One Displacement Field, Two Complications: A Kinematic Hypothesis Linking Interproximal Contact Loss and Implant Infraposition
- Paper type
- Hypothesis and Theory
- Status
- Original Work · Working Paper
- Keywords:
- interproximal contact loss · implant infraposition · mesial drift · continuous eruption · intraoral scanning · kinematics · dental implants
Contents
Abstract
Background. Loss of the interproximal contact between an implant restoration and an adjacent natural tooth is common, and so is infraposition of the implant crown relative to neighboring teeth. Each complication has its own literature, its own list of proposed predictors, and no tested mechanism. Reported predictors conflict between cohorts. Existing reviews attribute both to lifelong growth, continued eruption, and mesial drift, but the link is stated verbally and has not been formalized or tested within patients.
Objective and hypothesis. We propose that both complications are projections of one phenomenon: the three-dimensional displacement of the adjacent natural dentition relative to an ankylosed implant. The mesio-distal component of that displacement opens the contact once it exceeds the contact’s elastic tolerance; the vertical component is the infraposition. We therefore predict that the two components co-vary within patients where displacement direction is similar across patients, and that they share predictors once displacement is measured directly instead of inferred from age or sex.
Proposed approach. We specify a kinematic model in an implant-fixed frame, including a sign convention for mesial and distal contacts and a geometric coupling term. We give illustrative time-to-opening calculations under assumed drift rates. We propose serial intraoral scans registered on the implant restoration, with baseline contact gauging and digital occlusal recording, and we state six numbered predictions that would falsify the shared-mechanism account. A Monte Carlo illustration shows that correlation can vanish even when displacement magnitude is fully shared, if displacement direction varies enough between patients; falsification therefore requires an estimate of direction dispersion.
Expected contribution. The paper supplies a quantitative formulation and explicit refutation conditions for an idea that is already present in qualitative form in earlier reviews. It does not claim priority for the common-cause concept.
Conclusions. The hypothesis is testable with existing scanning technology, but its anterior and posterior evidence bases do not overlap, scanner precision at the required scale has not been verified here, and no data support the within-patient correlation. All numerical examples are assumption-driven illustrations.
1. Introduction
An osseointegrated implant behaves, mechanically, like an ankylosed tooth: it does not erupt, drift, or follow the jaw as the dentition remodels. Natural teeth do. The adjacent dentition therefore moves relative to the implant for as long as the patient lives, and the implant restoration, which looked correct on the day of delivery, gradually acquires a different relationship to its neighbors [1, 2].
Two clinical consequences of this relative motion are well described. The first is loss of the interproximal contact. Meta-analyses place its prevalence in the range of roughly one third to one half of sites, with the mesial contact affected more often than the distal contact and prevalence rising with time in service [3–6]. The second is infraposition, in which the implant crown sits increasingly apical to the incisal or occlusal plane of the neighboring teeth. A systematic review of 27 nonrandomized studies reported a pooled infraposition prevalence of about half of implants and a mean extent of about 0.6 mm, with certainty rated very low [7].
These complications are studied in separate literatures. Contact loss is studied mostly in fixed implant prostheses, largely posterior in the cohorts we could characterize, by feeler-gauge or strip methods [8–10]. Infraposition is studied mostly at single anterior maxillary implants by photographs, casts, radiographs, or cephalograms [11–14]. The proposed explanations also differ: anterior force and mesial drift for contacts [15], late facial growth and continued eruption for infraposition [16, 17]. Yet the physical situation is the same. A rigid reference sits in an arch of mobile teeth, and the teeth move.
This paper asks whether the two complications can be treated as two components of one displacement vector, and what data would show that they cannot. The question is narrower than it may first appear. We do not claim that nobody has suggested a common cause. Section 2 shows that several authors have. We claim only that no source seen in this search formalizes the relation, derives conditions under which co-variation should or should not appear, or measures both components at the same site in the same patients. The contribution is therefore a specification, not a discovery.
The evidence base was appraised at the level of PubMed abstracts retrieved on 17 November 2026. Full texts were not read. Embase, Scopus, and Web of Science were not accessible. Statements about study design, sample size, and bias are drawn from abstracts and may be incomplete.
2. Background and critical appraisal of the evidence
Origins of the displacement concept
Mesial drift, the tendency of teeth to move toward the midline over time, has a long experimental history. Studies in ankylosed deciduous molars were used to estimate the drift of neighboring human teeth [18], and work in adult monkeys examined drift after the tongue and cheek forces had been removed [19]. The idea that ankylosed units reveal movement of their neighbors was thus used well before dental implants. Studies in growing pigs examined the effect of osseointegrated implants on dentoalveolar development [20], and a review of growing patients later synthesized the clinical consequences of placing implants before growth is complete [21].
For adults, the key conceptual step was the recognition that craniofacial remodeling and tooth eruption do not stop at skeletal maturity. A review of facial development, continuous eruption, and mesial drift as compromising factors for implant placement made this argument explicitly: implants, like ankylosed teeth, retain their position while the surrounding structures change [2]. An observational series of long-term implant cases described opened contacts due to tooth migration and progressive discrepancy of the gingival margin as consequences of continued adult growth, while noting that the changes seemed random to the authors [22]. A recent critical review states that late growth and continuous eruption contribute to spatial discrepancies between implants and adjacent dentition in skeletally mature adults, with infraocclusion multifactorial [16]. Another recent review examines proximal contact loss and infraocclusion across age groups and stresses managing growth-related changes through timing of placement and attention to patient age and gender [23]. A review of open contacts lists tooth-migration, crown-related, and bone-growth-related causes [24].
These sources establish that the common-cause idea is not new. What they do not provide is a model that states what quantity is shared, how it maps to each complication, and what observation would contradict the mapping.
Contact loss: frequency and proposed mechanisms
Several meta-analyses agree that contact loss is frequent. One reported mesial loss in 44.2% and distal loss in 27.5% of contacts [3]; others reached similar conclusions about prevalence [4, 5]. A qualitative review judged contact loss to be frequent, progressive, and in practice inevitable, with duration of service and mesial location as the clearest factors and occlusion, adjacent tooth vitality, implant location, age, and splinting as less clear [6]. A pooled analysis of nonrandomized studies reported a contact-loss prevalence of 46.3%, an increase with observation time, and a mesial predilection, but rated the evidence very low [7].
