The impact of diabetes on dental implant failure: a systematic review and meta-analysis

Abstract

The aim of this study was to investigate the hypothesis that there is no difference in implant failure rate or marginal bone loss between type 1 or 2 diabetes subjects and non-diabetic subjects. An electronic search was conducted, without restrictions on date or language, in the PubMed/MEDLINE, Cochrane Central Register of Controlled Trials, Web of Science, and EMBASE databases, and in the grey literature, through August 2015. The eligibility criteria included prospective and retrospective cohort studies and randomized controlled trials. The initial search resulted in 1093 titles from PubMed/MEDLINE, 164 from the Cochrane Central Register of Controlled Trials, 134 from Web of Science, 228 from EMBASE, and four from the grey literature. Following the search and selection process, 14 studies published between 2000 and 2015 were included in this systematic review. According to the risk of bias analysis, all studies were classified as high quality. The results of this systematic review suggest that the number of implant failures does not differ between diabetic and non-diabetic subjects. Additionally, the results of the comparison between type 1 and 2 diabetes subjects showed no difference in the number of failures. With regard to marginal bone loss, there was a statistically significant difference favouring non-diabetic subjects.

Despite dental implants showing a high long-term success rate, certain risk factors can compromise the biological process of osseointegration or negatively impact the maintenance of peri-implant health. Diabetes is one of these factors and is characterized by hyperglycaemia resulting from a deficiency in insulin secretion, its mechanism of action, or both. Diabetes is classified as type 1 (insulin-dependent), type 2, or gestational. Studies have demonstrated that the aetiology of diabetes consists of a combination of genetic and environmental factors (including viral infections, an inadequate diet, and a sedentary lifestyle).

The epidemiological prevalence of diabetes is high. This disease may affect approximately 11% of the American population, with 90–95% of these cases diagnosed with type 2 diabetes, which is the most frequent type observed in patients older than 40 years of age.

As a result of microvascular complications in patients with diabetes, there is a delay in the tissue healing process; this is due to the lower cell concentration at the surgical site and subsequent lower release of growth factors and cytokines, and reduced collagen synthesis. Furthermore, diabetic patients may have a reduced immune response, which increases the possibility of post-surgical infection.

In studies evaluating the success or survival of dental implants in diabetic volunteers, a higher rate of early-onset failures has been observed compared to late-onset failures. Chronic hyperglycaemia can affect the synthesis of osteoblasts and stimulate increased osteoclast function. In addition, the metabolism of calcium and potassium may become altered. As a result of these phenomena, there will be decreased bone formation during the healing phase, which would explain a higher rate of early failure, i.e., during osseointegration. For these reasons, diabetes is considered a relative contraindication during dental implant treatment.

On the other hand, diabetic patients who maintain control of their glycaemic index appear to have implant success and survival rates similar to those of systemically healthy individuals. Thus, the aim of this review was to investigate the hypothesis that there is no difference in implant failure rate or marginal bone loss between type 1 or 2 diabetes subjects and non-diabetic subjects.

Materials and methods

The methodology of this systematic review followed the recommendations of the Cochrane Handbook for Systematic Reviews of Interventions. In order to increase the quality and transparency of the study, the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) and AMSTAR (Assessment of Multiple Systematic Reviews) checklist guidelines were followed. The clinical questions were formulated and organized using the PICOS process.

Focused question

Is there a difference in the failure rate and marginal bone loss level of dental implants between type 1 or 2 diabetes subjects and non-diabetic subjects?

Clinical relevance

Knowing the risk factors is essential to the success of treatment with implants. Accordingly, this systematic review will provide data to ensure that the decision-making process and case planning for diabetic patients is based on scientific evidence.

