Research & Evidence: All-on-4 vs All-on-6
Most All-on-4 vs All-on-6 content on the internet quotes a percentage and moves on. This page does the opposite: it sets out what published research on full-arch implants can and cannot actually tell you, defines the measures that get blurred together in marketing copy, and lists the specific studies behind the claims made across this site — each with a working identifier you can check yourself.
Written by All-on-4 vs All-on-6 — Editorial · Clinically reviewed by Dr Hasan Ali Ahmed, GDC 327607 · Last medically reviewed 20 July 2026
Is the evidence strong enough to declare a winner?
No — and any page that tells you otherwise is overstating what the literature supports. The strongest available evidence, including a 2026 systematic review and meta-analysis pooling both configurations and a large single-centre cohort comparing them directly, reports broadly comparable implant survival for All-on-4 and All-on-6 across the follow-up periods studied. Direct head-to-head comparisons remain scarce, definitions of "success" vary between studies, and follow-up length differs widely — all of which limit how precisely the two configurations can be compared. That is a genuine, current limitation of the evidence base, not an oversight of this page.
How strong is the evidence comparing four and six implants?
Full-arch implant dentistry has been studied for decades, and both the All-on-4 concept and six-implant full-arch restorations have a reasonable published track record individually. What is much thinner is evidence that directly compares them. Most of the literature reports outcomes for one configuration from one clinic or research group's patient population — not a randomised or even a matched comparison between four- and six-implant restorations in the same setting. Where comparative data does exist, such as the large retrospective cohort in the table below, patients were not randomly allocated to four or six implants; they were treated according to a clinical decision framework based on their own anatomy, which means the comparison reflects real-world case selection rather than a controlled trial. Systematic reviews that pool the separate All-on-4 and All-on-6 literatures together, like the 2026 review below, explicitly flag high heterogeneity between the studies they combine — different implant systems, patient populations, follow-up lengths and outcome definitions. None of this makes the evidence worthless; it means a fair reading has to hold two things at once: both configurations are generally well supported individually, and a precise, confident ranking between them is not something the current evidence base can respectably deliver.
What studies can and cannot tell us
Not all study designs carry the same weight, and it helps to know roughly where a given source sits before treating its numbers as settled fact:
- Systematic reviews and meta-analyses pool and critically appraise multiple independent studies, which generally gives the most reliable overall picture — but the result is only as good as the studies fed into it, and high heterogeneity between them (different populations, definitions, follow-up lengths) can make pooled figures more indicative than precise.
- Prospective and retrospective cohort studies follow a defined group of patients over time and can offer detailed, real-world outcome data, but without randomisation the patients in a "four-implant group" and a "six-implant group" were not allocated by chance — they were selected by clinical need, which can bias any direct comparison between the groups.
- Case series and retrospective reviews — common in this literature — describe a single clinic's or research group's experience and can under-represent patients lost to follow-up or who had early complications, which tends to skew reported figures toward the more favourable end.
- Professional-body guidance synthesises the evidence and clinical consensus into practical recommendations, but is not itself primary data and lags behind the newest published studies.
Confounding is a recurring issue throughout: smoking status, bruxism, diabetes control, bone quality, implant system, surgeon experience and prosthesis design all vary between study cohorts and are not held constant when a four-implant paper from one clinic is compared with a six-implant paper from another, published years apart. Industry or clinic funding of a study, where disclosed, is also worth noting when weighing a claim — not because it automatically invalidates a result, but because it is one more factor a careful reader accounts for. And, as covered throughout this site, survival is not success: a study that reports only how many implants remained in the jaw is answering a narrower question than one that reports whether they met defined clinical success criteria, functioned well and satisfied the patient.
Four measures, not one number
Before any percentage on this page — or anywhere else — is worth repeating, it helps to know exactly which of these four distinct measures it describes:
| Measure | What it actually captures |
|---|---|
| Implant survival | The implant fixture is still present in the jaw and has not been removed — a presence-only measure that says nothing about function, comfort or bone health around it. |
| Implant success | A stricter, criteria-based judgement — typically no mobility, no persistent pain or infection, bone loss within an accepted threshold, and no ongoing complications — assessed against a defined published framework such as the Albrektsson criteria. |
| Prosthesis survival | Whether the bridge sitting on the implants is still in place and functioning. A prosthesis can be remade while every implant beneath it remains successful, and vice versa — the two figures move somewhat independently. |
| Patient-reported outcomes (PROMs) | The patient's own rating of chewing ability, comfort, appearance and quality of life, usually captured with a validated questionnaire — a distinct measure from any clinician-assessed figure above it, and one that studies report far less consistently. |
These measures are not interchangeable, and a source that quotes only one — usually survival, because it is the easiest to define and produces the highest figure — without saying which one it means is giving you an incomplete picture, whether or not that is intentional.
