Before we describe what happened in the trials, here is who was in them: adult patients in selected research settings, with adequate bone for the planned implant positions, with case characteristics that allowed informed consent under a randomised protocol, with osseointegration assessed at standardised time points, and with follow-up at one to five years depending on which arm of the Cochrane review you are reading. The Cochrane systematic review by Esposito and colleagues — first published in 2007 and updated through several iterations — pooled the available RCT evidence on immediate, early, and conventional implant loading and reported, with GRADE-rated certainty assessment, that pooled implant failure rates between loading protocols were similar in the populations the included trials studied [1] [2].
That is the headline. It is widely cited in implant marketing, often as the unqualified statement that immediate loading is as good as delayed loading. The published Cochrane review itself does not quite say that. It says what the systematic-review evidence in the available trials supports, qualified by the certainty of evidence in each comparison, and the certainty is not high across every comparison. This is the procedural-quality companion to the All-on-4 patient guide and to the Scandinavian implant registry trial review. The trial-of-the-week format walks through the Cardiology-Trials four sections — Patients, Procedures and endpoints, Subgroups, Conclusions — and applies the same external-validity question this column applies to every trial it reviews.
Patients — who was in the included trials, and who was not
The Cochrane review’s inclusion criteria selected for randomised controlled trials of patients receiving dental implants under defined loading protocols [1]. The trials enrolled, in the aggregate, adult patients in research settings, with sufficient bone volume at the planned implant sites for the protocol being tested, with no major medical contraindications to implant placement, with no current heavy smoking in most of the included trials, with manageable periodontal status, and with the willingness and ability to participate in a follow-up protocol typically running between twelve months and five years post-loading.
The trials, in aggregate, did not enrol — or enrolled in much smaller numbers — patients with severely resorbed alveolar ridges; patients with uncontrolled diabetes; patients with current heavy smoking habits; patients with severe untreated periodontal disease; patients with documented severe bruxism unaddressed by occlusal-protection protocols; patients undergoing full-arch implant reconstruction at high-volume international clinics under one-week treatment timelines. The high-volume dental tourism patient population is, in the Cochrane review’s inclusion criteria, substantially under-represented [3] [4].
This is not a criticism of the review. A systematic review can only synthesise the published trial evidence available. The point is that the patient the review’s conclusions are being applied to, in the marketing on a high-volume tourism clinic website, is often not the patient the included trials enrolled. The mean implant survival figures in the trials are real numbers; the external-validity question is whether they transfer to the patient in front of the marketing.
Procedures and endpoints — what was tested, and what was measured
The Cochrane review compared three categories of implant loading [1]:
- Immediate loading — placement of the prosthesis (typically a provisional) within one week of implant placement, often within hours.
- Early loading — prosthesis placement between one week and two months after implant placement.
- Conventional (delayed) loading — prosthesis placement after a minimum of two months of unloaded osseointegration, which has historically been the standard protocol after the Brånemark school of osseointegration biology [3].
The pooled comparisons in the various Cochrane review iterations included immediate vs conventional, early vs conventional, and immediate vs early. The endpoints included implant failure (loss of the implant), prosthesis failure (loss of the prosthesis above an integrated implant), peri-implant bone loss (marginal bone changes on radiographs), and adverse events.
The follow-up periods in the included trials varied. Most were at twelve months. Some extended to three or five years. None of the included trials, in the Cochrane review’s published reading, provided the ten-year-plus follow-up that the Scandinavian implant registry data provides on the question of long-term implant survival, peri-implantitis incidence, and prosthesis maintenance burden. The Cochrane review is a synthesis of the trial evidence available; the trial evidence available, on long-term outcomes, is thinner than the published numbers might suggest at first reading [4].
Subgroups — where the pooled conclusion does and does not hold
The pooled result — similar implant failure rates between loading protocols, low or moderate GRADE certainty depending on the comparison [6] — does not hold uniformly across every subgroup the review considered.
The subgroup picture, in the Cochrane review’s reading and in subsequent literature, runs as follows. Implants placed in the mandible (lower jaw) have shown more consistent equivalence between immediate and conventional loading than implants placed in the maxilla (upper jaw), where the bone density is on average lower and the failure rates under immediate loading have been more variable. Implants placed with high primary stability (insertion torque values above defined thresholds, typically 30–35 Ncm or higher depending on the protocol) have shown more reliable immediate-loading outcomes than implants placed with marginal primary stability. Implants placed at sites without simultaneous grafting have shown more reliable outcomes than implants placed at simultaneously grafted sites. The single-tooth indication has shown more variable immediate-loading outcomes than the splinted multi-implant configuration where load is distributed across multiple fixtures.
The patient considering a procedure that involves immediate loading is, on the published Cochrane evidence, asking the right question if they ask whether their specific case (mandible vs maxilla, primary stability, grafting status, single vs splinted) falls within the subgroups where the pooled equivalence holds. A clinician offering immediate loading should be able to answer that question on the specific case. A clinician who treats the pooled result as universal — “the Cochrane review says immediate loading is as good as delayed” — has read past the subgroup work the review actually did.