Search strategy

An electronic search was conducted, without date or language restriction, in the PubMed/MEDLINE, Cochrane Central Register of Controlled Trials, Web of Science, and EMBASE databases through August 2015. In addition, a manual search was conducted in the following periodic journals: Journal of Periodontology , Journal of Clinical Periodontology , Journal of Periodontal Research , International Journal of Periodontics and Restorative Dentistry , Clinical Oral Implants Research , Clinical Implant Dentistry and Related Research , International Journal of Oral and Maxillofacial Implants , International Journal of Oral and Maxillofacial Surgery , Implant Dentistry , Journal of Dentistry , Journal of Prosthodontics , and Journal of Dental Research . A search of the so-called ‘grey literature’ in the Open-GRAY database, the ClinicalTrials.gov database ( www.clinicaltrials.gov ), and the references of the included studies (cross-referencing) was also conducted. The search strategy and PICOS framework can be seen in Table 1 .

Table 1
Systematic search strategy (PICOS strategy).
Search strategy
Population #1. (Partially edentulous[MeSH] OR edentulous[MeSH] OR edentulous maxilla OR edentulous mandible OR diabetic[MeSH] OR diabetes mellitus[MeSH] OR type 1 diabetes mellitus[MeSH] OR type 2 diabetes mellitus[MeSH] OR non-diabetic)
Intervention #2. (Dental implant[MeSH] OR dental implant surgery[MeSH] OR single implant[MeSH] OR multiple implant[MeSH])
Comparisons #3. (Diabetic type 1 vs. diabetic type 2 vs. non-diabetic)
Outcomes #4. (Cumulative survival rate[MeSH] OR survival OR dental implant survival OR dental implant failure OR failure OR marginal bone loss OR implant bone resorption OR dental implant bone loss)
Study design Prospective cohort studies, retrospective cohort studies, and randomized controlled trials
Search combination #1 AND #2 AND #3 AND #4
Database search
Language No restriction
Electronic databases PubMed/MEDLINE, Cochrane Central Register of Controlled Trials, Web of Science, and EMBASE

Selection criteria

This review searched for prospective and retrospective cohort studies and randomized controlled trials (RCTs) comparing implant failure rates and marginal bone loss between type 1 or 2 diabetes subjects and non-diabetic volunteers. This review considered implant failure as absolute implant loss. The exclusion criteria were animal studies, in vitro studies, clinical series, case reports, and reviews. Studies involving volunteers with other decompensated metabolic diseases or those with periodontal diseases without prior treatment were also excluded.

Screening process

The search and screening process was conducted by both authors/reviewers. The titles and abstracts were first analyzed. In the second stage, full-text articles were selected for careful reading and analysis against the eligibility criteria (inclusion/exclusion) for later data extraction. Disagreements between the reviewers were settled through detailed discussions. The concordance between the two reviewers for the search process was evaluated using Cohen’s kappa ( κ ) test. The authors of the studies included were contacted by e-mail to answer any questions, if necessary.

Risk of bias and quality assessment

The Newcastle–Ottawa scale (NOS) ( http://www.ohri.ca/programs/clinical_epidemiology/oxford.asp ) was used for the analysis of the quality of the non-randomized trials (prospective and retrospective cohort studies) included in this review. For the selection and outcome categories, the studies were awarded a star/point for each item. For the comparison category, two stars/points were awarded. The highest score that can be awarded to a study is nine stars/points. Studies that scored 6 stars or more were considered to be of high quality.

Data extraction

The following data were extracted from the selected studies (when available): authors, publication year, study design, follow-up period, number of subjects, sex, age of the subjects, diabetic condition, number of implants placed, number of failed implants, mean survival rates, implant characteristics, healing period, number of smokers, days of antibiotic prophylaxis, use of mouth rinse, marginal bone loss, and the authors’ conclusions.

Statistical analysis

Continuous and binary variables from the studies were subjected to meta-analysis when at least two of the studies assessed reported the same data type. For the binary outcomes (e.g., implant failure), the intervention effects estimated were expressed as a percentage risk ratio (RR) with a 95% confidence interval (CI). For the continuous outcomes (e.g., marginal bone loss), the mean and standard deviation were used to calculate the mean difference (MD) in millimetres with a 95% CI. The inverse-variance method was used for the random-effects model or the fixed-effects model. Heterogeneity was assessed using the χ 2 test and any impact on the meta-analysis was quantified via I 2 . Values of ≤25% were classified as low heterogeneity, values of up to 50% were classified as medium heterogeneity, and those above 70% were classified as high heterogeneity. When significant heterogeneity was found ( P < 0.10), the results of the random-effects model were validated. When low heterogeneity was verified, the fixed-effects model was considered. The level of statistical significance was set at P < 0.05.

