Implant Number, Antero-Posterior Spread, and Cantilever Length in Full-Arch Fixed Prostheses: A Dimensionless Load-Sharing Analysis and a Finite Element Protocol With Global Sensitivity Analysis and Uncertainty Quantification
- Paper type
- Technical methodology article with an analytic screening computation and a prespecified computational protocol (no clinical data)
- Status
- Original Work · Working Paper
- Keywords:
- full-arch fixed prosthesis · cantilever · antero-posterior spread · implant number · load sharing · finite element analysis · Sobol sensitivity indices · uncertainty quantification
Contents
- 1Introduction and Clinical Significance
- 2Critical Review of the Established Literature
- 3Identification of the Specific Research Gap
- 4Scientific Objective and Hypothesis
- 5Theoretical Foundation
- 6Methodology
- 7Results of the Analytic Model and Explicitly Labeled Theoretical Predictions
- 8Discussion: Mechanisms and Competing Explanations
- 9Implications for Implant Surgery and Restorative Dentistry
- 10Technical and Clinical Limitations
- 11Proposed Experimental and Clinical Validation
- 12Future Research Directions
- 13Evidence-Calibrated Conclusion
Abstract
Background: Full-arch fixed prostheses are supported by four to eight implants, frequently with distal cantilevers. Clinical guidance on cantilever length is expressed as multiples of the antero-posterior spread of the implants, a rule derived from early beam models. Finite element studies of implant number and tilting are numerous but mostly deterministic, each reporting stresses for a few fixed geometries, and they cannot apportion the outcome among design variables or state its uncertainty.
Objective: To derive the dependence of distal implant load on implant number, spread, and cantilever in a transparent analytic model, to quantify the contribution of each by global sensitivity analysis, and to specify a finite element protocol able to test whether the conclusions survive realistic compliance.
Methods: A rigid-plate, equal-spring load-sharing model was formulated for implants on a parabolic arch under a unilateral cantilever load, and first-order and total Sobol indices were computed by Monte Carlo sampling. A finite element protocol was written with probabilistic bone, framework, and loading inputs and with verification and validation steps.
Results of the analytic model: The maximum implant load, normalized by the applied load, depends only on implant number and the ratio of cantilever length to antero-posterior spread. For four implants it rose from 0.95 at zero cantilever to 1.51 at a ratio of 1 and 2.08 at a ratio of 2; for six implants the corresponding values were 0.82, 1.38, and 1.94. Over a ratio of 0 to 2 and four to eight implants, the ratio accounted for about 87 percent of output variance and implant number for about 12 percent. These are properties of an idealized model, not measurements.
Conclusions: In the rigid idealization the cantilever-to-spread ratio dominates and adding implants within a fixed spread has a secondary effect. We hypothesize that the same ranking holds for peak distal bone strain in a compliant finite element model; a first-order sensitivity index of 0.20 or more for implant number would refute it. Clinical inference requires validation that has not been performed.
Keywords: full-arch fixed prosthesis; cantilever; antero-posterior spread; implant number; load sharing; finite element analysis; Sobol sensitivity indices; uncertainty quantification
1. Introduction and Clinical Significance
A fixed prosthesis for the edentulous jaw is a curved beam on a small number of supports. Where the supports can be placed is dictated by anatomy: the mental foramina and inferior alveolar canal in the mandible, the maxillary sinuses in the maxilla. The dentition to be replaced extends beyond those limits, so the prosthesis commonly overhangs its most distal implant. How many implants are needed, how they should be distributed, and how long the overhang may be are among the oldest questions in implant prosthodontics, and they have acquired new urgency as protocols using four implants, with the distal pair tilted to move their platforms posteriorly, have become widespread [1, 2].
Long-term clinical series of four-implant mandibular rehabilitations report high implant and prosthesis survival over ten years and more [3, 4], and a systematic review of the concept summarized favorable medium-term outcomes while noting the limited number of independent long-term studies [5]. Systematic reviews asking how many implants a fixed reconstruction requires have not identified an optimal number: outcomes with fewer than five implants per arch were similar to those with more in the available studies, which were heterogeneous and largely observational [6–8]. Classic data from the original protocol showed comparable ten-year prosthesis survival on four or six implants [9].
Survival, however, is a blunt outcome for a mechanical question. The complications that burden full-arch patients are technical: veneer fracture, screw loosening and fracture, framework fracture, and wear. A meta-analysis of fixed prostheses in edentulous patients found prosthodontic complications to be frequent over observation periods of five years or longer [10], and a systematic review found that both biological and technical complications are common with fixed rehabilitations for edentulous patients [11]. For implant-supported fixed prostheses in general, ten-year prosthesis survival is substantially lower than implant survival, with technical complications predominant [12]. If design variables influence these events, they do so through load, and the relevant science is mechanics.
This paper is confined to that mechanics. It does not evaluate surgical protocols, immediate loading, or prosthetic materials. It asks how three geometric variables combine to determine load on the distal implant and strain in the bone around it, and how confidently that can be stated.
2. Critical Review of the Established Literature
Analytic models
Skalak's biomechanical analysis of osseointegrated prostheses treated the prosthesis as a rigid body on elastic implants and showed that a vertical load is shared according to the position of the load relative to the centroid and second moments of the implant group, and that a cantilevered load produces compression on the nearest implant exceeding the applied load and tension on implants on the far side [13]. Rangert and colleagues developed these ideas into clinical guidelines on forces and moments, describing the cantilever as a lever whose fulcrum is the distal implant and whose resisting arm is the distance to the anterior implants [14]. The antero-posterior spread, the distance from the center of the most anterior implant to a line joining the distal aspects of the most posterior implants, became the standard clinical measure of that resisting arm, and the rule that cantilever length should not exceed some multiple of it entered clinical teaching. Reviews of implant biomechanics summarize these principles and caution that the rigid model neglects the compliance of the framework and of bone [15].