How to read a survival percentage
A bare percentage on its own tells you very little. Before treating any figure as meaningful, it is worth checking:
- The timeframe. "98% survival" means something different at 1 year than at 10 years — longer follow-up is a harder test, and most full-arch literature has thinner data beyond 5–10 years for both configurations.
- The denominator. Is the percentage calculated per implant placed, per patient treated, or per prosthesis delivered? A per-implant figure in a multi-implant restoration can look higher than a per-patient figure, because one patient can lose an implant while every other implant in the study remains intact.
- Drop-outs and loss to follow-up. Patients who move away, stop attending reviews, or have their case handled elsewhere are not always tracked back into the final figure — and case series in particular can under-represent them, which tends to inflate the reported result.
- What "survival" actually required. Some studies count an implant as surviving even after it needed replacement of the restoration on top, or additional treatment for a complication, as long as the fixture itself stayed in place.
- Whether it is independently published or internally tracked. A figure from a peer-reviewed journal has been through external scrutiny; a figure quoted on a clinic's own marketing page has not, whatever its accuracy might turn out to be.
Applying these five checks to any percentage — including every figure in the table below — is the difference between reading evidence and repeating a slogan.
Verified study table
The table below lists 6 entries drawn from 4 distinct published sources, each independently checked against its original PubMed record before being added. It is a working list, not a full literature review — see the sections above for how the wider evidence base should be read alongside it.
| Study | Type | n | Arch / configuration | Follow-up | Implant survival | Prosthesis survival | Complications | Marginal bone | Limitations | Source |
|---|---|---|---|---|---|---|---|---|---|---|
| All-on-4 and All-on-6 implant-supported fixed prostheses for the edentulous jaw: a systematic review and meta-analysis Shao W-H, Chen R, Wang S, Duan S-Y, Zhang X-D, Tang Y-L · International Journal of Oral and Maxillofacial Surgery (2026) | systematic review & meta-analysis | 55 pooled studies (implant/patient totals not aggregated into one figure by the authors) | Maxilla and mandible (pooled) All-on-4 (four implants) | — | Pooled: 99.20% at 1 year; 99.66% at 1–5 years; 98.14% at ≥5 years | Not separately pooled in this review | Biological: 6.54% pooled Technical: 5.91% pooled (mechanical) | Pooled 0.77 mm at short-term follow-up; 1.28 mm at ≥5 years | Meta-analysis of 55 heterogeneous studies (different implant systems, outcome definitions and patient populations); the authors report high heterogeneity between studies, so pooled figures are indicative rather than precise. Most of the 55 included studies were not direct head-to-head All-on-4-vs-All-on-6 comparisons within the same cohort — this review pools separate literatures for each configuration rather than a single randomised comparison. | PMID 42031576 / DOI 10.1016/j.ijom.2026.04.002 |
| All-on-4 and All-on-6 implant-supported fixed prostheses for the edentulous jaw: a systematic review and meta-analysis Shao W-H, Chen R, Wang S, Duan S-Y, Zhang X-D, Tang Y-L · International Journal of Oral and Maxillofacial Surgery (2026) | systematic review & meta-analysis | 55 pooled studies (implant/patient totals not aggregated into one figure by the authors) | Maxilla and mandible (pooled) All-on-6 (six implants) | — | Pooled: 100% at 1 year; 98.55% at 1–5 years; 97.50% at ≥5 years | Not separately pooled in this review | Biological: 7.41% pooled Technical: 6.64% pooled (mechanical) | Pooled 0.85 mm at short-term follow-up; 0.94 mm at ≥5 years | Same pooled meta-analysis as the All-on-4 row above — see that row for the shared heterogeneity and non-head-to-head caveats. The All-on-6 literature pooled here generally comprises fewer, smaller studies than the All-on-4 literature, which the authors note as a further source of imprecision. | PMID 42031576 / DOI 10.1016/j.ijom.2026.04.002 |
| Four vs. Six Implant Full-Arch Restorations—A Direct Comparative Retrospective Analysis in a Large Controlled Treatment Cohort Caramês JMM, Francisco HCO, Vieira FA, Caramês GB, Martins JNR, Marques DNS · Journal of Clinical Medicine (2025) | retrospective cohort | 943 patients / 5,989 implants (four- and six-implant groups combined) | Maxilla and mandible (pooled) Four-implant fixed complete denture | 60 months | 98.6% at 2 years; 98.8% at 5 years (average follow-up 5.0 ± 3.2 years, range 0–17 years) | Not evaluated — outside this study's stated scope | Biological: Not evaluated — outside this study's stated scope Technical: Not evaluated — outside this study's stated scope | Not reported | Single-centre retrospective analysis using one clinical decision-support framework to allocate patients to four or six implants (not randomised), so allocation itself is not independent of case selection. Average follow-up was 5.0 ± 3.2 years but individual follow-up ranged from 0 to 17 years, so long-term figures rest on a smaller subset of patients than the topline numbers suggest. Complications and prosthesis-level survival were not evaluated in this paper. | PMID 40565982 / DOI 10.3390/jcm14124237 |
