) and (iii) the number and type of outcome variables chosen for evaluation. A wide heterogeneity was also observed regarding histological processing such as different methods for tissue preservation, embedding and cutting methods and also with regard to the type of microscopical technique, (e.g. light microscopy, fluorescence microscopy or electron microscopy).

Although the main outcome variables reported in the studies included in the stratified random sample were different types of qualitative and quantitative histological assessments, clinical, surgical,radiographical, microbiological and bio-mechanical assessments were also made.

Histological assessments included mainly histomorphometric measurements of bone to implant contact, bone area, bone volume, marginal bone level, soft tissues dimensions, and tissue composition. Bio-mechanical assessments included removal torque test, push in and push out tests and resonance frequency analysis of implant stability. Radiographical assessments included conventional radiography or micro CT to evaluate bone loss or bone to implant contact.

Histology was a common method of analysis in all animal models, whereas microbiological techniques were used mainly in monkeys and bio-mechanical tests were most common in rabbit studies.

Dog studies

A total of 78 dog studies were included in the stratified random sample. Seventy-one studies used the intra-oral research model to investigate on (i) healing/tissue integration (35 studies), (ii) healing of compromised sites/subjects (33 studies) and (iii) study/treatment of mucositis/peri-implantitis. Seven studies used the extra oral model to investigate on healing/tissue integration. The number of animals included in each dog study varied and up to 36 animals were included. In 58 of 78 studies ≤ 8 animals were included. Most of the studies included a number of treatment groups with different healing times that varied from 0 to 12 months. It was therefore, difficult to identify the real n for the statistical analysis. In ambition to test for many different variables in each experiment a general trend was that the n was low. Histological methods were used in 72 of 78 studies. Other methods used for assessing the outcome variables were micro-CT (3 studies), conventional radiographs (16 studies), biomechanical testing (10 studies) surgical and clinical examinations (6 studies) and laser Doppler (1 study).

Rabbit studies

A total of 38 rabbit studies were included in the stratified random sample. Thirty-seven studies used the extra oral research model to investigate (i) healing/tissue integration (31 studies), (ii) healing of compromised sites/subjects (six studies), whereas no studies were encountered on research focus 3. One study used the intra-oral experimental model to investigate on healing/tissue integration. The number of animals included in the rabbit studies varied and up to 48 animals were included. Most of the studies included a number of treatment groups with different healing times that ranged between 1 week and 12 months. Histological methods were used in 30 of 38 studies. Other methods used for assessing the outcome variables were micro-CT (2 studies), conventional radiographs (1 study), biomechanical testing (21 studies) and clinical examinations (1 study).

Monkey studies

Twelve monkey studies were included in the stratified random sample. Eleven studies used the intra-oral research model to investigate on (i) healing tissue/integration (four studies), (ii) healing of compromised sites/subjects (three studies) and (iii) study/treatment of mucositis/peri-implantitis (four studies). One study used the extra oral experimental model to investigate on healing/tissue integration. The number of animals included varied and up to 18 animals were included. Healing times ranged between 4 weeks and 18 months. Histological methods were used in 8 of 12 studies. Microbiological samples of submucosal plaque for phase contrast microscopic analysis were obtained in two studies. Other methods used for assessing the outcome variables were conventional radiographs (five studies), biomechanical testing (one study) and clinical/surgical examinations (four studies).

Goat/sheep studies

Eight goat/sheep studies were included in the stratified random sample. All but two studies used the extra oral research model to investigate on healing and tissue integration. The number of animals included varied between 3 and 12 animals. Healing times ranged between 1 h and 3 months. Histological methods were used in all studies and biomechanical testing was included in two studies. No other methods were used in this animal model.

Minipig studies

A total of 14 minipig studies were included in the stratified random sample. Fifty percentage of the studies used the extra oral research model to investigate on healing and tissue integration. The number of animals included varied between 3 and 12 animals. Healing times ranged between 1 h and 3 months. Histological methods were used in all studies and biomechanical tests were included in two studies. No other methods were used for evaluations in this animal model.