Measured loads
In vivo strain-gauge measurement on abutments supporting fixed prostheses confirmed the qualitative predictions: loading a cantilever produced the highest axial forces and bending moments on the implant nearest the load, and distribution depended on the number and position of supporting implants [16]. A subsequent in vivo study varying implant number, distribution, and prosthesis material found that distribution, and not merely number, governed the loads recorded, and that a stiffer prosthesis distributed load more evenly [17]. These studies involved few patients and static or simple chewing tasks; they establish that the mechanics is real in the mouth and do not provide population estimates.
Finite element studies
Finite element analysis has been applied to implant dentistry for decades, and reviews describe both its uses and its recurrent weaknesses: simplified geometry, homogeneous isotropic bone, bonded interfaces, static loads, and absent validation [18, 19]. For cantilevered implant prostheses, an early three-dimensional analysis showed that stress in the framework and around the distal implant increased with cantilever length [20]. A two-dimensional analysis found that tilting splinted distal implants did not increase peri-implant bone stress relative to axial implants and reduced it when tilting shortened the cantilever [21]. Three-dimensional models comparing tilted and non-tilted configurations for the edentulous mandible reached the same conclusion, attributing the benefit to cantilever reduction [22], and a maxillary model reported that inclining the distal implants and shortening the cantilever together lowered stress in bone and framework [23]. Analyses of the number and inclination of implants in four-implant configurations [24], of prostheses on four versus six implants [25], and of several implant configurations for a mandibular fixed prosthesis [26] reported that additional implants and wider distribution reduced peak stresses to varying degrees.
The conclusions agree in direction and cannot be compared in magnitude. Each study fixed its own geometry, material properties, load magnitude and direction, and output measure (von Mises stress in cortical bone in one, principal strain in another, stress in the framework in a third). None varied its inputs over their plausible ranges, so none can say whether a difference between four and six implants is large or small relative to the effect of bone quality or load direction. A review addressing how far finite element models of dental implants can be expected to predict biological phenomena concluded that few were validated and that outputs should be interpreted comparatively [27].
Clinical evidence on cantilevers and tilting
Systematic reviews of implant-supported fixed prostheses with cantilever extensions, mostly in partially edentulous patients, found five-year prosthesis survival above 90 percent, with technical complications such as veneer fracture and screw loosening more frequent than in prostheses without cantilevers and little evidence of increased marginal bone loss [28, 29]. An early study relating survival of fixed implant prostheses to cantilever length reported better survival with shorter cantilevers [30], and five-year results of mandibular fixed rehabilitations with distal cantilevers showed high implant survival with technical complications [31]. Clinical series have recorded cantilever lengths and antero-posterior spreads of interim and definitive full-arch prostheses and examined their relationship to prosthetic complications; in one, about 18 percent of interim all-acrylic prostheses required at least one repair during treatment [32, 33]. Meta-analyses comparing tilted with axially placed implants found no overall difference in implant failure or marginal bone loss, although a subgroup analysis restricted to the maxilla showed a higher failure risk for tilted implants [34, 35].
Bone response to load
Whether elevated load on a distal implant harms bone depends on the strain it produces and on the tissue's condition. The mechanostat concept proposes strain ranges associated with disuse, maintenance, modeling, and pathological overload [36]. For implants specifically, a systematic review found little unbiased evidence that occlusal overload causes loss of osseointegration or marginal bone loss in well-integrated implants [37]. Mechanical analysis of full-arch designs is therefore better motivated by technical complications of components and frameworks, for which load is unambiguously causal, than by bone loss, for which the link is uncertain.
3. Identification of the Specific Research Gap
It is established that a cantilever load is amplified at the distal implant and reversed at remote implants, that tilting distal implants reduces the cantilever and with it the amplification, and that clinical outcomes over ten years are acceptable with four or more implants. It is uncertain how the three design variables rank in importance, whether a fifth and sixth implant placed within the same spread materially reduce distal load, and how sensitive any such conclusion is to bone properties, framework stiffness, and loading.
Existing computational methods are inadequate for those questions in specific ways. Deterministic, one-geometry-at-a-time analyses yield point values without ranges. Comparisons across studies are confounded by modeling choices. Results are reported in dimensional form for a particular arch, so they do not generalize. Validation against measurement is rare. And the standard clinical rule, a maximum cantilever expressed as a multiple of spread, has been transmitted as a rule of thumb without a published sensitivity analysis showing how much of the mechanical outcome it captures.
The specific gap is a global sensitivity analysis: a quantitative apportionment of variance in a clinically meaningful mechanical output among design and patient variables, with uncertainty propagated from inputs to output, and with the model's credibility established by verification and validation appropriate to its intended use [38, 39].
4. Scientific Objective and Hypothesis
The objective is to quantify the contributions of implant number, antero-posterior spread, and cantilever length to distal implant load and distal crestal bone strain in a full-arch fixed prosthesis.
Hypothesis H7: peak principal strain in crestal bone at the distal implant scales primarily with the dimensionless ratio of cantilever length to antero-posterior spread (C/AP); implant number in the range four to eight, at fixed C/AP, accounts for less than 20 percent of the variance in that output when all inputs are varied over clinically plausible ranges.