Conclusions — what the published evidence supports, and what it does not
The Cochrane review’s published conclusion, in the iteration most widely cited, is that there is some evidence that immediate and early loading protocols may achieve outcomes similar to conventional loading in certain conditions, with low to moderate certainty depending on the comparison [1]. The review explicitly notes the limitations of the evidence base — trial heterogeneity, follow-up duration, variation in operator experience, variation in implant systems, variation in surgical protocols. The certainty of evidence is not uniformly high. The conclusion is not “immediate loading is as good as delayed loading in all cases.” It is “in the cases the included trials enrolled, the pooled failure rates were similar, with the certainty caveats specified” [6].
This is the kind of conclusion that is difficult to translate into a clinic’s marketing copy without losing the qualifications. The headline that survives the translation is “immediate loading is supported by Cochrane evidence.” The qualifications that do not survive the translation are “in selected cases, with adequate primary stability, in the subgroups studied, at one-to-five-year follow-up, with low-to-moderate GRADE certainty.” Both versions are based on the same review. Only the second is what the review actually says.
External validity — the question this column applies to every trial it reviews
The external-validity question is whether the trial population the review synthesised resembles the patient sitting in the marketing of the clinic the patient is evaluating. For the high-volume dental tourism patient considering an All-on-4 procedure with immediate loading at an Antalya or Hanoi clinic, the external-validity question has a specific shape.
The Cochrane review’s included trials were predominantly conducted in specialist or university research settings, by operators with documented training, on cases selected against published inclusion criteria, with follow-up protocols enforced by the trial design. The high-volume dental tourism procedure is performed in a clinic with high case throughput, by operators whose individual case experience and continuing-education credentials may or may not be transparent on the clinic’s marketing surface (see the publication’s clinic reviews for what credential-trace work looks like in practice), on patients selected at least partly against the price they were willing to pay rather than against clinical inclusion criteria, with follow-up dependent on the patient’s return travel and on the destination clinic’s continuity-of-care protocols.
The trials’ conclusions transfer to the dental tourism patient to the extent that the dental tourism case resembles the trial inclusion criteria. Where the cases diverge — on operator experience, case selection, follow-up cadence, or the conditions the trials excluded — the pooled trial result is the upper bound of what the patient should expect, not the expected outcome.
For the cost arithmetic that drives most of these decisions, see the implant cost-by-country reference. For the upstream procedural decision on whether the natural tooth recommended for extraction in order to enable implant placement should in fact be saved, see when to save a tooth and when to replace it and vital pulp therapy vs root canal. For the procedural critique of the high-volume version of the full-arch procedure that most often pairs with immediate loading in dental tourism, see zirconia full-arch when it’s the wrong answer. For the complementary registry-grade trial review, see the Scandinavian implant registry trial. For the long-read on the structural reasons a patient cannot, from the marketing surface, evaluate any of these subgroup questions on their own case, see the dental tourism trust gap.
What would change my reading
I read the Cochrane review’s conclusion as a low-to-moderate-certainty statement of equivalence within the populations the included trials studied, with the caveats the review’s own GRADE rating documents [6]. The evidence that would update this reading:
A registered, multi-centre, multi-country RCT with N > 2,000, ten-year follow-up, mandatory parafunction screening, documented primary-stability thresholds at placement, and pre-specified subgroup analyses on the mandible-vs-maxilla, single-vs-splinted, and grafted-vs-non-grafted comparisons, that reported either materially better or materially worse outcomes for immediate vs conventional loading than the current Cochrane synthesis. I am not aware of such a trial in active enrolment.
A registry analysis (along the lines of the Scandinavian implant registry data) reporting immediate-vs-delayed loading outcomes at ten and twenty years across a population including high-volume dental tourism cases, with explicit external-validity comparison to the Cochrane-included trial populations. The published registry data at long horizons is thin on this comparison; the question would be settled by a registry that included it.
A methodology critique of the Cochrane review’s pooling decisions — specifically whether trials with materially different primary-stability thresholds or operator-experience criteria should have been pooled, and what the result would look like under stricter inclusion criteria. The critiques I have read so far accept the pooling decisions and focus on the certainty caveats the review itself acknowledges.
If any of these emerges, the reading I take of the Cochrane synthesis shifts. Until then, the synthesis is the strongest summary we have of the available trial evidence on this question, the qualifications the review itself documents are the qualifications the patient deserves to hear, and the external-validity question for the high-volume dental tourism case is the question every clinic offering immediate loading should be able to answer on the specific patient in front of them.
Sources
- Esposito M, Grusovin MG, Maghaireh H, Worthington HV. Interventions for replacing missing teeth: different times for loading dental implants. Cochrane Database of Systematic Reviews 2013, Issue 3. https://www.cochranelibrary.com/cdsr/doi/10.1002/14651858.CD003878.pub5/full
- Cochrane Library. https://www.cochranelibrary.com/
- Wikipedia. Osseointegration. https://en.wikipedia.org/wiki/Osseointegration
- Wikipedia. Dental implant. https://en.wikipedia.org/wiki/Dental_implant
- Wikipedia. Systematic review. https://en.wikipedia.org/wiki/Systematic_review
- Wikipedia. GRADE approach. https://en.wikipedia.org/wiki/GRADE_approach