Rodent studies

A total of 24 studies were included in the stratified random sample. Eighteen studies used the extra oral research model to investigate on (i) healing/tissue integration (10 studies), (ii) healing of compromised subjects (eight studies) whereas no study on study/treatment of mucositis/peri-implantitis was encountered. Six studies used the intra-oral model to investigate on (i) healing/tissue integration (five studies) and (ii) healing of compromised subjects (one study). The number of animals included in the rodent studies varied and up to 152 animals were included. Healing times varied between 1 week and 12 months. Histological methods were used in all the presented studies. Other methods used for assessing the outcome variables were biomechanical testing (five studies), micro-CT (two studies) and conventional radiographs (two studies).

Human studies

A total of 19 studies were included in the stratified random sample. All studies used the intra-oral research model to investigate on (i) healing/tissue integration (eight studies), (ii) healing of compromised sites/subjects (nine studies) and (iii) study on study/treatment of mucositis/peri-implantitis (two studies). The number of subjects included in the human experiments varied between 4 and 48 included subjects. Follow-up times ranged between 7 days and 12 months. Histological methods were used in all the presented studies. Other methods used for assessing the outcome variables were biomechanical testing (one study), conventional radiographs (one study) and measurement of clinical variables (two studies).

Qualitative assessment

The results from the qualitative assessments of the stratified sample (193) are reported in Table 7 and Fig. 2. The mean score achieved by all the selected papers was 18.33. The score for the different animal models ranged between 16.08 for the experimental studies in monkeys and 19.35 for the experimental studies in dogs. The score depending on the grouping of research focus varied between 17.93 and 19.40 for the Focus 1 (Healing) and Focus 2 (Healing of compromised sites/subjects) studies, respectively. The studies published before 1991 reached the lowest mean score i.e. 13.75, whereas the studies published between 1991 and 2000 and between 2001 and 2011 reached scores amounting to 16.9 and 19.2 respectively.

Figure 2.

Histograms presenting the frequency distribution (%) of the scores assessed for each item of the modified ARRIVE guidelines (Tables 2 and 3). Items 1–3 and 16 scored 0, 1 or 2 (poor, adequate or good). All other items scored 0 or 1 (yes or no). The data are presented for all included studies.

Table 7. Results from the qualitative assessments of the stratified sample of selected papers (193) divided and grouped according to type of animal model, research focus and publication period. Scores according to the modified ARRIVE guidelines (Tables 2, 3)
GroupMeanMedianSDMinMax
All studies18.33193.75727
Before 199113.75142.981017
1991–200016.9174.13725
2001–201119.2203.211127
Dog19.35203.78727
Rabbit16.6173.8824
Monkey16.08174.69822
Goat19.14203.571223
Minipig17.78193.061020
Rodents18182.751123
Humans19.2192.581624
Extraoral17.52183.70824
Intraoral18.91193.66727
Focus 117.93193.68825
Focus 219.4203.561027
Focus 318205.24722

The frequency distribution of the scores, item by item, according to the modification of the ARRIVE guidelines (Tables 2, 3) was assessed and is presented for all studies (Fig. 2) and for the different subgroups, separately (Table 8). Four items of the checklist were scored 0, 1 or 2 (poor, adequate or good) whereas all other items were scored 0 or 1 (yes or no).