| Four vs. Six Implant Full-Arch Restorations—A Direct Comparative Retrospective Analysis in a Large Controlled Treatment Cohort Caramês JMM, Francisco HCO, Vieira FA, Caramês GB, Martins JNR, Marques DNS · Journal of Clinical Medicine (2025) | retrospective cohort | 943 patients / 5,989 implants (four- and six-implant groups combined) | Maxilla and mandible (pooled) Six-implant fixed complete denture | 60 months | 98.4% at 2 years; 98.7% at 5 years (average follow-up 5.0 ± 3.2 years, range 0–17 years) | Not evaluated — outside this study's stated scope | Biological: Not evaluated — outside this study's stated scope Technical: Not evaluated — outside this study's stated scope | Not reported | Same single-centre retrospective cohort as the four-implant row above — see that row for the shared allocation, follow-up-range and scope caveats. The study's own conclusion is that four- and six-implant restorations showed high and comparable mid-to-long-term survival under its decision-support protocol, not that either configuration outperformed the other. | PMID 40565982 / DOI 10.3390/jcm14124237 |
| The All-on-Four Treatment Concept: A Systematic Review Patzelt SBM, Bahat O, Reynolds MA, Strub JR · Clinical Implant Dentistry and Related Research (2014) | systematic review | 13 studies included (from 487 screened) / 4,804 implants / 1,201 prostheses | Maxilla and mandible (pooled across included studies) All-on-4 (four implants, two posterior tilted) | 36 months | 99.0% ± 1.0% cumulative survival at 36 months (74 of 4,804 implants failed; 74% of failures occurred within the first 12 months) | 99.9% ± 0.3% at 36 months | Biological: Not separately pooled beyond implant failure in this review Technical: All-acrylic fixed prosthesis fracture was the most frequently reported technical complication | Mean 1.3 mm ± 0.4 mm at 36 months; no significant difference reported between maxilla/mandible or axial/tilted implant placement | Short-term evidence base — most pooled data extends only to 36 months, and the authors note a paucity of studies with 5+ years of follow-up at the time of writing. Covers the All-on-4 protocol only; it is not a head-to-head comparison with All-on-6. Now over a decade old — later reviews (see rows above) draw on a larger and longer-follow-up literature. | PMID 23560986 / DOI 10.1111/cid.12068 |
| The All-on-4 treatment concept for the rehabilitation of the completely edentulous mandible: A longitudinal study with 10 to 18 years of follow-up Maló P, de Araújo Nobre M, Lopes A, Ferro A, Botto J · Clinical Implant Dentistry and Related Research (2019) | retrospective cohort | 471 patients / 1,884 implants (mandible only) | Mandible only All-on-4 (four implants, two posterior tilted) | — | 93% cumulative implant survival and 91.7% cumulative implant success up to 18 years (120–216 months) — the gap between the two figures over a long follow-up illustrates why survival and success are reported separately | 98.8% cumulative prosthetic survival | Biological: 11.8% of implants had a biological complication over the study period Technical: 36.7% of cases had at least one mechanical/technical complication over the study period | Mean 1.72 mm at 10 years; 2.32 mm at 15 years | Mandible only — does not cover the maxilla or All-on-6, and is not a comparison between configurations. Single-centre, single-protocol case series from the clinical team that originated the All-on-4 concept, not an independent or randomised study. The authors identify smoking, prior implant failure and biological complications as significant risk factors, which the headline survival figure does not show on its own. | PMID 30924309 / DOI 10.1111/cid.12769 |
Only human-verified entries appear in this table. Each row was checked against the original PubMed record (title, authors, journal, year and a working PMID/DOI) before being added. Nothing here is copied from clinic marketing pages. Where a single comparative study reports separate figures for All-on-4 and All-on-6, it appears as two rows sharing the same identifier so each configuration's figures stay attributable. This is a working list, not an exhaustive literature review — see the page text for what the wider evidence base can and cannot support.Table last checked 20 July 2026.
