Every clinic that places implants publishes a survival rate. Ninety-eight percent is common. Ninety-nine point two is not unusual in a brochure. These numbers are not necessarily false. They are almost certainly not comparable to each other, not adjusted for the populations they came from, and almost certainly drawn from short follow-up in selected patients at a single centre.
The corrective to this is registry data. A national or regional registry captures implants placed across a full population — by specialists and generalists, in straightforward and complex cases, with compliant and non-compliant patients, with brands sold under promotional arrangements and brands chosen on clinical grounds. The registry does not have a marketing department. Its outcome numbers include the failures the single-clinic study left in a desk drawer.
This week I am reading a Scandinavian implant registry analysis — specifically the 2016 Derks study from the Swedish national cohort [1], cross-referenced against the Pjetursson 2012 systematic review of implant-supported fixed prostheses at five and ten years [2]. Neither paper tells you everything. Together they tell you more than any clinic brochure has ever told you.
Disclosure. Dr. Maloney has no commercial relationship with any registry, implant manufacturer, or clinic referenced in this piece. The publication’s standing disclosures are documented at /disclosures/. [Last reviewed: 2026-05-10.]
Patients
The Derks 2016 cohort was drawn from the Swedish Social Insurance Agency register — the national body that reimburses implant treatment through the Swedish public dental health system [1]. The study identified all patients who received implant-supported prostheses in Sweden between 2003 and 2004, then followed them to a clinical examination nine years later.
2,765 patients were eligible. 596 were randomly selected for a clinical follow-up examination approximately nine years after implant placement. Among those examined, 827 implants across 596 patients were assessed.
The population was:
- Mean age at placement: mid-fifties (consistent with the Swedish population of implant recipients at that time).
- Mix of single-tooth, short-span bridge, and full-arch reconstructions.
- Implants placed in both specialist and general dental practice settings — the Swedish system does not confine implant placement to specialists.
- Multiple implant systems represented, reflecting the commercial market in Sweden in 2003–2004 rather than any one brand’s clinical programme.
- High follow-up compliance relative to international norms, consistent with a publicly funded reimbursement system where patients have a standing relationship with the treating clinician.
What this is not: a sample of patients who sought implants privately, in a high-volume cosmetic clinic, or abroad. The Swedish Social Insurance Agency population skews toward patients whose treatment was planned within a publicly funded framework, with referral pathways and follow-up built into the system.
That qualification will matter substantially by the time we reach the external validity section.
Procedures and endpoints
Primary endpoint: implant survival. The Pjetursson 2012 systematic review — which pooled data from 87 studies covering 5,640 implants in 2,625 patients at five years, and 1,149 implants in 485 patients at ten years — reported estimated 5-year survival for implant-supported fixed dental prostheses of 95.6% [2]. Ten-year survival was estimated at 93.1% [2]. These numbers refer to the prosthesis remaining in function; they do not mean the implant itself was lost in every failure event.
For implant fixture survival specifically (the titanium fixture remaining osseointegrated and in situ), five-year rates across the implant literature cluster above 95%, with most registry-calibre analyses landing between 95% and 98%. The Pjetursson pooled estimate of 95.6% at five years for implant-supported prostheses is, if anything, conservative relative to fixture-level data — some prosthetic failures involve surviving fixtures carrying failed superstructures.
Osseointegration — the direct structural and functional connection between the titanium fixture and living bone — is the biological process that underpins all of these numbers [Wikipedia: Osseointegration]. When it fails in the first weeks post-placement, that is primary non-osseointegration. When the implant is lost after successful integration, the causes are different: peri-implantitis, biomechanical overload, prosthetic failure. The registry data captures both.
Secondary endpoint: peri-implantitis incidence — and the two numbers you need to hold together.
This is where the Derks 2016 data does something most clinic marketing never does: it names both numbers.
At nine-year follow-up, 45% of patients in the examined cohort showed evidence of peri-implantitis by the study’s definition (bleeding on probing plus radiographic bone loss exceeding the threshold set by the consensus definition in use at the time) [1]. That is the number that gets cited in newspaper articles about implant complications, and it is a real number from a real registry study.
Here is the second number: 14.5% of implants had peri-implantitis characterised by bone loss greater than 2 mm — the threshold the authors considered clinically significant bone destruction [1].
