Dental Literature Watch · Systematic review · Implants
Implants at 10-25 years: about 95% still in, about 85% "successful", and one in ten with peri-implantitis
TL;DR
- Pooled implant failure beyond 10 years was about 5% whatever the study design: 5.1% in prospective studies (95% CI 3.0-8.7%), 4.9% retrospective (3.3-7.1%), 4.8% in randomised trials (3.2-7.1%).
- That is roughly 95% survival at 10-25 years across 34 studies, 7,522 patients and 20,445 implants.
- Success averaged 85.1% across 14 studies, but success definitions varied substantially between them.
- Descriptive long-term means: probing depth 2.99 mm, marginal bone loss 1.30 mm, implant-level peri-implantitis 10.8% (range 3.0-39.9%).
- The authors' thesis: retention is a necessary but insufficient endpoint; high survival does not establish long-term biological health.
- Sensitivity analyses covered influential studies, risk of bias, implant-surface era and clustering assumptions.
The study
| Element | Detail |
|---|---|
| Design | Systematic review and meta-analysis; RCTs, prospective and retrospective cohorts with at least 10 years of follow-up; random-effects pooling by study design with clustering corrections and prediction intervals |
| Included | 34 studies, 7,522 patients, 20,445 implants |
| Failure: prospective | 5.1% (95% CI 3.0-8.7%) |
| Failure: retrospective | 4.9% (95% CI 3.3-7.1%) |
| Failure: RCTs | 4.8% (95% CI 3.2-7.1%) |
| Success | 85.1% average across 14 studies; definitions varied substantially |
| Biological outcomes | Mean probing depth 2.99 mm; mean marginal bone loss 1.30 mm; implant-level peri-implantitis 10.8% (range 3.0-39.9%) |
💎 Insight — what this means chairside
Three numbers for the consent conversation
About 95 in 100 implants are still in place after 10 years or more.
About 85 in 100 meet a success definition, and about 1 in 10 carries peri-implantitis.
About 95 in 100 implants are still in place after 10 years or more.
About 85 in 100 meet a success definition, and about 1 in 10 carries peri-implantitis.
Say "still in" and "healthy" as two different things
The ten-point gap between survival and success is where bone loss, deep pockets and bleeding live.
A patient who hears only 95% will feel misled when their retained implant needs peri-implantitis treatment.
The ten-point gap between survival and success is where bone loss, deep pockets and bleeding live.
A patient who hears only 95% will feel misled when their retained implant needs peri-implantitis treatment.
The peri-implantitis range is the maintenance argument
3% to nearly 40% between studies is not measurement noise; it tracks populations, maintenance and case definitions.
That spread is the evidence for recall intervals and hygiene access around implants.
3% to nearly 40% between studies is not measurement noise; it tracks populations, maintenance and case definitions.
That spread is the evidence for recall intervals and hygiene access around implants.
⚠️ The catch
Why this sits at 3 of 5
It updates numbers most clinicians already quote and does not change who gets an implant.
Its value is in making the survival-versus-success distinction explicit, which changes what is said at consent.
It updates numbers most clinicians already quote and does not change who gets an implant.
Its value is in making the survival-versus-success distinction explicit, which changes what is said at consent.
Success is whatever each study said it was
The 85.1% figure averages 14 studies with different success criteria, so it is an illustration, not a pooled estimate.
Probing depth, bone loss and peri-implantitis were summarised descriptively, not meta-analysed.
The 85.1% figure averages 14 studies with different success criteria, so it is an illustration, not a pooled estimate.
Probing depth, bone loss and peri-implantitis were summarised descriptively, not meta-analysed.
Survivorship bias in long cohorts
Patients still attending at 10-25 years are the ones whose implants did well enough to keep them attending.
Retrospective designs made up part of the pool, and the abstract does not give the dropout rates behind the survival figures.
Patients still attending at 10-25 years are the ones whose implants did well enough to keep them attending.
Retrospective designs made up part of the pool, and the abstract does not give the dropout rates behind the survival figures.
Abstract only
Prediction intervals, surface-era subgroup results and risk-of-bias ratings are in the paywalled full text.
Prediction intervals, surface-era subgroup results and risk-of-bias ratings are in the paywalled full text.
⚡ Bottom line
- Patient asks "will it last?" → about 95% are still in place at 10 years and beyond; about 85% meet a success definition; roughly 1 in 10 develops peri-implantitis.
- Writing the consent form → separate retention from biological health; quote both.
- Setting a recall interval → the 3-40% peri-implantitis range across studies is the case for structured maintenance.
- Comparing implant systems or surfaces → this review pools across them; it does not rank products.
🧪 Questions
Tap an answer — instant grade, deciding line, and the trap.
A 55-year-old asks what the chance is that a single posterior implant will still be in place after ten years. Which figure does this review support?
Pooled failure beyond 10 years was 4.8-5.1% across study designs, which is roughly 95% survival at 10-25 years. 85% is the average success rate, a stricter and definition-dependent measure, not retention. 75% and 65% are below every confidence interval reported. Trap: 85%, which is a real number from the review attached to the wrong question.
The same review reports roughly 95% survival but an average success of 85.1%. What best explains the gap?
Success definitions layer biological and technical criteria on top of retention, so an implant can be present and still count as a failure of success. Late losses count against both measures, not one. Randomised trials reported essentially the same failure (4.8%) as cohorts, so design does not explain the gap. Loading protocol is not how the review partitioned either outcome. Trap: blaming late failures, which is intuitive and wrong because survival already includes them.