Individual cohorts illustrate both the consistency and the disagreement. A Japanese cohort of 146 prostheses in 105 patients, using a 50 micrometer gauge, classified 43% of contacts as lost, more often mesial than distal, and found that age, opposing dentition, vitality of the adjacent tooth, and splinting influenced mesial loss [8]. A Korean prospective study of 234 posterior prostheses reported loss at 59.9% of 299 contacts and identified adjacent bone level, maxillary position, and mesial site as significant [9]. A cross-sectional study of 94 patients found loss at 34% of embrasures, increasing with follow-up, with follow-up duration the only multivariable factor and no demonstrable effect on periodontal or peri-implant tissue, although food impaction was more frequent [10]. A 10-year retrospective study reported loss at half of sites, with deeper probing [25]. A retrospective study of 83 restorations found that a night occlusal device did not prevent loss [26], and a recent cohort again examined risk factors [27]. Whether open contacts harm peri-implant tissue is disputed: one study found no tissue effect [10], while a recent meta-analysis found more peri-implant disease with open contacts [28].
The mechanistic explanation most often offered for contact loss is mesial drift driven by the anterior component of occlusal force. That component was measured in classic work [29], and a more recent direct measurement in 30 young adults reported mean values near 21 N during maximal biting, positively related to bite force [15]. The same paper stressed that older explanations do not account for the concurrent rise of contact tightness at anterior and posterior contacts [15]. The only study in this search that tied the force explanation to implant patients was a preliminary 28-patient survey using three-dimensional occlusal analysis, which found higher anterior and lingual force components in patients with contact loss [30]. It is cross-sectional, small, and cannot separate cause from consequence.
Two observations strain a pure mesial-drift account. First, distal contacts also open, yet a tooth distal to the implant that drifts mesially would press against the implant crown and tighten the contact. Second, the preventive trial that would test force explanations has not been done: ClinicalTrials.gov searches with the key terms returned no trial of prevention of contact loss.
Infraposition: frequency and contested predictors
Infraposition is documented in mature adults. A retrospective study of the anterior maxilla reported vertical steps of 0.12 to 1.86 mm in mature adults [11]. A cohort of 28 implant crowns in 25 patients followed for about 16 years reported clinical stability at 11 of 20 sites in males and at none in females, and linked this to greater increases in anterior face height and posterior mandibular rotation in a female reference group [12]. In a 57-patient study at 17 to 19 years, women were at higher risk and no association with age or facial shape was found [13]. In 76 adults followed for up to 15 years, radiographic infraocclusion increased by about 0.08 mm per year, more in younger adults and after delayed placement, with no sex effect [14]. A study of continuous craniofacial growth in adults found high prevalence and no identifiable risk factor [31], and a cross-sectional study found the discrepancy independent of sex and age [32]. A prospective study of up to 6 years reported no infraposition at about half of implants and small changes in most others [33], and a retrospective study at 11 to 21 years reported a high proportion with infraposition while satisfaction stayed high [34].
A three-dimensional study of 30 patients at 14 to 20 years measured adjacent tooth movement of 0 to 2.5 mm, with incisal and palatal components most pronounced, and associated incisal movement above 1 mm with a lower anterior face height of 70 mm or more [35]. This is the study closest to a displacement-vector approach, but it did not measure contacts.
The independent eruption evidence supports the physical premise. In 24 women followed for 10 years with digital cast superimposition on the palate and rugae, maxillary anterior teeth erupted by about 0.3 mm on average [17]. A 50-year digital study of 30 untreated subjects from age 13 to 62 reported significant decreases in arch length, decreases in overbite, and increases in clinical crown height [36]. These data show that natural arches change after maturity in the directions the hypothesis requires, including a reduction of arch length that is consistent with mesial movement of teeth.
What the appraisal shows
Table 1 collects the pivotal sources. Three points matter for the hypothesis.
First, the two outcomes have been reported together in one meta-analysis [7], but as separate pooled estimates. The review did not test whether patients with one outcome more often had the other. Second, the proposed predictors of each outcome are inconsistent, as the sex, age, and face-height findings above show. Inconsistent predictors are what one expects when the true driver (displacement) is unmeasured and age, sex, and facial type are noisy proxies for it. This is an argument of plausibility, not proof. Third, the two evidence bases sit at different sites. Contact loss is mostly reported for posterior prostheses; infraposition for anterior maxillary single crowns. A within-patient test of the hypothesis is therefore feasible mainly in the anterior maxilla, where both outcomes occur, or in designs that measure vertical step at posterior sites.
| Source | Design and size (from abstract) | Finding relevant here | Main bias | Level |
|---|---|---|---|---|
| Papageorgiou 2018 [7] | Meta-analysis, 27 nonrandomized studies, 1,572 patients | Infraposition about 50%; contact loss 46.3%; very low certainty | Heterogeneity, nonrandomized | A |
| Sheba 2024 [3] | Systematic review and meta-analysis | Mesial 44.2%, distal 27.5% | Variable contact definitions | A |
| Koori 2010 [8] | Cohort, 146 prostheses, 105 patients | 43% of contacts lost; mesial > distal | Gauge-based; retrospective follow-up visits | B |
| Pang 2017 [9] | Prospective, 234 prostheses | 59.9% of contacts lost; bone level, maxilla, mesial | Posterior only | B |
| Wei 2008 [30] | Cross-sectional, 28 patients | Higher anterior and lingual force in loss group | Small, cross-sectional | C |
| Jemt 2007 [12] | Cohort, 25 patients, 28 crowns | Movement more frequent in females; link to face-height change | Clinical index, no scans | B |
| Polymeri 2020 [14] | Retrospective, 76 adults | About 0.08 mm/year infraocclusion | Two-dimensional radiographs | B |
| Winitsky 2021 [35] | Retrospective, 30 patients, 3-D | 0-2.5 mm movement; incisal and palatal | Small, no contacts | B |
| Huanca Ghislanzoni 2017 [17] | Longitudinal, 24 women | Anterior eruption about 0.3 mm in 10 years | Women only, cast registration | B |
| Daftary 2013 [22] | Observational case series | Opened contacts and misalignment ascribed to adult growth | Selected cases | D |
Note. Evidence levels are operational categories used in this manuscript, not GRADE: A strong (multiple consistent studies or pooled analysis), B moderate (single cohort), C preliminary, D mechanistic or theoretical, E speculative hypothesis. Levels reflect abstract-level reading only.