Both numbers are true. They measure different things. The 45% patient-level figure captures anyone who showed any peri-implant bone loss beyond the threshold, including mild cases that may not progress to implant loss without intervention. The 14.5% implant-level figure captures the subset with substantial bone destruction. The distance between these two numbers — 45% and 14.5% — is the space in which careful monitoring, early intervention, and competent follow-up operates. That space is not small. Managing what lives in that space is what follow-up appointments are for.
Bone loss thresholds. The study used ≥2 mm of radiographic bone loss from the expected bone level as its threshold for peri-implantitis at the moderate-to-severe end. This is consistent with the Chapple-Tonetti-Schwarz consensus classification of peri-implant diseases in use across the literature. It is not a trivial threshold — 2 mm of crestal bone loss at an implant represents a meaningful change in the supporting architecture.
Prosthesis-level complications. The Pjetursson 2012 review documented that beyond survival, biological and technical complications were common: at five years, any complication had occurred in roughly 33% of implant-supported fixed prostheses [2]. Most were not catastrophic. Screw loosening, minor chipping of ceramic veneers, and peri-implant soft-tissue issues were the predominant events. The ten-year complication picture was correspondingly heavier.
The message in the complication data is not that implants fail routinely. It is that implants are not inert posts you place and forget. They require the same attention over a decade that a well-maintained car requires — routine check-ups, early attention to small problems before they become large ones, and a clinician who can read the follow-up film rather than simply confirming the implant is still there.
Subgroups
Registry data allows us to ask which patients did worse. The Derks analysis and the broader Scandinavian implant literature identify four patient-level risk factors with consistent support across studies:
Smokers. Smoking is associated with higher rates of early implant failure (primary non-osseointegration) and higher rates of peri-implantitis progression over time. The biological mechanism involves impaired vascularity at the implant-bone interface and altered immune response to bacterial biofilm at the peri-implant sulcus. Implants in smokers survive — the survival rate does not fall off a cliff — but the complication burden is substantially higher and the trajectory of marginal bone loss is steeper.
Patients with a history of periodontitis. This is the subgroup that deserves the most attention in the dental tourism context. A patient who has previously lost teeth to periodontal disease carries the same bacterial ecology, the same host response pattern, and often the same compliance behaviour into their implant case. Multiple studies, including registry analyses, show that the prevalence of peri-implantitis is approximately doubled in patients with a periodontal history compared to periodontally healthy patients. Placing an implant in someone with uncontrolled periodontal disease is not a cure for the periodontal disease. The implant inherits the environment.
Poorly controlled diabetics. Hyperglycaemia impairs osseointegration and alters wound healing. The evidence suggests that well-controlled diabetics (HbA1c below 7–8%, depending on the study) have outcomes close to non-diabetics. Poorly controlled diabetics show higher early failure rates and compromised soft-tissue healing. A clinic that does not ask about diabetes status and HbA1c is not doing a complete pre-operative assessment.
Full-arch reconstructions versus single units. The Pjetursson data shows that full-arch implant-supported fixed prostheses (All-on-X type reconstructions) have a higher complication burden than single-tooth implants. The mechanical complexity of a twelve-unit bridge on four to six implants is substantially higher than a single crown on a single fixture. Occlusal loads are distributed differently. Screw access is more complex. Fracture of the prosthetic framework, while uncommon, is catastrophic and expensive when it occurs. I have covered the specific failure patterns of the zirconia full-arch in a separate treatment option review.
The external validity question
This is the Cardiology Trials signature question, and it is the most important thing in this article for someone who is planning to have implants placed outside their home country.
These were not the patients you probably are.
The Derks 2016 cohort is Swedish, publicly funded, placed in 2003–2004, examined at nine years in a system with structured recall. The patients had standing relationships with Swedish dental clinics. Follow-up was achieved at high rates because the treating clinician could contact the patient through the reimbursement system. When peri-implantitis was detected, it was detected in a clinic that had also placed the implant and had access to the original treatment records, the pre-operative CBCT, the surgical notes, and nine years of annual radiographs.
The Pjetursson 2012 data is drawn from published studies — which are subject to their own selection effects. Published studies over-represent specialist centres, protocol-compliant patients, and well-resourced settings. Registry data corrects for some of this. It does not correct for the difference between Scandinavian public dental health and the private international dental tourism market.
The dental tourist’s profile differs in five specific ways:
Follow-up compliance is structurally compromised. The patient returns home after treatment. The treating clinic is on a different continent. The home dentist did not place the implant and does not have the surgical records. The peri-implantitis that would have been caught at a nine-month recall in Gothenburg goes undetected until it reaches the threshold where the patient notices symptoms — which, for peri-implantitis, is often after substantial bone loss.