H7 is falsified if, in the validated finite element model, the first-order Sobol sensitivity index of implant number is 0.20 or greater. It is also weakened, though not strictly falsified, if the index for C/AP falls below those of patient variables such as cortical thickness, which would mean that design matters less than the host. The threshold of 0.20 is a prespecified convention for a non-negligible contribution.
The hypothesis concerns a compliant model. Section 7 reports its counterpart in the rigid idealization, where it can be evaluated exactly; that result is a prior expectation and does not test H7.
5. Theoretical Foundation
The two-support lever
The simplest representation has two supports, an anterior and a distal implant line separated by spread a, and a vertical load F on a cantilever of length c behind the distal support. Equilibrium gives a compressive reaction at the distal support and a tensile reaction at the anterior support:
R_d = F (1 + c/a), R_a = -F c/a (1)
The distal reaction exceeds the applied load by the factor 1 + c/a, and the ratio c/a is the only geometric quantity involved. A cantilever equal to the spread doubles the distal load and places a tensile load equal to the bite force on the anterior implants.
Rigid plate on N elastic supports
Let N implants of equal axial stiffness support a rigid prosthesis, with platform coordinates (xi, yi) in the occlusal plane measured from the centroid of the implant group along its principal axes. A vertical load P applied at (xP, yP) produces axial implant forces
F_i = P (1/N + (x_P x_i)/(∑ x_j^2) + (y_P y_i)/(∑ y_j^2)) (2)
following Skalak's formulation [13]. The first term is the equal share; the second and third are the contributions of the moments about the two principal axes, resisted in proportion to each implant's distance from the centroid and inversely to the second moment of the implant group. Additional implants reduce the equal share and increase the second moments, but an implant placed near the centroid adds almost nothing to the second moments and so does little to resist a cantilever moment. The model thus predicts that where implants are placed matters more than how many there are.
Dimensionless form
If implants are distributed along an arch whose shape is fixed up to scale, with the distal implants at y = 0 and the most anterior point at y = AP, and the load is applied on one distal cantilever at distance C behind the distal implant, then every coordinate in Equation 2 scales with AP in the antero-posterior direction and with arch width in the transverse direction. The transverse term is independent of arch width because numerator and denominator scale together, and the antero-posterior term depends on C and AP only through their ratio. The normalized maximum implant force can therefore be written
(F_(max))/P = g(N, C/AP) (3)
for a given arch shape and spacing rule. This is the theoretical justification for expressing cantilever limits as a multiple of spread, and it shows what the clinical rule omits: the function g also depends on N and on how implants are distributed along the arch.
What the rigid model leaves out
Real frameworks bend. A compliant cantilever transfers load preferentially to the nearest implant and less to remote ones than the rigid model predicts, which raises distal compression and reduces anterior tension. Implants and bone are not equal springs: stiffness varies with bone density, implant length and diameter, and inclination. Loads are not purely vertical. The output of clinical interest, strain in crestal bone or stress in a prosthetic screw, depends on bending moments at the implant as well as on axial force, and tilted implants receive a bending component from vertical load. These are the reasons a finite element model is required and the reasons its conclusion might differ from the rigid one.
Variance-based sensitivity
For a model output Y depending on uncertain inputs X1 to Xk, the first-order Sobol index of input i is
S_i = (Var[E(Y | X_i)])/(Var(Y)) (4)
the fraction of output variance that would be removed if Xi were known. The total index STi adds the contribution of all interactions involving Xi [40, 41]. Unlike one-at-a-time parameter variation, these indices are defined over the whole input space and account for nonlinearity and interaction. They require an explicit statement of the range and distribution of each input, which is itself a discipline that deterministic studies lack.
6. Methodology
Analytic screening model (performed)
The rigid-plate model of Equation 2 was evaluated for implants placed at equal transverse intervals along a parabolic arch, y = AP[1 − (x/W)2], between distal implants at x = ±W, with a unit vertical load on one cantilever at distance C behind the distal implant. Implant numbers of 4, 5, 6, and 8 were considered. The normalized maximum and minimum implant forces were tabulated for C/AP from 0 to 2. Sobol indices for the normalized maximum force were then estimated by Monte Carlo sampling with 100,000 base samples using standard estimators [41], with implant number drawn uniformly from the four values, C/AP uniformly from 0 to 2, and AP uniformly from 6 to 18 mm; a second run sampled C (0 to 20 mm), AP (6 to 18 mm), and arch half-width (18 to 26 mm) as separate inputs. The computation is deterministic apart from sampling error and is fully specified by this paragraph and Equation 2. Its results are reported in Section 7.1.
Finite element protocol (proposed)
The finite element study is specified as follows and has not been performed.
Geometry. A parametric edentulous mandible is generated from segmented cone-beam computed tomography of at least ten atrophic mandibles, to provide a population of shapes; implant positions are defined by N (4, 5, 6, 8), antero-posterior spread, and distal implant inclination (0 to 45 degrees). Implants, multi-unit abutments, prosthetic screws, and a framework of parametric cross-section are modeled explicitly. Cantilever length is varied to give C/AP from 0 to 2.
Materials. Cortical and cancellous bone are assigned elastic moduli sampled from distributions spanning the reported ranges, with cortical thickness as a separate parameter. Framework modulus is sampled over the range from acrylic-reinforced through titanium and cobalt-chromium to zirconia. The bone-implant interface is modeled as bonded for the primary analysis and as frictional contact in a sensitivity analysis.