The data were analyzed using Review Manager version 5.2.8 statistical software (The Nordic Cochrane Centre, The Cochrane Collaboration, Copenhagen, Denmark, 2014).

Publication bias was explored graphically using a funnel plot. An asymmetric funnel plot (studies falling outside the funnel) may indicate a possible publication bias.

Results

Literature search

The initial search resulted in 1093 titles from PubMed/MEDLINE, 164 from the Cochrane Central Register of Controlled Trials, 134 from Web of Science, 228 from EMBASE, and four from the grey literature. After the first assessment, 22 full-text articles were selected. Following careful reading, eight studies were excluded as they failed to conform to the eligibility criteria of this review ( Table 2 ). Thus, 14 studies published between 2000 and 2015 were included in this systematic review. The article search and selection process is shown in Fig. 1 .

Table 2
Excluded studies.
Reason for rejection Authors
Case report Balshi et al. (2007)
Duplicate study Oates et al. (2009)
Animal study Keller et al. (1999)
Did not state the type of diabetes and/or failures in each group Fiorellini et al. (2000)
van Steenberghe et al. (2002)
Moy et al. (2005)
Farzad et al. (2002)
Omran et al. (2015)

Fig. 1
Flow diagram (PRISMA format) of the screening and selection process.

The κ value for concordance between the two authors/reviewers for the potential articles to be included (titles and abstracts) was 0.97 and for the selected articles was 0.90, demonstrating excellent concordance.

Study characteristics

The characteristics of the selected studies are presented in Table 3 . Two prospective cohort studies, three retrospective cohort studies, and nine controlled clinical trials were included in this systematic review. The number of volunteers in the studies ranged from 10 to 663, and the volunteers were aged between 15 and 89 years. Eight hundred and two diabetic volunteers and 1532 non-diabetic volunteers were assessed, with the installation of 1551 and 6353 implants, respectively. Only one study assessed subjects with type 1 diabetes. All selected articles included only volunteers with diabetes that was under control at the time of surgery. Five articles did not report whether participants who smoked were included. The monitoring period in the studies ranged from 3 to 204 months, with an average of 45.1 months.