3. Hypothesis and theoretical framework
Formal statement
Hypothesis H. Let an implant crown be rigidly fixed to bone and let an adjacent natural tooth move over time relative to it. Define the displacement of a reference point on the adjacent tooth, expressed in a coordinate frame fixed to the implant, as the vector
d(t)=(d_x(t), d_y(t), d_z(t))
with all components in micrometers and t in years since crown delivery. The x axis runs mesio-distally along the arch tangent at the contact, positive toward mesial; y runs bucco-lingually, positive toward buccal; z runs along the implant axis, positive toward occlusal. Then:
- The infraposition of the implant crown relative to the adjacent tooth is the vertical component net of differential wear,
I(t)=d_z(t)−Δw(t),
where Δw(t) is the net differential occlusal or incisal wear in micrometers (zero if wear is negligible).
- The signed gap at the contact on side k, with k=M for the mesial contact and k=D for the distal contact, is
g_k(t)=g_(0,k)+s_k d_(x,k)(t)+κ_k d_(z,k)(t),
where g_(0,k) is the gap at delivery (negative when the contact is compressed), s_M=+1 and s_D=−1, and κ_k is a dimensionless geometric coupling coefficient equal to the tangent of the local slope of the proximal surfaces relative to the implant axis (its sign depends on whether the vertical movement slides the contact toward a wider or narrower part of the embrasure).
- A contact is clinically “lost” when the gap reaches the gauge threshold g^* (50 micrometers in two of the cohorts reviewed above [8, 9]). This is equivalent to
s_k d_(x,k)(t)+κ_k d_(z,k)(t)≥τ_k, τ_k=g^*−g_(0,k),
where τ_k in micrometers is the contact’s elastic tolerance: the compression at delivery that the displacement must first remove, plus the gauge threshold.
The sign convention matters. A tooth mesial to the implant that moves mesially moves away from the implant crown and opens the contact (s_M=+1). A tooth distal to the implant that moves mesially presses on the crown and tightens the contact (s_D=−1). Pure mesial drift therefore predicts mesial contact opening and distal contact closure. The observation that distal contacts open too (27.5% in one meta-analysis [3]) does not refute hypothesis H, because H does not assume that displacement is purely mesial. It does require that distal opening be accompanied by distally directed displacement of the distal tooth, or else by a change on the restoration side (Section 5).
Statement of the shared-mechanism claim. Both I(t) and the contact gap depend on the same vector d(t). If displacement magnitude differs between patients for reasons common to both components (for example, patient-level eruption and drift tendency, residual growth, or occlusal loading), then d_x and d_z will co-vary between patients, and predictors of one will predict the other.
Assumptions
The formulation rests on explicit assumptions, each of which can fail.
- A1, rigid reference. The implant does not move relative to the bone that carries the registration reference. Implants in growing patients do not satisfy this [20, 21]; in adults it is an approximation.
- A2, small displacements. Displacements are small compared with tooth dimensions, so that rotations can be ignored. Tipping would add a rotation term and a point-dependent displacement; we track a single reference point per tooth.
- A3, linear superposition. The gap depends linearly on d_x and d_z at a given contact. This is a first-order approximation of the proximal geometry.
- A4, wear is measurable or negligible. Differential wear of the crown and the neighbor is either measured or small over the observation period.
- A5, restoration stability. The crown itself does not change shape or position relative to the implant platform. Abutment settling, screw loosening, and contact wear violate this. Axial settling of conical connections is documented on the bench and shifted crowns apically while tightening contacts after torquing [37], but a clinical study found that relative occlusal force on implant crowns kept rising for 24 months [38]. A5 is therefore testable, not assumed true.
Illustrative computation under stated assumptions
Illustrative computation under stated assumptions — not empirical data. The script is saved as sim.py in the paper folder and its output as sim_output.txt. No drift rate in the sources seen for this paper is specified for the mesio-distal direction at implant-adjacent teeth. All mesio-distal rates below are assumed values chosen to span a plausible range for illustration. The vertical rates (0.03 and 0.08 mm per year) are taken from the cast-superimposition and radiographic studies cited above, which reported about 0.3 mm of anterior eruption in 10 years and about 0.08 mm per year of infraocclusion [14, 17].
Time to contact opening. With constant rates and no coupling (κ_k=0), a mesial contact opens at
t^*=τ/r_x,
where r_x is the mesial rate in micrometers per year. The script computed t^* for assumed tolerances τ of 50, 100, and 150 micrometers and assumed rates r_x of 10, 20, 40, and 80 micrometers per year (Table 2). For example, a tolerance of 100 micrometers and a rate of 20 micrometers per year gives 5.0 years. A distal contact with the same mesial drift would not open at all under this model.
| Tolerance τ (micrometers) | rx = 10 micrometers/year | 20 | 40 | 80 |
|---|---|---|---|---|
| 50 | 5.00 | 2.50 | 1.25 | 0.62 |
| 100 | 10.00 | 5.00 | 2.50 | 1.25 |
| 150 | 15.00 | 7.50 | 3.75 | 1.88 |
Two features of this table are worth stating, though neither is empirical. First, the spread in t^* across plausible inputs is more than tenfold, which is enough to produce the cohort-to-cohort differences in time-dependent prevalence even if the underlying drift were identical. Second, the tolerance τ is not directly observable from a feeler-gauge check at one visit; it requires the delivery-time contact tightness and a displacement record. That is a reason for baseline measurement in the proposed protocol.
Vertical component. At the cited rates, 10 years of eruption would give 0.3 and 0.8 mm of vertical displacement (the script’s arithmetic: 0.03×10 and 0.08×10). If the same patient also had an assumed mesial rate of 40 micrometers per year and a tolerance of 100 micrometers, the contact would open at 2.5 years, when the vertical step would be only 75 micrometers at an assumed 30 micrometers per year. The model therefore predicts that contact loss usually becomes detectable before the vertical step is visible, a testable ordering (P4 below), which depends entirely on the assumed rates.