Loads and constraints. A unilateral load is applied on the cantilever at the first molar position, with magnitude and direction sampled from distributions (vertical component dominant, with lateral components up to a stated fraction), and muscle and condylar constraints represented in a simplified physiological scheme and, in a sensitivity analysis, by fixed inferior border constraints to show the effect of that common simplification.
Outputs. The primary output is the 95th percentile of maximum principal strain in crestal cortical bone within a defined region around the loaded-side distal implant; percentile values are used to avoid mesh-dependent singular peaks. Secondary outputs are axial force and bending moment at each implant platform, peak von Mises stress in prosthetic screws and framework at the distal connector, and tensile force on the contralateral anterior implant.
Verification
Element type and mesh density are fixed by a convergence study in which the primary output changes by less than 5 percent on successive refinement. The solver's implementation is checked by reproducing Equation 2 in a limiting case with a rigid framework and equal linear springs, where the analytic answer is known. Reporting follows published considerations for finite element studies in biomechanics [42] and guidance on extracting clinically relevant data from such simulations [43].
Validation
A physical model is built for at least three configurations spanning the design space (for example, four implants at high C/AP, four at low, six at intermediate): implants in bone-analog blocks or cadaveric mandibles, a framework fabricated from the model geometry, and strain gauges on abutments and on the analog cortical surface adjacent to the distal implant, following the approach of published in vivo and in vitro strain-gauge studies [16, 17]. The finite element model of each physical specimen is run with measured material properties, and predicted and measured strains are compared. The validation metric and acceptance threshold are fixed in advance according to the model's context of use, as required by the credibility framework for computational models of medical devices [38, 39].
Uncertainty propagation and sensitivity analysis
Because each finite element run is costly, a surrogate model (Gaussian-process or polynomial chaos expansion) is trained on a space-filling design of several hundred runs and checked by cross-validation; Sobol indices and output distributions are then computed on the surrogate, with bootstrap confidence intervals. Inputs and their ranges are listed in Table 1. H7 is evaluated on the first-order index of implant number for the primary output.
| Input | Proposed range or levels | Type |
|---|---|---|
| Implant number | 4, 5, 6, 8 | Design |
| Ratio of cantilever to antero-posterior spread | 0 to 2 | Design |
| Antero-posterior spread | 6 to 18 mm | Design and anatomy |
| Distal implant inclination | 0 to 45 degrees | Design |
| Framework elastic modulus and cross-section | Acrylic-reinforced to zirconia; two section heights | Design |
| Cortical thickness at crest | Thin to thick, from imaging population | Patient |
| Cortical and cancellous elastic moduli | Literature ranges | Patient |
| Load magnitude and direction | Distribution with dominant vertical component | Patient |
| Interface condition | Bonded; frictional (sensitivity analysis) | Modeling |
| Boundary constraints | Physiological; fixed border (sensitivity analysis) | Modeling |
7. Results of the Analytic Model and Explicitly Labeled Theoretical Predictions
Computed results of the rigid-plate model
The following values were computed for this manuscript from Equation 2 under the assumptions in Section 6.1. They are outputs of an idealized analytic model. They are not finite element results, not measurements, and not clinical data.
The normalized maximum implant force depended only on implant number and C/AP; antero-posterior spread at fixed ratio and arch half-width had no effect, as Equation 3 requires. Table 2 gives the values. For four implants, the most heavily loaded implant carried 0.95 of the applied load with no cantilever, 1.51 at C/AP of 1, and 2.08 at C/AP of 2. For six implants the values were 0.82, 1.38, and 1.94, and for eight implants 0.71, 1.22, and 1.73. At a C/AP of 1, increasing from four to six implants within the same spread reduced the maximum force by about 9 percent, and from four to eight by about 19 percent. Reducing C/AP from 1 to 0.5 with four implants reduced it by about 19 percent. The most negative (tensile) implant force for four implants was 0.71 of the applied load at C/AP of 1 and 1.27 at C/AP of 2.
In the sensitivity analysis with implant number and C/AP as inputs, the first-order index of C/AP was 0.87 and that of implant number 0.12; the index for spread at fixed ratio was zero. When cantilever length and spread were sampled as separate inputs, cantilever length had a first-order index of 0.63 and a total index of 0.71, spread 0.18 and 0.27, implant number 0.08 and 0.09, and arch width zero. The difference between first-order and total indices for cantilever and spread reflects their interaction, which is absorbed when they are combined as a ratio.
| C/AP | N = 4, max | N = 4, min | N = 6, max | N = 6, min | N = 8, max |
|---|---|---|---|---|---|
| 0 | 0.95 | −0.15 | 0.82 | −0.14 | 0.71 |
| 0.5 | 1.23 | −0.43 | 1.10 | −0.37 | 0.96 |
| 1.0 | 1.51 | −0.71 | 1.38 | −0.59 | 1.22 |
| 1.5 | 1.79 | −0.99 | 1.66 | −0.81 | 1.47 |
| 2.0 | 2.08 | −1.27 | 1.94 | −1.04 | 1.73 |
Theoretical predictions for the finite element study
The following are predictions; the finite element study has not been done.
Prediction 1. Framework compliance will increase the share of load carried by the distal implant relative to the rigid model and will reduce tensile loading of anterior implants, with the effect greatest for low-modulus frameworks and long cantilevers.
Prediction 2. The first-order index of C/AP for peak distal crestal strain will remain the largest among design variables, and that of implant number will be below 0.20 (H7).
Prediction 3. Patient variables, in particular crestal cortical thickness and load magnitude, will have indices comparable to or greater than those of design variables for bone strain, and smaller indices for screw and framework stress, which depend chiefly on design and load.