Table 8. Table presenting the frequency distribution of the scores assessed for each item of the modified ARRIVE guidelines (Tables 2 and 3). Items 1–3 and 16 scored 0, 1 or 2 (poor, adequate or good). All other items scored 0 or 1 (yes or no). The data are presented for all included studies and for different study subgroups and research foci (RF)
ArriveScoreDogsRabbitsMonkeysGoats/sheepMinipigsRodentsHumansAll studies (%)RF 1RF 2RF 3
1043010109 (4.66)810
1493185111411129 (66.8)82407
22544239855 (28.4)33193
206102011020 (10.3)1550
1592596111716143 (74)89468
21331226330 (15.5)1992
30393020017 (8.8)1232
147279582014130 (67.3)85387
22820344546 (23.3)26191
4020001104 (2)301
17638128132319189 (98)120609
50282074411882 (42.4)58213
1501854101311111 (57.6)65397
6a0521002111 (5.6)650
17336118142218182 (94.4)1175510
6b0352592111710109 (56.4)79255
143133637984 (43.6)44355
7a041110018 (4.1)521
17437117142418185 (95.9)118589
7b0625110015 (7.7)1212
1723677132419178 (92.3)111598
7c020011116 (3.1)411
17638127132318187 (96.9)119599
802362143645 (23.3)3492
15532107102113148 (76.7)89518
90532581131114125 (64.7)85355
1251347113568 (35.3)38255
10a01500143023 (11.9)9122
1773312802119170 (98.1)114488
10b07638127132419189 (98)119602
120011004 (2)400
110733114019 (9.8)1261
1713597132019174 (90.2)111249
12a01476238545 (23.3)27162
1643166111614148 (76.7)96448
12b041249482013119 (61.6)73397
137143464674 (38.4)5013
12c0572992131717144 (74.6)99405
12193617249 (25.4)24205
130242452416176 (39.3)53194
154147610818117 (60.7)70416
140720144018 (9.3)1341
17136127102019175 (90.7)110559
150162461713471 (36.7)47195
162146771115122 (63.3)76415
160636110017 (8.8)1052
1433055111610120 (62.1)80355
22951228956 (29.1)33203
17013162229044 (22.7)3842
16522106121519149 (77.3)85568
180201541451160 (31)36213
158238710198133 (69)87397

For the ARRIVE items 5, 6b, 9, 10b, 12b and 12c, more than 40% of the studies scored 0. Hence, 42.7% of the studies did not contain any information on ethical considerations, 56.25% of the studies did not present information regarding randomization or blinding procedures, 66.66% of the studies did not report any information of housing and husbandry conditions (animal studies) or provided details of the dental history of the patients (human studies). 97.91% of the studies failed to report on any sample size calculation. 61.97% of the studies did not specify the unit of analysis used for the statistical assessment and 76.56% of the studies did not provide details of or discuss the validation of the statistical method used. No major differences regarding the score distribution of these items were observed between the different experimental models or research foci (Table 8).

Discussion

In this systematic review the quality of reporting of experimental research was investigated. In addition, aspects on study design, outcome assessments and model validation are provided.

Due to the huge amount of experimental studies meeting our inclusion criteria, found in the systematic search, using the PubMed database of the US National Library of Medicine (MEDLINE via Pubmed), a stratified random representative sample was extracted and used for the analysis. A quantitative as well as a qualitative analysis was performed.

The quantitative analysis was based on 982 publications reporting of experimental research within the field of implant dentistry. The data extracted from this pool of studies revealed that experiments on the study and treatment of mucositis/peri-implantitis were few. These studies were performed in dogs (49 studies), monkeys (13 studies) or humans (2 studies). The choice of animal model probably reflects a limitation for the other animal models for this type of research. Studies with a research focus on healing in compromised subjects or sites were performed in all animal models included in this review. The majority of these were dog studies, manly focused on healing in compromised sites, while many of the rabbit and rodent studies focused on healing in systemically compromised subjects. For the study of healing and tissue integration of implants, all animal models included in this review have been used. The proportion of studies, within each animal model used for this research focus, however, varies. 45.4% of the dog studies, 75.8% of the rabbit studies, 58.5% of the monkey studies, 60.2% of the goat studies, 76.4% of the minipig studies and 50.7% of the rodent studies were used for the study of healing and tissue integration of implants. Furthermore, it has to be taken into consideration that most of rabbit, rodent and goat studies used extra oral surgical models. Common outcome assessments associated with the extra oral model, were, beside histology, different methods for mechanical testing of implant retention or stability, such as removal torque tests, push out or push in tests. Anatomical and morphological differences between extra and intra-oral models and the type of analysis related to these models makes the interpretation of the results, difficult. The extrapolation of such results into clinical settings are, therefore, often not possible or should be interpreted with caution.

The qualitative analysis was based on data extraction from 193 publications reporting of experimental research within the field of implant dentistry. The data extracted were chosen mainly according to the ARRIVE guidelines. In our systematic review we selected 24 items from the ARRIVE guidelines and added another seven items for our qualitative analysis. It should however be kept in mind that the items from the ARRIVE guidelines are not intended as quality descriptors of a manuscript. Instead the intention is to use them as a checklist or guideline when reporting from experimental research. For our review, we modified the original ARRIVE guidelines (Kilkenny et al. 2010) to be more suited for quality assessments. Despite the modification and depending on the type of study that is analysed, all items are not always appropriate to include and they may not have an equal quality value compared to other items. One example of this may be that when assessing the items regarding statistics, a frequent finding was that although the n number was reported, it was not properly calculated or it was very low due to a study design with many different experimental groups.