Why correlation may vanish even when the mechanism is shared. A Monte Carlo illustration (200,000 simulated patients; random seed 25) varied the structure of the displacement field. The assumed inputs were: patient-level displacement intensity lognormal with log-standard deviation 0.6; mean mesial scale 40 and mean vertical scale 50 micrometers per year over 10 years; and independent lognormal site noise. Three results follow.
- With a fully shared intensity, the correlation between d_x and d_z was 0.83. With independent intensities it was 0.00. Adding Gaussian measurement error of 100 micrometers (standard deviation) lowered the observed correlation of the shared case from 0.83 to 0.77, and error of 30 micrometers left it at 0.82.
- If displacement magnitude was shared but the direction angle θ (in the vertical-mesial plane) varied between patients around 45 degrees, the correlation fell sharply with the angular standard deviation: 0.95 at 5 degrees, 0.63 at 15 degrees, 0.10 at 30 degrees, and −0.16 at 60 degrees.
- Hence a near-zero observed correlation is compatible with a perfectly shared mechanism if direction varies by about 30 degrees or more between patients, and the shared-mechanism hypothesis cannot be rejected on correlation alone. The proposal therefore makes the test conditional on estimated direction dispersion (P1) and adds tests that do not depend on it (P2, P3).
The simulation is a statement about the model’s logic, not about patients. Its parameters were not estimated from any source.
Numbered, falsifiable predictions
The following predictions apply to the proposed cohort of Section 4. Each states the result that would refute it.
P1 (conditional within-patient co-variation). In the anterior maxillary stratum, the mesio-distal displacement at the mesial contact (d_x) and the vertical displacement (d_z) of the same adjacent tooth are positively correlated across patients, with a magnitude of at least about 0.3, provided that the estimated circular standard deviation of the direction angle is below about 20 degrees. Refuted if the 95% confidence interval for the correlation lies entirely below 0.2 while the estimated dispersion is below 20 degrees.
P2 (shared predictors). Candidate predictors (age, sex, facial height, bite force, time since crown delivery) act on d_x and d_z through a common latent factor. A two-outcome model with one latent factor fits better than a model with independent factors. Refuted if the independent-factor model fits at least as well by a prespecified information criterion, or if the predictor effects on d_x and d_z have opposite signs with non-overlapping confidence intervals.
P3 (patient-level displacement tendency). Displacement of natural teeth in the same arch, not adjacent to any implant (for example, the homologous contralateral tooth), correlates positively with the displacement of the implant-adjacent tooth. Refuted if that correlation is null with a confidence interval excluding 0.3.
P4 (tolerance and ordering). Time to contact loss increases with delivery-time contact tightness after adjustment for the measured d_x rate, and contact loss at the anterior mesial site precedes a vertical step of 100 micrometers in most patients in whom both occur. Refuted if delivery-time tightness is unrelated to time to loss after adjustment, or if the vertical step typically precedes contact loss.
P5 (the contact follows d_x, not d_z). After adjusting for d_x, d_z adds little to the explained variance of the contact gap (small κ_k). This discriminates a shared biological driver from a purely geometric coupling. Refuted if d_z explains the gap better than d_x, which would mean that the contact opens mostly because of the proximal geometry.
P6 (sign of distal opening). Distal contacts that open are accompanied by distally directed displacement of the distal tooth in the implant frame, or by an identified restoration-side change. Refuted if distal contacts open while the distal tooth moves mesially in the implant frame and no restoration-side change is found, because the model would then have no mechanism for opening.
An additional exploratory prediction links the first mesio-distal result to the force explanation: baseline anterior force components from digital occlusal recording [15, 30] predict the rate of d_x. This is exploratory because the supporting evidence is a single small cross-sectional survey.
4. Proposed methodology
No data have been collected. This section specifies a study that would test hypothesis H. All sample sizes are computed from stated assumptions, not from pilot data.
Design and population
We propose a prospective observational cohort with annual follow-up for 5 years, nested in routine implant maintenance, with an optional retrospective arm that uses archived intraoral scans from clinics that already run digital workflows. Participants would be adults with a single-tooth implant restoration in contact with at least one natural tooth, recruited at crown delivery. Two strata are needed because the existing evidence bases occupy different sites. The anterior maxillary stratum (incisor to canine positions) supplies both outcomes at the same tooth and is the primary test. The posterior stratum supplies contact data and a vertical step at the occlusal surface, and tests whether the model holds beyond the esthetic zone.
Exclusion criteria would include planned or ongoing orthodontic treatment, planned restoration of the adjacent tooth, splinting of the adjacent tooth to the implant, and implant placement before skeletal maturity, since in growing patients the implant itself is displaced relative to the skeleton [20, 21]. Periodontal status, adjacent tooth vitality, and splinting of adjacent teeth would be recorded as covariates because earlier cohorts found them associated with contact loss [8, 9].
Measurement of displacement
The proposal depends on one idea: the implant restoration is a stable reference. Each visit would produce two intraoral scans of the full arch by the same operator and scanner. The scans would be registered to the baseline scan by best-fit alignment restricted to the surface of the implant restoration and its scan-derived platform geometry, and not to the teeth under study. Displacement of each adjacent tooth would then be read as the shift of anatomical landmarks (the contact point, the incisal edge or cusp tip, and the gingival margin) in the implant-fixed frame of Section 3.1. Scan superimposition has been used for single-tooth monitoring in a recent case report of an ankylosed canine [39], and digital cast superimposition on the palate has been used to quantify adult tooth eruption [17]. Neither is a validation of the method at implant restorations.
Three validity problems must be handled before the method can carry the study.
- Reference stability. The crown surface can wear or be replaced; the abutment can settle [37]. A secondary registration on stable, non-implant structures such as the palatal rugae [17] would be used as a sensitivity analysis, and any divergence between the two registrations would be reported as a reference-stability check, not averaged away.
- Measurement precision. The scanner’s reproducibility for this registration was not determined from the sources read for this paper. We therefore specify a reproducibility substudy: duplicate scans at each visit in all participants, giving a standard error of measurement (SEM) for each component. The minimum detectable change is then MDC_(95)=1.96√(2) SEM. If the MDC exceeds the expected displacement over the scan interval, the interval must be lengthened or the component dropped from analysis.