The treatment setting is different. Registry patients in Sweden were placed by a mix of specialists and generalists within a regulated credentialling system. The international dental tourism market ranges from highly trained specialists with structured training equivalent to any Australian or UK programme, to clinics where the qualification behind the title is not verifiable. I discuss how to evaluate the latter in the clinical standards framework.
The patient risk profile may be different. Registry patients are whoever presented in the Swedish system. Dental tourists are a self-selected group — often patients who delayed treatment due to cost, who may arrive with more complex cases and higher complication risk than the average registry patient. The access crisis long-read explains why cost-driven delay is structurally built into the source markets. The dental tourism trust gap describes what this produces in the consulting room.
Implant brand and component traceability. The Swedish registry is composed of brands that were commercially prominent in 2003–2004 and have decade-plus post-market safety data. Some implants placed in high-volume international clinics — particularly on the lower end of the price range — use systems with shorter commercial histories and less extensive registry representation. This does not mean they are worse; it means the evidence base is shorter. I have discussed this in the cost comparison here.
The prosthetic-surgical handoff. In well-resourced Scandinavian placements, the surgeon who places the implant often works closely with the prosthodontist who designs and fits the crown. In high-volume international settings — particularly where patients are managed in condensed itinerary formats — the handoff between surgical and prosthetic phases is more compressed. As I noted in the grafting piece, a treatment plan that was designed around a patient’s holiday schedule rather than around sound clinical sequencing carries a different risk profile than one designed in the other order. The prior decision — whether an implant is even the right treatment — is discussed in the save-a-tooth piece.
What this means for you
The five-year survival figure above 95% is real. It is not a clinic’s marketing invention. The evidence for it comes from the kind of careful, large-scale data compilation that takes years to produce and has been replicated consistently across the implant literature. If a competent clinician places an implant in a suitable patient, with adequate bone, in an appropriate position, with a sound prosthetic plan — the probability of that implant being in function five years later is above 95%. That is genuinely reassuring.
The peri-implantitis burden is also real. The 45% patient-level figure at nine years in the Derks cohort is not an anomaly. It is consistent with what systematic reviewers and registry analysts have found across the literature. Most of those cases are at the mild end of the spectrum, manageable with early intervention. The 14.5% with bone loss exceeding 2 mm — the clinically significant threshold — represents the minority. But 14.5% of a large installed base of implants is still a large number of patients who will, at some point over the decade following placement, need clinical attention that goes beyond a routine check.
The implications for someone choosing an international clinic are practical and specific:
Ask about the follow-up protocol before you commit. Not “do you have a follow-up protocol” — every clinic will say yes. Ask what happens when you are back home and your new home dentist finds a 2 mm pocket with bleeding on probing at the implant site at your twelve-month recall. Who does the CBCT? Who reads it? Who decides on treatment? Who pays for the decontamination procedure if peri-implantitis is diagnosed?
Establish a home dentist who will own the follow-up. Before you leave for treatment, confirm that a dentist at home is willing to monitor your implant, is willing to contact the placing clinic if needed, and has a protocol for peri-implant disease management. This is not an administrative nicety. It is the mechanism by which the space between 45% and 14.5% gets managed.
Disclose your full medical and dental history to the placing clinic. Smoking, diabetes, periodontal history — every one of these modifies the outcome in a way the registry data shows clearly. A clinic that places implants without this information is not in a position to give you an individualised survival estimate.
Understand what brand is being placed and why. You do not need to specify the brand. You do need to understand that the brand has a demonstrable commercial track record, that components will be available in your home country if the crown needs remaking, and that the implant system is one your home prosthodontist recognises.
Falsification condition
If a study emerges from a national implant registry — Scandinavian or otherwise — with adequate follow-up (minimum five years), patient-level rather than publication-level selection, and a survival rate below 90% for fixture-level outcomes across the general population, the assessment in this article would need revision. Similarly, if a well-powered prospective cohort of dental tourism patients showed peri-implantitis prevalence substantially lower than the 14.5% moderate-to-severe figure in Derks 2016, I would revise the follow-up emphasis accordingly. The registry evidence as it currently stands supports both the high survival rate and the significant peri-implantitis burden. Neither finding is optional.
Sources are listed in the frontmatter. This article was clinically reviewed by Dr. Maloney on 2026-05-10. It is not individual treatment advice. For the methodology behind Trial of the Week reviews, see /methodology/.