What this means for your decision
Taken together, the strongest available evidence supports a small number of genuinely defensible conclusions: both All-on-4 and All-on-6, in well-selected patients treated by experienced clinicians, are reported with high implant survival across the periods studied; direct, precise comparisons between the two remain limited; and factors such as planning accuracy, hygiene, bruxism management and clinic follow-up are more consistently associated with long-term outcomes than the raw implant count. None of this tells you, on its own, which configuration might suit your own case — that depends on your bone, bite, budget and priorities, assessed individually. For how those individual factors are weighed, see who needs which. For the fuller explanation of survival versus success and what determines how long a restoration lasts, see longevity & success rates. For what can go wrong and the risk factors involved, see failure & complications. For how available bone specifically affects the choice, see bone loss & bone grafting. Return to the All-on-4 vs All-on-6 comparison home page for the full overview.
Frequently asked questions
Is there solid research proving All-on-6 lasts longer than All-on-4?
No. Reading the study table above and the wider literature together, there is no strong, direct evidence base showing that All-on-6 reliably outlasts All-on-4, or vice versa. The largest recent meta-analysis pooling both configurations reports broadly comparable survival at every follow-up window studied, and a large single-centre cohort explicitly comparing the two concluded the two approaches were "high and comparable" rather than one outperforming the other. Claims that six implants categorically last longer are not something the current published evidence can responsibly support.
Why do so few studies compare All-on-4 and All-on-6 directly?
Running a genuine head-to-head study — the same patient population, implant system, surgeon and follow-up protocol, with patients allocated to four or six implants for comparison rather than by clinical need — is difficult to do ethically and practically, because the number of implants a clinician recommends is itself a clinical decision based on that patient's anatomy. Most published research instead reports outcomes for one configuration or the other, from different clinics, patient groups and years, which is why the evidence table above shows several rows from the same comparative papers rather than dozens of independent head-to-head trials.
What is the difference between implant survival and implant success in these studies?
Survival means the implant fixture is still present and has not been removed — it says nothing about how well it is functioning. Success is a stricter, criteria-based judgement covering mobility, pain, infection and bone loss thresholds. The 2019 Maló mandible study in the table above illustrates the gap directly: 93% cumulative survival against 91.7% cumulative success over the same 18-year follow-up. See longevity & success rates for the fuller explanation of why these numbers diverge.
Can I trust a clinic that quotes a specific success percentage?
It depends on whether the figure is sourced. A clinic that can say which measure it means (survival, success or prosthesis survival), over what follow-up period, based on how many of its own patients, and whether it is citing independently published research or an internal figure, is giving you something you can actually evaluate. A bare percentage with none of that context — however confident it sounds — is not verifiable and should be read with the same caution as any unsourced statistic.
Does a systematic review or meta-analysis outrank a single clinic's case series?
Generally, yes, in terms of how much weight the finding can bear — a systematic review or meta-analysis pools and critically appraises multiple independent studies, which reduces (though does not eliminate) the risk that one clinic's particular patient selection or follow-up practices are driving the result. See what studies can and cannot tell us below for how this hierarchy works and where its limits are.
How often is this evidence table updated?
The table is reviewed periodically as new systematic reviews and comparative studies are published and independently verified against their original source. Each row states the identifier it was checked against; if you believe a row is inaccurate or out of date, see the corrections policy for how to flag it.