Prediction 4. Distal implant inclination will have a small first-order index once C/AP is accounted for, because its principal mechanical effect is to change C/AP; a residual effect through bending of the tilted implant will be detectable in implant-level bending moment more than in bone strain.
Prediction 5. Fixed inferior-border constraints will overstate the stiffness of the mandible and understate distal strain relative to physiological constraints.
8. Discussion: Mechanisms and Competing Explanations
The rigid-model result has a simple mechanical reading. A cantilever load must be resisted by a couple, and the arm of that couple is the antero-posterior extent of the implant group. Implants added between existing ones contribute to the equal-share term, which is small when the moment term dominates, and add little second moment because they sit near the centroid. The reduction from four to six implants, about 9 percent at C/AP of 1 in the present geometry, is real and modest. An equal reduction is obtained by shortening the cantilever by roughly a quarter of the spread. That exchange rate, an implant pair for a few millimeters of cantilever, is the kind of statement deterministic studies have not been able to make.
Whether it holds in a compliant system is uncertain, and there are specific reasons it might not. With a flexible framework, the distal implant behaves more like an isolated support, and the benefit of remote implants in resisting the moment diminishes; in that regime neither the spread nor the number of anterior implants helps much, and the dominant variables become cantilever length in absolute terms and framework stiffness. The frequency with which interim acrylic prostheses require repair [32] is what a framework-limited regime would produce, although that series does not by itself establish dependence on the cantilever ratio. Conversely, additional implants shorten the unsupported spans between implants and reduce framework bending within the arch, a benefit the rigid model cannot represent. If such effects are large, implant number could account for more variance in framework and screw stress than in distal bone strain, and the answer to "how many implants" would depend on which failure mode is at issue. H7 is stated for bone strain; a separate index is computed for each secondary output for this reason.
A competing clinical explanation for the adequacy of four implants is that loads on full-arch prostheses in edentulous patients are lower than assumed, because sensory feedback and opposing dentition limit bite force, so that any reasonable design operates below thresholds of harm and design differences are mechanically real and clinically irrelevant. The similar implant survival across implant numbers in systematic reviews [6, 8] is compatible with this. So is the systematic-review finding that overload has not been shown to cause bone loss around integrated implants [37]. The mechanical analysis would then matter mainly for the minority of patients with high forces and for technical complications, which is where clinical data point [10, 11].
Redundancy is a consideration outside the scope of load-sharing in the intact system. With four implants, the loss of one, particularly a distal one, converts the prosthesis into a long cantilever on three supports; with six, the system degrades more gracefully. A sensitivity analysis of the intact configuration does not capture this. It could be added by computing outputs for each single-implant-loss scenario and weighting by an assumed failure probability, and the protocol includes that as a secondary analysis.
The limitations of sensitivity indices should be acknowledged. They depend on the input ranges. Had the cantilever ratio been restricted to 0 to 1, as conservative practice would dictate, its index would fall and that of implant number would rise, without any change in the mechanics. The indices describe the population of designs and patients defined by the ranges and should be reported with them. They apportion variance, not harm: an input with a small index may still move a particular patient across a failure threshold. For decisions, the output distribution and the probability of exceeding a stated limit are more informative than the indices, and the protocol reports both.
The choice of output also shapes conclusions. Peak stress at a single node is mesh-dependent and has been a source of irreproducibility in the dental finite element literature [27]. Strain energy density, volume-averaged or percentile strain, and platform-level forces and moments are more robust. Platform forces and moments have the further advantage that they can be measured, which is why they anchor the validation.
9. Implications for Implant Surgery and Restorative Dentistry
The clinical inferences that the present analysis supports are limited by its idealization and are consistent with existing guidance. Cantilever length should be judged relative to antero-posterior spread, and the theoretical basis for that practice is sound. Measures that increase spread or reduce cantilever, including distal tilting where anatomy permits, address the dominant term [21, 22]; clinical meta-analyses give no overall indication that tilting harms implants, with the caveat of one maxillary subgroup finding [34, 35]. Adding implants within the same spread reduces distal load modestly in the rigid model; its main clinical justification may be redundancy and reduced framework spans, which the model does not address. A stiff definitive framework brings the system closer to the rigid idealization in which remote implants share the moment; flexible interim prostheses do not, and their cantilevers should be correspondingly short, a precaution in keeping with the repair rates reported for interim prostheses [32].
No numerical limit on cantilever ratio is proposed. The rigid model shows a smooth, approximately linear increase of distal load with the ratio and no threshold; a clinically meaningful limit would have to come from the strength of components and tissue and from the patient's forces, which the model does not contain.
Patients with signs of high occlusal force warrant a more conservative design on mechanical grounds, though trial evidence for specific design modifications is lacking.
10. Technical and Clinical Limitations
The analytic computation assumes a rigid prosthesis, identical linear implant stiffness, purely vertical load, a single arch shape, and a particular spacing rule; different spacing rules, for instance clustering implants at the extremes of the arch, would change the dependence on implant number. It yields axial implant force only. The sensitivity indices are conditional on the stated input ranges. The finite element protocol, when executed, will carry the usual limitations: linear elastic, often isotropic bone; idealized interfaces; static loading; and a validation limited to a few physical configurations and to surrogate bone. Neither model contains biology, and neither can predict bone loss or prosthesis fracture, only the mechanical quantities presumed to drive them. Maxillary arches, with different bone quality, arch form, and implant inclinations, are not specifically modeled in the screening computation. Finally, the work addresses one loading scenario, a unilateral posterior load; anterior loading, bilateral clenching, and dynamic chewing would have different distributions.