Table 3
Main characteristics of the studies selected.
Authors Study design
Follow-up (months)
No. of subjects
Number per group
Age range
Mean age
Sex
Diabetes type
Nature of diabetes
Implants placed/implants failed Mean survival rate (%) Implant system Implant size (diameter × length) (mm)
Morris et al. (2000) CCT
36
663
255 (D2)/408 (ND)
40–89
NR
NR
Type 2
Controlled
255/20 (D2)
2632/180 (ND)
92.2 (D2)
93.2 (ND)
Spectra system (NR × NR)
Olson et al. (2000) Prospective
60
89
89 (D2)
40–78
62.7
89 M
Type 2
Controlled
178/16 (D2) 91.0 (D2) Paragon Implant Company, Nobel Biocare, Interpore Corporation
(NR × 8, 10, 10.5, 11, 13, 15, 16, 18, 20)
Accursi (2000) Retrospective
204
45
15 (D2)/30 (ND)
15–83
53.9
16 M/29 F
Type 2
Controlled
59/4 (D2)
111/7 (ND)
93.2 (D2)
93.7 (ND)
Brånemark
(NR × NR)
Peled et al. (2003) CCT
60
41
41 (D2)
NR
NR
26 M/15 F
Type 2
Controlled
141/4 (D2) 97.2 (D2) Medical Implant System
(3.75 × 10 to 16)
Dowell et al. (2007) CCT
4
35
25 (D2)/10 (ND)
29–81
NR
NR
Type 2
Controlled
39/0 (D2)
11/0 (ND)
100 (D2)
100 (ND)
Straumann
(4.1 × 10, 12)
Alsaadi et al. (2008) Retrospective
24
412
9 (D1)/1 (D2)/402 (ND)
NR
NR
NR
Types 1 and 2
Controlled
33/0 (D1)
1/0 (D2)
1480/101 (ND)
100 (D1)
100 (D2)
93.2 (ND)
Nobel Biocare
(3.3, 3.75, 4, 5 × 10)
Tawil et al. (2008) CCT
42.4
90
45 (D2)/45 (ND)
29–85
62.1
57 M/33 F
Type 2
Controlled
255/6 (D2)
244/1 (ND)
97.6 (D2)
99.6 (ND)
Nobel Biocare
(NR × NR)
Loo et al. (2009) CCT
3
278
138 (D2)/140 (ND)
38–50
45.5
119 M/159 F
Type 2
Controlled
255/174 (D2)
346/48 (ND)
31.8 (D2)
86.1 (ND)
Straumann
(3.5 × NR)
Anner et al. (2010) Retrospective
114
475
49 (D2)/426 (ND)
NR
51.9
176 M/299 F
Type 2
Controlled
177/5 (D2)
1449/72 (ND)
97.2 (D2)
95.0 (ND)
NR
(NR × NR)
Turkyilmaz (2010) Prospective
12
10
10 (D2)
45–71
58
6 M/4 F
Type 2
Controlled
23/0 (D2) 100 (D2) Astra Tech
(4, 4.5 × 9, 11, 13, 15)
Erdogan et al. (2015) CCT
12
24
12 (D2)/12 (ND)
NR
51
12 M/12 F
Type 2
Controlled
22/0 (D2)
21/0 (ND)
100 (D2)
100 (ND)
Straumann
(4.1 × 10, 12)
Gómez-Moreno et al. (2015) CCT
36
67
46 (D2)/21 (ND)
59–64
59.5
33 M/34 F
Type 2
Controlled
46/0 (D2)
21/0 (ND)
100 (D2)
100 (ND)
Straumann
(3.3, 4.1 × 10 to 14)
Conte et al. (2015) CCT
12
54
35 (D2)/19 (ND)
37–70
55.3
25 M/29 F
Type 2
Controlled
35/0 (D2)
19/0 (ND)
100 (D2)
100 (ND)
S.I.N.
(NR × NR)
Ghiraldini et al. (2015) CCT
12
51
32 (D2)/19 (ND)
37–70
54.2
28 M/23 F
Type 2
Controlled
32/0 (D2)
19/0 (ND)
100 (D2)
100 (ND)
S.I.N.
(3.75 × 8.5 to 11.5)
Authors Healing period No. of smokers Antibiotics/mouth rinse (days) Marginal bone loss (mean ± SD) (mm) Author conclusions
Morris et al. (2000) NR NR NR NR The use of endosseous dental implants in type 2 diabetic patients involves a marginal risk to long-term implant survival
Olson et al. (2000) 4 months 34 NR/NR NR This study supports the use of dental implants in type 2 diabetic patients
Accursi (2000) NR NR NR/NR 0.25 ± 0.07 (D2)
0.06 ± 0.03 (ND)
The diabetic patients were no more likely to experience implant failure than the non-diabetic patients
Peled et al. (2003) 3 months NR 5/NR NR The clinical outcome of dental implant placement in a selected group of patients with well-controlled type 2 diabetes mellitus is encouraging
Dowell et al. (2007) 4 months 0 10 to D2 and 3 for ND/NR NR We found no evidence of diminished clinical success or significant early healing complications associated with implant therapy based on the glycaemic control levels of patients with type 2 diabetes mellitus
Alsaadi et al. (2008) NR 61 NR NR Systemic health factors do not seem to be prominent players in the aetiology of late implant loss
Tawil et al. (2008) NR 40 7/14 0.5 ± 0.71 (D2)
0.21 ± 0.3 (ND)
No statistically significant difference was found for patients or for implants for the advanced surgery cases or the conventional approach in diabetic patients compared to non-diabetic patients
Loo et al. (2009) 3 months NR NR/14 NR Implant failure in diabetics was significantly greater than that in non-diabetics when multiple adjoining implants were placed
Anner et al. (2010) NR 63 NR NR This study found no evidence of diminished clinical success or significant early healing complications associated with implant therapy in patients with controlled type 2 diabetes mellitus
Turkyilmaz (2010) 3 months NR 5/14 0.3 ± 0.2 (D2) This clinical report supports the use of dental implants in patients with well- or moderately well-controlled type 2 diabetes mellitus as a dental treatment modality
Erdogan et al. (2015) 4 months 0 NR/NR 2.96 ± 0.59 (D2)
2.86 ± 0.64 (ND)
The results of the present study suggest that type 2 diabetic patients may undergo staged guided bone regeneration procedures securely
Gómez-Moreno et al. (2015) 4 months 0 7/14 0.5 ± 0.14 (D2)
0.41 ± 0.18 (ND)
Implant therapies for diabetic patients can be predictable, providing these patients fall within controlled ranges of glycaemia over time
Conte et al. (2015) 4 months 0 Single dose 1 h before/7 NR The patient’s glycaemic status appears to modulate bone-related genes in a different manner
Ghiraldini et al. (2015) 4 months 0 Single dose 1 h before/7 NR Poor glycaemic control negatively modulated the bone factors during healing
CCT, controlled clinical trial; D1, type 1 diabetes; D2, type 2 diabetes; ND, non-diabetes; NR, not reported; M, male; F, female; SD, standard deviation.