- Contact gap definition. The primary contact outcome is gauge-based, with a 50 micrometer threshold, to match two earlier cohorts [8, 9]. A scan-derived minimum interproximal distance is a secondary, continuous outcome. The two must be reconciled in a calibration substudy.
At baseline we would also record: contact tightness at delivery by gauge (for g_(0,k)), a digital occlusal force recording including its anterior and lateral components [15, 30], bone level on the adjacent tooth, and anterior facial height from an existing lateral cephalogram or equivalent clinical record. No radiation would be added for research purposes.
Controls
Homologous contralateral natural-tooth pairs, where available, supply an internal control for natural drift without an implant in the contact (P3). A second control is the registration contrast of Section 4.2, which separates implant-side change from arch-side change. Patients with a stable, tight contact at the last visit serve as the comparison group for survival analysis of contact loss.
Sample size reasoning
The primary hypothesis is a correlation. For a target correlation ρ, two-sided significance level α, and power 1−β, the Fisher-z approximation gives
n=((z_(1−α/2)+z_(1−β))/(artanh(ρ)))^2+3.
With α=0.05 and power 80%, the script computed n=194 for ρ=0.2, 85 for ρ=0.3, 47 for ρ=0.4, and 30 for ρ=0.5; with power 90% the values were 259, 113, 62, and 38. We propose ρ=0.3 as the smallest correlation of interest, giving 85 complete anterior-stratum patients at 80% power. With 20% attrition, enrolment would be 107 (85/0.8 rounded up). P1’s refutation rule requires a confidence interval whose upper bound lies below 0.2; its precision is greater than the Fisher-z calculation for a single test requires, and a pilot would be needed to settle on the final enrollment.
Measurement error lowers the observed correlation by roughly √(R_xR_z), where R_x and R_z are the reliabilities of the two components. For reliabilities of 0.9 and 0.9 the factor is 0.90; for 0.7 and 0.7 it is 0.70; for 0.5 and 0.7 it is 0.59. A true correlation of 0.3 would then appear as 0.27, 0.21, or 0.18. Sample size should therefore be recomputed with the observed reliabilities from the reproducibility substudy; low reliability could make the primary test infeasible at the planned size. Primary analysis uses one tooth per patient, so that clustering does not distort the test.
Statistical analysis plan
Primary analysis. Patient-level displacement rates for d_x and d_z would be estimated by linear mixed models with random intercepts and slopes, using all annual visits. The primary test of P1 is the correlation of the two estimated rates, corrected for attenuation using the reproducibility substudy. The direction angle θ=atan2(d_x,d_z) would be computed for patients whose total displacement exceeds the MDC, and its circular standard deviation reported. P1 is evaluated only in the dispersion range stated in Section 3.4.
Secondary analyses. (i) P2 by a two-outcome structural equation model with one latent factor versus two independent factors, compared by a prespecified information criterion. (ii) P3 by correlation of implant-adjacent and contralateral displacement. (iii) P4 and P5 by interval-censored survival models for contact loss with time-varying d_x and d_z and delivery-time tightness as covariates. (iv) P6 by cross-tabulating distal contact status with the sign of d_x at the distal tooth. Missing visits would be handled by multiple imputation, with a sensitivity analysis under informative drop-out. The six predictions would be handled in a prespecified hierarchy: P1 is primary, and P2 to P6 are secondary with a Holm correction.
Bias control, ethics, and feasibility
The operator who performs registration and landmarking would be blind to contact status and to the patient’s covariates. Scanner type, software version, and registration algorithm would be fixed and reported, and the analysis plan registered before enrollment. Ethics approval and participant consent would be required; none has been sought or obtained for this paper. Feasibility is plausible because digital impressions are already routine at crown delivery in many clinics, but a prospective cohort of this length has an attrition risk that the sample-size calculation covers only approximately.
5. Competing explanations and how the design discriminates among them
Hypothesis H competes with at least five explanations. Table 3 summarizes which predictions separate them.
| Competing explanation | What it claims | Observation that favors it over H | Prediction that tests it |
|---|---|---|---|
| Restoration-side change | Contact loss comes from crown wear, abutment settling, or screw loosening, not tooth movement | Contact gap grows while adjacent-tooth displacement in the implant frame is below MDC; divergence between implant-based and rugae-based registration | P6, reference-stability check |
| Local periodontal and bone-level mechanism | Loss of attachment or inflammation shifts the adjacent tooth only at the contact | Contact loss tracks bone level and probing depth, not arch-level displacement | P3 |
| Geometric coupling only | Vertical movement alters the gap through the proximal slope; no shared driver is needed | dz explains the gap better than dx | P5 |
| Independent drivers | Vertical eruption and mesial drift have unrelated causes | No latent factor; opposite predictor effects | P2, P1 (when dispersion is low) |
| Force-driven drift specific to contacts | Anterior force vector alone moves teeth mesially; vertical change is separate | Baseline force predicts dx but not dz | Exploratory force prediction |
The restoration-side explanation deserves specific attention because it could make contact loss appear tooth-driven when it is not. A night occlusal device failed to prevent contact loss in one cohort [26], which argues against a purely parafunctional mechanism but says nothing about restoration-side change. Displacement measured in the implant frame cannot distinguish between a tooth that moves and a crown that wears; the secondary registration and the wear measurement Δw are needed for that distinction.
The local-periodontal explanation also has support. Lower adjacent-tooth bone support was associated with contact loss in one prospective cohort [9], and a systematic review reported that bone loss and probing depth had unclear associations with contact loss [6]. Under H, bone level would modify the tolerance τ or the rate r_x rather than act as an alternative cause, so bone level is entered as a covariate. A truly separate mechanism would show contact loss without arch-level displacement.
The most difficult discrimination is between H and independent drivers. The two make the same prediction when direction dispersion between patients is high: both predict that d_x and d_z are uncorrelated. The Monte Carlo result of Section 3.3 shows this limit. P2 and P3 are included for that reason. A latent factor shared across components, and a patient-level tendency shared across arch teeth, would be expected under H and under no version of the independent-drivers alternative that lacks a common cause of arch-level movement.
A final alternative is that both complications are real but rarely co-occur in the same patients for reasons unrelated to displacement, for example because contact loss is mostly posterior and infraposition mostly anterior. The anterior stratum addresses this only partly, since anterior contact data are sparse in the studies reviewed. If anterior contact loss proves rare in the cohort, the primary test loses power, and P1 may have to be tested on the continuous scan-derived gap instead of the dichotomous gauge outcome.