11. Proposed Experimental and Clinical Validation
Validation is hierarchical. At the first level, the analytic model is checked against the finite element model in its rigid limit, which verifies both. At the second, the finite element model is compared with strain-gauge measurements on physical specimens as described in Section 6.4. At the third, predictions are compared with in vivo abutment-level measurements where published data permit [16, 17]. At the fourth, clinical relevance is examined in a cohort: for patients with full-arch prostheses, C/AP, implant number, framework material, and opposing dentition are recorded at delivery from the digital design files, and technical complications are recorded prospectively. The model predicts that complication incidence rises with C/AP more steeply than it falls with implant number; a cohort of adequate size, analyzed with patient-level clustering and reported to STROBE, could test that. The falsification criterion for H7 itself is computational: a first-order index for implant number of 0.20 or greater in the validated model.
12. Future Research Directions
The same framework extends naturally in three directions. Patient-specific analysis, in which the surrogate model is evaluated for an individual's imaging-derived anatomy and measured bite force, would turn a population sensitivity analysis into a planning aid, subject to its own validation. Fatigue assessment of screws and frameworks requires coupling the computed load distributions with load spectra and component fatigue data. And design optimization under uncertainty, choosing implant positions that minimize the probability of exceeding a stress limit across the plausible range of patient variables, is a tractable computational problem once the surrogate exists. Measurement of functional loads on full-arch prostheses in larger numbers of patients would improve every one of these.
13. Evidence-Calibrated Conclusion
Clinical evidence indicates acceptable long-term implant survival for full-arch fixed prostheses on four or more implants and a substantial burden of technical complications; it does not identify an optimal implant number. Mechanical analysis, analytic and finite element, consistently indicates that cantilever loads are amplified at the distal implant. The computation reported here shows that, in a rigid-prosthesis idealization, the amplification depends only on implant number and the cantilever-to-spread ratio, with the ratio accounting for most of the variance over the ranges examined. This is a mathematical property of a simplified model. Whether the ranking persists for bone strain and component stress in a compliant, validated finite element model is the hypothesis, with a stated refutation criterion. Until that work and its clinical corroboration are done, the result supports the reasoning behind existing clinical guidance and should not be used to set numerical design limits.
Cite this paper
Jain S, Gupta A. Implant Number, Antero-Posterior Spread, and Cantilever Length in Full-Arch Fixed Prostheses: A Dimensionless Load-Sharing Analysis and a Finite Element Protocol With Global Sensitivity Analysis and Uncertainty Quantification. Working paper. Center for Implant Dentistry; 2026. Available from: https://bayareaimplantdentistry.com/clinician-research-and-publications/cantilever-mechanics-in-full-arch-rehabilitation
References
- Maló P, Rangert B, Nobre M. "All-on-Four" immediate-function concept with Brånemark System implants for completely edentulous mandibles: a retrospective clinical study. Clin Implant Dent Relat Res. 2003;5 Suppl 1:2-9. doi:10.1111/j.1708-8208.2003.tb00010.x (opens in a new tab) PMID: 12691645 (opens in a new tab).
- Krekmanov L, Kahn M, Rangert B, Lindström H. Tilting of posterior mandibular and maxillary implants for improved prosthesis support. Int J Oral Maxillofac Implants. 2000;15(3):405-14. PMID: 10874806 (opens in a new tab).
- Malo P, de Araújo Nobre M, Lopes A, Moss SM, Molina GJ. A longitudinal study of the survival of All-on-4 implants in the mandible with up to 10 years of follow-up. J Am Dent Assoc. 2011;142(3):310-20. doi:10.14219/jada.archive.2011.0170 (opens in a new tab) PMID: 21357865 (opens in a new tab).
- Maló P, de Araújo Nobre M, Lopes A, Ferro A, Botto J. 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. Clin Implant Dent Relat Res. 2019;21(4):565-577. doi:10.1111/cid.12769 (opens in a new tab) PMID: 30924309 (opens in a new tab).
- Soto-Penaloza D, Zaragozí-Alonso R, Penarrocha-Diago M, Penarrocha-Diago M. The all-on-four treatment concept: Systematic review. J Clin Exp Dent. 2017;9(3):e474-e488. doi:10.4317/jced.53613 (opens in a new tab) PMID: 28298995 (opens in a new tab).
- Heydecke G, Zwahlen M, Nicol A, Nisand D, Payer M, Renouard F, et al. What is the optimal number of implants for fixed reconstructions: a systematic review. Clin Oral Implants Res. 2012;23 Suppl 6:217-28. doi:10.1111/j.1600-0501.2012.02548.x (opens in a new tab) PMID: 23062144 (opens in a new tab).
- Mericske-Stern R, Worni A. Optimal number of oral implants for fixed reconstructions: a review of the literature. Eur J Oral Implantol. 2014;7 Suppl 2:S133-53. PMID: 24977248 (opens in a new tab).
- Daudt Polido W, Aghaloo T, Emmett TW, Taylor TD, Morton D. Number of implants placed for complete-arch fixed prostheses: A systematic review and meta-analysis. Clin Oral Implants Res. 2018;29 Suppl 16:154-183. doi:10.1111/clr.13312 (opens in a new tab) PMID: 30328199 (opens in a new tab).
- Brånemark PI, Svensson B, van Steenberghe D. Ten-year survival rates of fixed prostheses on four or six implants ad modum Brånemark in full edentulism. Clin Oral Implants Res. 1995;6(4):227-31. doi:10.1034/j.1600-0501.1995.060405.x (opens in a new tab) PMID: 8603114 (opens in a new tab).