The mean success rate among diabetic volunteers ranged from 31.8% to 100% and for non-diabetic volunteers ranged from 86.1% to 100%. The mean annual failure was 3.92% for diabetic subjects and 1.65% for non-diabetic subjects.

The healing period from implant installation to prosthetic loading ranged from 3 to 4 months. All implants used in the studies were surface-treated. Implants were installed in bone regeneration areas in two studies. A marginal bone loss assessment was conducted in five studies using peri-apical and/or panoramic radiographs, in which the implant platform in relation to the alveolar bone crest was used as the reference. Antibiotic medication was administered in seven studies, either before or after surgery. Six studies reported prescribing chlorhexidine mouthwash postoperatively.

Risk of bias and quality assessment

The quality analysis results for the studies included in the review are presented in Table 4 . No study scored fewer than 6 points.

Table 4
Risk of bias and quality assessment of the studies included.
Authors Selection Comparability Outcome Total 9/9
Representativeness of the exposed cohort Selection of external control Ascertainment of exposure Outcome of interest not present at start Comparability of cohorts on the basis of the design or analysis a Assessment of outcome Follow-up was long enough for outcome to occur b Adequacy of follow-up of cohorts
Morris et al. (2000) 0 ★ 0 0 6/9
Olson et al. (2000) 0 ★ 0 7/9
Accursi (2000) ★ 0 0 7/9
Peled et al. (2003) 0 ★ 0 0 6/9
Dowell et al. (2007) 0 ★ 0 0 6/9
Alsaadi et al. (2008) 0 ★ 0 0 6/9
Tawil et al. (2008) 0 ★ 0 0 6/9
Loo et al. (2009) ★★ 0 0 7/9
Anner et al. (2010) 0 ★ 0 7/9
Turkyilmaz (2010) 0 ★ 0 0 6/9
Erdogan et al. (2015) ★ 0 8/9
Gómez-Moreno et al. (2015) 0 ★★ 0 7/9
Conte et al. (2015) 0 ★★ 0 7/9
Ghiraldini et al. (2015) 0 ★★ 0 7/9
Only gold members can continue reading. Log In or Register to continue

Stay updated, free dental videos. Join our Telegram channel

Dec 15, 2017 | Posted by in Oral and Maxillofacial Surgery | Comments Off on The impact of diabetes on dental implant failure: a systematic review and meta-analysis

VIDEdental - Online dental courses

Get VIDEdental app for watching clinical videos