6. Clinical and engineering implications
These implications are conditional. They apply only if the within-patient predictions survive testing.
Risk identification. If displacement is the shared driver, a clinician who records scans at delivery and at the first or second maintenance visit could estimate a patient’s displacement rate directly. Such a rate would be a better indicator of later contact loss and infraposition than age, sex, or facial type, which the cohorts reviewed above show to be inconsistent proxies [11–14, 32]. It would not replace those proxies until it were shown to predict outcomes prospectively in a new sample.
Timing and counseling. Reviews already advise delaying implant placement where residual growth is expected and counseling patients about long-term change [2, 16]. A measured displacement rate could refine that advice for individual adults, but the present hypothesis does not itself show that delayed placement reduces either complication. One cohort found greater infraocclusion with delayed than with immediate placement [14], which indicates that timing effects are not simple.
Maintenance intervals and correction. Contact loss is associated with food impaction [10], and options for its correction have been reviewed [24, 40]. If displacement is monitored, the interval between maintenance visits and the decision to modify the contact could rest on the measured gap and the rate at which it widens, not on the time in service alone. Whether contact loss harms peri-implant tissue remains disputed [10, 28], so a measured rate would not by itself justify intervention.
Prosthetic design. If contacts open because of arch-level movement, designs that allow later adjustment of the proximal surface could be considered. The hypothesis does not test any design. Conversely, if restoration-side change dominates, the engineering focus would shift to abutment seating and screw preload [37, 38], and H would have limited clinical reach.
Measurement engineering. The proposal makes the registration reference a design requirement: the implant restoration would have to remain geometrically stable, wear-resistant in the registration zone, and scannable. Whether commercial scanning software offers a registration restricted to a single implant restoration, and with what precision, was not established here.
7. Limitations
- Abstract-level appraisal. All evidence statements derive from PubMed abstracts. Full-text details, such as the exact definition of contact loss, the method of vertical measurement, and handling of dropouts, may change the interpretation of individual studies.
- Site mismatch. Contact-loss evidence is mostly posterior; infraposition evidence is almost entirely anterior maxillary. The hypothesis is tested at sites where the literature gives it least support for contacts.
- No empirical estimate of the key parameters. The mesio-distal drift rates, the tolerance τ, the coupling coefficient κ_k, and the direction dispersion are unmeasured. The time-to-opening table and the Monte Carlo output follow from assumed inputs and do not support any epidemiological claim.
- Measurement precision unverified. Whether intraoral scanning with implant-based registration reaches the precision needed to detect displacements of tens of micrometers per year was not established from the sources read. If it does not, the design fails at the first step.
- Model simplifications. The model treats each tooth as a single reference point, ignores tipping and rotation, and assumes linear gap response. Tipping is a plausible mode of mesial movement and would need a more complete rigid-body formulation with six degrees of freedom per tooth.
- Correlation limits. Section 3.3 shows that a null correlation cannot reject the shared-mechanism claim when direction varies widely. A positive correlation, in turn, could arise from geometric coupling alone, which P5 addresses only partly.
- Heterogeneous outcomes. Contact loss is defined differently across studies (gauge thickness, strip thickness), so the clinical threshold for the dichotomous outcome is arbitrary.
- Unmeasured causes of drift. The study does not intervene on forces, so it cannot establish that the anterior component of force causes drift. The preventive trial that would do so has not been done, as far as registry searches in this work could show.
- Search limits. Embase, Scopus, and Web of Science were not accessible. Europe PMC and ClinicalTrials.gov were searched. Relevant work indexed elsewhere, including orthodontic retention literature on late incisor crowding and drift, may have been missed.
8. Conclusions
Interproximal contact loss and implant infraposition share a physical setting: an immobile implant beside teeth that continue to move. Existing reviews have attributed both to late growth, continued eruption, and mesial drift, and a meta-analysis has reported both, but no study seen in this search measured the two components together or derived conditions for their co-variation. We have proposed a kinematic formulation in which the adjacent tooth’s displacement vector, expressed in an implant-fixed frame, produces infraposition through its vertical component and contact opening through its mesio-distal component, with a sign convention that explains why distal contacts demand a different account. We have shown, under stated assumptions, that correlation between the components can disappear when direction varies between patients, and we have therefore tied falsification to estimated direction dispersion and to tests that do not rely on it. The hypothesis is unsupported by data of its own. It could be refuted by a modest, well-designed cohort, which is its principal merit.
Cite this paper
Jain S, Gupta A. One Displacement Field, Two Complications: A Kinematic Hypothesis Linking Interproximal Contact Loss and Implant Infraposition. Working paper. Center for Implant Dentistry; 2026. Available from: https://bayareaimplantdentistry.com/clinician-research-and-publications/contact-loss-infraposition-kinematics
References
- Jemt T. Measurements of tooth movements in relation to single-implant restorations during 16 years: a case report. Clin Implant Dent Relat Res. 2005;7(4):200-8. doi:10.1111/j.1708-8208.2005.tb00065.x (opens in a new tab). PMID: 16336911 (opens in a new tab).
- Heij DG, Opdebeeck H, van Steenberghe D, Kokich VG, Belser U, Quirynen M. Facial development, continuous tooth eruption, and mesial drift as compromising factors for implant placement. Int J Oral Maxillofac Implants. 2006;21(6):867-78. PMID: 17190296 (opens in a new tab).
- Sheba M, Floriani F, Nimmo A, Ercoli C, Hosney S. Interproximal contact loss between implant restorations and adjacent natural teeth: A systematic review and meta-analysis. J Prosthodont. 2024;33(4):313-323. doi:10.1111/jopr.13780 (opens in a new tab). PMID: 37794763 (opens in a new tab).
- Manicone PF, De Angelis P, Rella E, Papetti L, D’Addona A. Proximal Contact Loss in Implant-Supported Restorations: A Systematic Review and Meta-Analysis of Prevalence. J Prosthodont. 2022;31(3):201-209. doi:10.1111/jopr.13407 (opens in a new tab). PMID: 34263959 (opens in a new tab).