- Bozini T, Petridis H, Garefis K, Garefis P. A meta-analysis of prosthodontic complication rates of implant-supported fixed dental prostheses in edentulous patients after an observation period of at least 5 years. Int J Oral Maxillofac Implants. 2011;26(2):304-18. PMID: 21483883 (opens in a new tab).
- Papaspyridakos P, Chen CJ, Chuang SK, Weber HP, Gallucci GO. A systematic review of biologic and technical complications with fixed implant rehabilitations for edentulous patients. Int J Oral Maxillofac Implants. 2012;27(1):102-10. PMID: 22299086 (opens in a new tab).
- Pjetursson BE, Thoma D, Jung R, Zwahlen M, Zembic A. A systematic review of the survival and complication rates of implant-supported fixed dental prostheses (FDPs) after a mean observation period of at least 5 years. Clin Oral Implants Res. 2012;23 Suppl 6:22-38. doi:10.1111/j.1600-0501.2012.02546.x (opens in a new tab) PMID: 23062125 (opens in a new tab).
- Skalak R. Biomechanical considerations in osseointegrated prostheses. J Prosthet Dent. 1983;49(6):843-8. doi:10.1016/0022-3913(83)90361-x (opens in a new tab) PMID: 6576140 (opens in a new tab).
- Rangert B, Jemt T, Jörneus L. Forces and moments on Branemark implants. Int J Oral Maxillofac Implants. 1989;4(3):241-7. PMID: 2700747 (opens in a new tab).
- Brunski JB, Puleo DA, Nanci A. Biomaterials and biomechanics of oral and maxillofacial implants: current status and future developments. Int J Oral Maxillofac Implants. 2000;15(1):15-46. PMID: 10697938 (opens in a new tab).
- Duyck J, Van Oosterwyck H, Vander Sloten J, De Cooman M, Puers R, Naert I. Magnitude and distribution of occlusal forces on oral implants supporting fixed prostheses: an in vivo study. Clin Oral Implants Res. 2000;11(5):465-75. doi:10.1034/j.1600-0501.2000.011005465.x (opens in a new tab) PMID: 11168239 (opens in a new tab).
- Ogawa T, Dhaliwal S, Naert I, Mine A, Kronstrom M, Sasaki K, et al. Impact of implant number, distribution and prosthesis material on loading on implants supporting fixed prostheses. J Oral Rehabil. 2010;37(7):525-31. doi:10.1111/j.1365-2842.2010.02076.x (opens in a new tab) PMID: 20236236 (opens in a new tab).
- Geng JP, Tan KB, Liu GR. Application of finite element analysis in implant dentistry: a review of the literature. J Prosthet Dent. 2001;85(6):585-98. doi:10.1067/mpr.2001.115251 (opens in a new tab) PMID: 11404759 (opens in a new tab).
- Korioth TW, Versluis A. Modeling the mechanical behavior of the jaws and their related structures by finite element (FE) analysis. Crit Rev Oral Biol Med. 1997;8(1):90-104. doi:10.1177/10454411970080010501 (opens in a new tab) PMID: 9063627 (opens in a new tab).
- Sertgöz A, Güvener S. Finite element analysis of the effect of cantilever and implant length on stress distribution in an implant-supported fixed prosthesis. J Prosthet Dent. 1996;76(2):165-9. doi:10.1016/s0022-3913(96)90301-7 (opens in a new tab) PMID: 8820808 (opens in a new tab).
- Zampelis A, Rangert B, Heijl L. Tilting of splinted implants for improved prosthodontic support: a two-dimensional finite element analysis. J Prosthet Dent. 2007;97(6 Suppl):S35-43. doi:10.1016/S0022-3913(07)60006-7 (opens in a new tab) PMID: 17618932 (opens in a new tab).
- Bellini CM, Romeo D, Galbusera F, Taschieri S, Raimondi MT, Zampelis A, et al. Comparison of tilted versus nontilted implant-supported prosthetic designs for the restoration of the edentuous mandible: a biomechanical study. Int J Oral Maxillofac Implants. 2009;24(3):511-7. PMID: 19587875 (opens in a new tab).
- Bevilacqua M, Tealdo T, Menini M, Pera F, Mossolov A, Drago C, et al. The influence of cantilever length and implant inclination on stress distribution in maxillary implant-supported fixed dentures. J Prosthet Dent. 2011;105(1):5-13. doi:10.1016/S0022-3913(10)60182-5 (opens in a new tab) PMID: 21194582 (opens in a new tab).
- Takahashi T, Shimamura I, Sakurai K. Influence of number and inclination angle of implants on stress distribution in mandibular cortical bone with All-on-4 Concept. J Prosthodont Res. 2010;54(4):179-84. doi:10.1016/j.jpor.2010.04.004 (opens in a new tab) PMID: 20452854 (opens in a new tab).
- Silva GC, Mendonça JA, Lopes LR, Landre J Jr. Stress patterns on implants in prostheses supported by four or six implants: a three-dimensional finite element analysis. Int J Oral Maxillofac Implants. 2010;25(2):239-46. PMID: 20369081 (opens in a new tab).
- Fazi G, Tellini S, Vangi D, Branchi R. Three-dimensional finite element analysis of different implant configurations for a mandibular fixed prosthesis. Int J Oral Maxillofac Implants. 2011;26(4):752-9. PMID: 21841984 (opens in a new tab).