Although a trend towards higher mean ARRIVE scores was observed for more recent studies, the results from the qualitative analysis revealed that the main shortcomings when reporting of experimental research in implant dentistry were related to basic aspects of methodology. Limited information was provided on randomization and blinding procedures, as well as statistical methods. This was a general finding, which was observed in all studies of the stratified random sample as well as within the different animal models and research foci. Indeed, between 56% and 98% of the studies did not present information regarding randomization or blinding procedures, sample size calculations, did not specify the unit of analysis used for the statistical assessment or provide details of or discuss the validation of the statistical method used. The lack of such critical information needs to be considered when analyzing data from experimental research and points out critical issues that must be addressed in the future. This is in agreement with a recently published systematic review (Faggion et al. 2011). In this review the risk of bias of animal studies on regenerative procedures was analysed and it was concluded that methodological aspects of animal experiments in periodontology and implantology should be improved in order to reduce such risks.

Data from human studies are often considered as the highest level of scientific evidence. Despite this general opinion, human experimental studies in implant dentistry are few. The reasons for this may be many; the heterogeneity of human samples in terms of age, general health and different site related factors often make comparisons, conclusions and reaching adequate statistical power difficult unless a very high number of subjects are recruited. In many cases such experiments are also difficult to perform for ethical reasons since histological samples are often required.

The main findings in this systematic review were:

  • The dog model was the commonly most used experimental model (44.6%) followed by the rabbit model (21.1%).
  • Human experimental studies were few (1.9%).
  • Although intra-oral experimental sites dominated in human (100%), monkey (98.6%), dog (95.2%) and mini-pig (76.5%) studies, extra oral sites dominated in rabbit (98.1%), rodent (85.4%) and goat/sheep (71.4%) studies.
  • Studies on healing and tissue integration of dental implants was the most investigated research focus, representing 57.5% of all experimental research in implant dentistry. This focus was followed by research on healing of compromised sites or systematically compromised subjects (35.9%).
  • Experimental studies on the pathogenesis and treatment of peri-implant diseases were few and made up 6.4% of all experimental research within implant dentistry.
  • All animal models included in this review were used for research focus 1 and 2, whereas only dogs, monkeys and humans were used for research focus 3.
  • The dog was the most utilized animal model for investigations on healing of compromised sites, whereas rodents were mainly used in studies on healing of systematically compromised subjects.
  • Both intra- and extra- oral surgical models have been used for research on healing and tissue integration.
  • Intra-oral surgical models dominated in experiments on the healing of compromised sites and the study or treatment of mucositis or peri-implantitis, whereas extra-oral models dominated in experiments on healing of systematically compromised subjects.
  • The mean score in the qualitative analysis of the stratified sample (193) achieved by all the selected papers was 18.33.
  • The score for the different animal models ranged between 16.1 for the experimental studies in monkeys and 19.4 for the experimental studies in dogs.
  • Focus 2 studies (Healing in compromised sites/subjects) reached the highest mean score (19.4), whereas Focus 1 studies (Healing) had the lowest mean score (17.9).
  • Studies published before 1991 reached the lowest mean score i.e. 13.8, while the studies published between 1991 and 2000 and between 2001 and 2011 reached scores amounting to 16.9 and 19.2, respectively.
  • Between 56% and 98% of the studies did not present information regarding randomization or blinding procedures, or provided sufficient information about the statistical methods used.

In conclusion, a wide range of animal models, experimental protocols (intraoral/extra-oral) as well as methods of analysis has been used to address different areas of experimental research in implant dentistry. In a high number of these studies there were limitations in reporting on methodology, mainly regarding randomization/blinding procedures as well as statistical methods. Hence, results and conclusions should be interpreted in relation with the limitations inherent in the methodology and specific animal model used.

Ancillary