- Bento VAA, Gomes JML, Lemos CAA, Limirio JPJO, Rosa CDDRD, Pellizzer EP. Prevalence of proximal contact loss between implant-supported prostheses and adjacent natural teeth: A systematic review and meta-analysis. J Prosthet Dent. 2023;129(3):404-412. doi:10.1016/j.prosdent.2021.05.025 (opens in a new tab). PMID: 34294422 (opens in a new tab).
- Abduo J, Lau D. Proximal contact loss between implant prostheses and adjacent natural teeth: A qualitative systematic review of prevalence, influencing factors and implications. Heliyon. 2022;8(8):e10064. doi:10.1016/j.heliyon.2022.e10064 (opens in a new tab). PMID: 35992004 (opens in a new tab).
- Papageorgiou SN, Eliades T, Hämmerle CHF. Frequency of infraposition and missing contact points in implant-supported restorations within natural dentitions over time: A systematic review with meta-analysis. Clin Oral Implants Res. 2018;29 Suppl 18:309-325. doi:10.1111/clr.13291 (opens in a new tab). PMID: 30306689 (opens in a new tab).
- Koori H, Morimoto K, Tsukiyama Y, Koyano K. Statistical analysis of the diachronic loss of interproximal contact between fixed implant prostheses and adjacent teeth. Int J Prosthodont. 2010;23(6):535-40. PMID: 21209989 (opens in a new tab).
- Pang NS, Suh CS, Kim KD, Park W, Jung BY. Prevalence of proximal contact loss between implant-supported fixed prostheses and adjacent natural teeth and its associated factors: a 7-year prospective study. Clin Oral Implants Res. 2017;28(12):1501-1508. doi:10.1111/clr.13018 (opens in a new tab). PMID: 28391648 (opens in a new tab).
- Byun SJ, Heo SM, Ahn SG, Chang M. Analysis of proximal contact loss between implant-supported fixed dental prostheses and adjacent teeth in relation to influential factors and effects. A cross-sectional study. Clin Oral Implants Res. 2015;26(6):709-14. doi:10.1111/clr.12373 (opens in a new tab). PMID: 24712313 (opens in a new tab).
- Bernard JP, Schatz JP, Christou P, Belser U, Kiliaridis S. Long-term vertical changes of the anterior maxillary teeth adjacent to single implants in young and mature adults. A retrospective study. J Clin Periodontol. 2004;31(11):1024-8. doi:10.1111/j.1600-051X.2004.00574.x (opens in a new tab). PMID: 15491321 (opens in a new tab).
- Jemt T, Ahlberg G, Henriksson K, Bondevik O. Tooth movements adjacent to single-implant restorations after more than 15 years of follow-up. Int J Prosthodont. 2007;20(6):626-32. PMID: 18069373 (opens in a new tab).
- Andersson B, Bergenblock S, Fürst B, Jemt T. Long-term function of single-implant restorations: a 17- to 19-year follow-up study on implant infraposition related to the shape of the face and patients’ satisfaction. Clin Implant Dent Relat Res. 2013;15(4):471-80. doi:10.1111/j.1708-8208.2011.00381.x (opens in a new tab). PMID: 21834859 (opens in a new tab).
- Polymeri A, Li Q, Laine ML, Loos BG, Wang HL. Occlusal Migration of Teeth Adjacent to Implant Prostheses in Adults: A Long-Term Study. Int J Oral Maxillofac Implants. 2020;35(2):342-349. doi:10.11607/jomi.7784 (opens in a new tab). PMID: 32142571 (opens in a new tab).
- Al Qassar SS, Mavragani M, Psarras V, Halazonetis DJ. The anterior component of occlusal force revisited: direct measurement and theoretical considerations. Eur J Orthod. 2016;38(2):190-6. doi:10.1093/ejo/cjv028 (opens in a new tab). PMID: 25888533 (opens in a new tab).
- Valente NA, D’Amato S, Farella M. Late facial growth and continuous dentoalveolar eruption: Implications for optimal dental implant placement. Periodontol 2000. 2025. doi:10.1111/prd.70024 (opens in a new tab). PMID: 41321054 (opens in a new tab).
- Huanca Ghislanzoni L, Jonasson G, Kiliaridis S. Continuous eruption of maxillary teeth and changes in clinical crown length: A 10-year longitudinal study in adult women. Clin Implant Dent Relat Res. 2017;19(6):1082-1089. doi:10.1111/cid.12545 (opens in a new tab). PMID: 28982223 (opens in a new tab).
- Yilmaz RS, Darling AI, Levers BG. Mesial drift of human teeth assessed from ankylosed deciduous molars. Arch Oral Biol. 1980;25(2):127-31. doi:10.1016/0003-9969(80)90088-6 (opens in a new tab). PMID: 6931561 (opens in a new tab).
- Moss JP, Picton DC. Mesial drift of teeth in adult monkeys (Macaca irus) when forces from the cheeks and tongue had been eliminated. Arch Oral Biol. 1970;15(10):979-86. doi:10.1016/0003-9969(70)90093-2 (opens in a new tab). PMID: 4992597 (opens in a new tab).
- Odman J, Gröndahl K, Lekholm U, Thilander B. The effect of osseointegrated implants on the dento-alveolar development. A clinical and radiographic study in growing pigs. Eur J Orthod. 1991;13(4):279-86. doi:10.1093/ejo/13.4.279 (opens in a new tab). PMID: 1915616 (opens in a new tab).
- Bohner L, Hanisch M, Kleinheinz J, Jung S. Dental implants in growing patients: a systematic review. Br J Oral Maxillofac Surg. 2019;57(5):397-406. doi:10.1016/j.bjoms.2019.04.011 (opens in a new tab). PMID: 31076220 (opens in a new tab).
- Daftary F, Mahallati R, Bahat O, Sullivan RM. Lifelong craniofacial growth and the implications for osseointegrated implants. Int J Oral Maxillofac Implants. 2013;28(1):163-9. doi:10.11607/jomi.2827 (opens in a new tab). PMID: 23377062 (opens in a new tab).
- Alsaloum M, Almohefer SA, Alzaid A, Alburidy F, Alanazi F, Tarabzoni L, et al. Proximal Contact Loss and Interproximal Caries in Natural Teeth Adjacent to Dental Implants. A Review. Clin Cosmet Investig Dent. 2026;18:585610. doi:10.2147/CCIDE.S585610 (opens in a new tab). PMID: 41939798 (opens in a new tab).