- Chang Y, Tambe AA, Maeda Y, Wada M, Gonda T. Finite element analysis of dental implants with validation: to what extent can we expect the model to predict biological phenomena? A literature review and proposal for classification of a validation process. Int J Implant Dent. 2018;4(1):7. doi:10.1186/s40729-018-0119-5 (opens in a new tab) PMID: 29516219 (opens in a new tab).
- Aglietta M, Siciliano VI, Zwahlen M, Brägger U, Pjetursson BE, Lang NP, et al. A systematic review of the survival and complication rates of implant supported fixed dental prostheses with cantilever extensions after an observation period of at least 5 years. Clin Oral Implants Res. 2009;20(5):441-51. doi:10.1111/j.1600-0501.2009.01706.x (opens in a new tab) PMID: 19522975 (opens in a new tab).
- Romeo E, Storelli S. Systematic review of the survival rate and the biological, technical, and aesthetic complications of fixed dental prostheses with cantilevers on implants reported in longitudinal studies with a mean of 5 years follow-up. Clin Oral Implants Res. 2012;23 Suppl 6:39-49. doi:10.1111/j.1600-0501.2012.02551.x (opens in a new tab) PMID: 23062126 (opens in a new tab).
- Shackleton JL, Carr L, Slabbert JC, Becker PJ. Survival of fixed implant-supported prostheses related to cantilever lengths. J Prosthet Dent. 1994;71(1):23-6. doi:10.1016/0022-3913(94)90250-x (opens in a new tab) PMID: 8120840 (opens in a new tab).
- Gallucci GO, Doughtie CB, Hwang JW, Fiorellini JP, Weber HP. Five-year results of fixed implant-supported rehabilitations with distal cantilevers for the edentulous mandible. Clin Oral Implants Res. 2009;20(6):601-7. doi:10.1111/j.1600-0501.2008.01699.x (opens in a new tab) PMID: 19302389 (opens in a new tab).
- Drago C. Cantilever Lengths and Anterior-Posterior Spreads of Interim, Acrylic Resin, Full-Arch Screw-Retained Prostheses and Their Relationship to Prosthetic Complications. J Prosthodont. 2017;26(6):502-507. doi:10.1111/jopr.12426 (opens in a new tab) PMID: 26848820 (opens in a new tab).
- Drago C. Ratios of Cantilever Lengths and Anterior-Posterior Spreads of Definitive Hybrid Full-Arch, Screw-Retained Prostheses: Results of a Clinical Study. J Prosthodont. 2018;27(5):402-408. doi:10.1111/jopr.12519 (opens in a new tab) PMID: 27416302 (opens in a new tab).
- Chrcanovic BR, Albrektsson T, Wennerberg A. Tilted versus axially placed dental implants: a meta-analysis. J Dent. 2015;43(2):149-70. doi:10.1016/j.jdent.2014.09.002 (opens in a new tab) PMID: 25239770 (opens in a new tab).
- Monje A, Chan HL, Suarez F, Galindo-Moreno P, Wang HL. Marginal bone loss around tilted implants in comparison to straight implants: a meta-analysis. Int J Oral Maxillofac Implants. 2012;27(6):1576-83. PMID: 23189313 (opens in a new tab).
- Frost HM. A 2003 update of bone physiology and Wolff's Law for clinicians. Angle Orthod. 2004;74(1):3-15. doi:10.1043/0003-3219(2004)074<0003:AUOBPA>2.0.CO;2 (opens in a new tab) PMID: 15038485 (opens in a new tab).
- Naert I, Duyck J, Vandamme K. Occlusal overload and bone/implant loss. Clin Oral Implants Res. 2012;23 Suppl 6:95-107. doi:10.1111/j.1600-0501.2012.02550.x (opens in a new tab) PMID: 23062133 (opens in a new tab).
- American Society of Mechanical Engineers. ASME V&V 40-2018: Assessing credibility of computational modeling through verification and validation: application to medical devices. New York: ASME; 2018. †
- U.S. Food and Drug Administration. Assessing the credibility of computational modeling and simulation in medical device submissions: guidance for industry and Food and Drug Administration staff. Silver Spring (MD): FDA; 2023. †
- Sobol IM. Global sensitivity indices for nonlinear mathematical models and their Monte Carlo estimates. Math Comput Simul. 2001;55(1-3):271-80. †
- Saltelli A, Annoni P, Azzini I, Campolongo F, Ratto M, Tarantola S. Variance based sensitivity analysis of model output. Design and estimator for the total sensitivity index. Comput Phys Commun. 2010;181(2):259-70. †
- Erdemir A, Guess TM, Halloran J, Tadepalli SC, Morrison TM. Considerations for reporting finite element analysis studies in biomechanics. J Biomech. 2012;45(4):625-33. doi:10.1016/j.jbiomech.2011.11.038 (opens in a new tab) PMID: 22236526 (opens in a new tab).
- Viceconti M, Olsen S, Nolte LP, Burton K. Extracting clinically relevant data from finite element simulations. Clin Biomech (Bristol). 2005;20(5):451-4. doi:10.1016/j.clinbiomech.2005.01.010 (opens in a new tab) PMID: 15836931 (opens in a new tab).
- Reference verification note. 39 of 43 references are PubMed-indexed; their bibliographic records (authors, title, journal, year, volume, pages, DOI, PMID) were retrieved programmatically from the PubMed database on November 17, 2026. † Source not indexed in PubMed (book, standard, guidance or non-indexed journal); bibliographic details are cited from the published record and were not database-verified in this drafting session. Database retrieval confirms that each record exists as cited; it does not replace full-text reading by the responsible authors of every statement attributed to a source.