- Varthis S, Tarnow DP, Randi A. Interproximal Open Contacts Between Implant Restorations and Adjacent Teeth. Prevalence - Causes - Possible Solutions. J Prosthodont. 2019;28(2):e806-e810. doi:10.1111/jopr.12980 (opens in a new tab). PMID: 30350332 (opens in a new tab).
- Gasser TJW, Papageorgiou SN, Eliades T, Hämmerle CHF, Thoma DS. Interproximal contact loss at implant sites: a retrospective clinical study with a 10-year follow-up. Clin Oral Implants Res. 2022;33(5):482-491. doi:10.1111/clr.13908 (opens in a new tab). PMID: 35194854 (opens in a new tab).
- Bompolaki D, Edmondson SA, Katancik JA. Interproximal contact loss between implant-supported restorations and adjacent natural teeth: A retrospective cross-sectional study of 83 restorations with an up to 10-year follow-up. J Prosthet Dent. 2022;127(3):418-424. doi:10.1016/j.prosdent.2020.09.034 (opens in a new tab). PMID: 33342612 (opens in a new tab).
- Liu L, Jiang Y, Xu H, Zheng Y, Han XJ. Risk factors for proximal contact loss after implant-supported prosthetic restoration in adjacent natural teeth. BMC Oral Health. 2025;26(1):129. doi:10.1186/s12903-025-07480-3 (opens in a new tab). PMID: 41402875 (opens in a new tab).
- Atieh MA, Shah M, Hakam A, Bohamedi K, Tawse-Smith A, Alsabeeha NHM. Impact of Open Proximal Contacts on Peri-Implant Diseases: A Systematic Review and Meta-Analysis. Clin Exp Dent Res. 2026;12(1):e70278. doi:10.1002/cre2.70278 (opens in a new tab). PMID: 41587408 (opens in a new tab).
- Southard TE, Behrents RG, Tolley EA. The anterior component of occlusal force. Part 1. Measurement and distribution. Am J Orthod Dentofacial Orthop. 1989;96(6):493-500. doi:10.1016/0889-5406(89)90116-9 (opens in a new tab). PMID: 2589259 (opens in a new tab).
- Wei H, Tomotake Y, Nagao K, Ichikawa T. Implant prostheses and adjacent tooth migration: preliminary retrospective survey using 3-dimensional occlusal analysis. Int J Prosthodont. 2008;21(4):302-4. PMID: 18717086 (opens in a new tab).
- Cocchetto R, Pradies G, Celletti R, Canullo L. Continuous craniofacial growth in adult patients treated with dental implants in the anterior maxilla. Clin Implant Dent Relat Res. 2019;21(4):627-634. doi:10.1111/cid.12790 (opens in a new tab). PMID: 31037820 (opens in a new tab).
- Sauvin G, Nurdin N, Bischof M, Kiliaridis S. Assessment and aesthetic impact of a long-term vertical discrepancy between the single anterior maxillary implant-supported crown and adjacent teeth: A retrospective cross-sectional study. Clin Exp Dent Res. 2022;8(5):1109-1116. doi:10.1002/cre2.629 (opens in a new tab). PMID: 36029165 (opens in a new tab).
- Nilsson A, Johansson LÅ, Stenport VF, Wennerberg A, Ekfeldt A. Infraposition of anterior maxillary implant-supported single-tooth restorations in adolescent and adult patients-A prospective follow-up study up to 6 years. Clin Implant Dent Relat Res. 2019;21(5):953-959. doi:10.1111/cid.12818 (opens in a new tab). PMID: 31381250 (opens in a new tab).
- Landin N, Kaya C, Papia E, Eliasson A. Infraposition, Esthetics, Survival, and Complications of Implant-Supported Crowns in the Anterior Maxilla: A Retrospective Study with 11-21 Years Follow-up. Int J Oral Maxillofac Implants. 2025;0(0):1-26. doi:10.11607/jomi.11430 (opens in a new tab). PMID: 41201968 (opens in a new tab).
- Winitsky N, Naimi-Akbar A, Nedelcu R, Jemt T, Smedberg JI. 3-D tooth movement adjacent to single anterior implants and esthetic outcome. A 14- to 20-year follow-up study. Clin Oral Implants Res. 2021;32(11):1328-1340. doi:10.1111/clr.13833 (opens in a new tab). PMID: 34403160 (opens in a new tab).
- Basim A, Naoumova J, Leffler K, Johansson K, Persson M, Al-Taai N. Digital analysis of age-related dentoalveolar changes: a 50-year longitudinal study. Prog Orthod. 2026;27(1). doi:10.1186/s40510-026-00643-0 (opens in a new tab). PMID: 42726424 (opens in a new tab).
- Yilmaz B, Seidt JD, McGlumphy EA, Clelland NL. Displacement of screw-retained single crowns into implants with conical internal connections. Int J Oral Maxillofac Implants. 2013;28(3):803-6. doi:10.11607/jomi.3005 (opens in a new tab). PMID: 23748312 (opens in a new tab).
- Luo Q, Ding Q, Zhang L, Xie Q. Analyzing the occlusion variation of single posterior implant-supported fixed prostheses by using the T-scan system: A prospective 3-year follow-up study. J Prosthet Dent. 2020;123(1):79-84. doi:10.1016/j.prosdent.2018.12.012 (opens in a new tab). PMID: 31079880 (opens in a new tab).
- Nguyen VA, Pham TH, Joda T, Mai HN, Hoang V, Marya A. Management of a partially ankylosed mandibular canine using controlled luxation, serial intraoral scan superimposition, and a customized lingual bracket: A case report. Medicine (Baltimore). 2026;105(30):e49869. doi:10.1097/MD.0000000000049869 (opens in a new tab). PMID: 42499045 (opens in a new tab).
- Greenstein G, Carpentieri J, Cavallaro J. Open contacts adjacent to dental implant restorations: Etiology, incidence, consequences, and correction. J Am Dent Assoc. 2016;147(1):28-34. doi:10.1016/j.adaj.2015.06.011 (opens in a new tab). PMID: 26562738 (opens in a new tab).

