Cause or Consequence? A Translational Research Framework for Establishing the Temporal and Causal Relationship Between Titanium Particle Release, Microbial Dysbiosis, and Peri-Implant Inflammation
- Paper type
- Translational research framework with preclinical and prospective clinical study protocols
- Status
- Original Work · Working Paper
Contents
Abstract
Background. Titanium particles and dissolved titanium are found in greater amounts at implants with peri-implantitis than at healthy implants. They are also found at healthy implants. In vitro, titanium ions and particles activate inflammasome signaling in macrophages and alter the composition of oral biofilms. A critical review concluded that higher titanium concentrations in lesions could be a consequence of biofilm and inflammation and need not be their trigger. All human evidence is cross-sectional, and cross-sectional data cannot establish which came first.
Objective. To specify the competing causal models explicitly, to show what each predicts, and to design experiments and observational studies capable of distinguishing them.
Framework. Four models are defined: particles as initiator, particles as amplifier of biofilm-induced disease, particles as consequence, and particles as a marker of a common cause. A two-variable kinetic model shows that in steady state titanium burden and inflammation are correlated under every model, and that the cross-sectional slope identifies a different parameter depending on which upstream factor varies between sites.
Proposed methods. A two-by-two factorial animal experiment (particles by biofilm) with dose levels, endotoxin-controlled particles, a microbiota-depleted arm, and inflammasome blockade. A reverse-direction experiment measuring titanium release as a function of induced inflammation without instrumentation. A prospective human cohort with banked baseline samples and a nested case-control analysis of incident peri-implantitis, with cross-lagged analysis of repeated titanium and inflammatory measures.
Central hypothesis. Particles alone do not initiate marginal bone loss at osseointegrated implants but increase the bone loss produced by a biofilm challenge, and titanium burden measured in health does not predict incident peri-implantitis after adjustment for plaque and inflammation.
Conclusions. The hypothesis is one of several defensible positions. The contribution of this manuscript is a design in which each can be refuted. No results are reported.
Keywords: titanium particles; peri-implantitis; biocorrosion; dysbiosis; causal inference; directed acyclic graph; nested case-control study; animal model
1. Introduction
Titanium is released from dental implants. Particles are shed when the surface is abraded during insertion [1, 2], when components move against one another, when the surface is instrumented during maintenance or treatment, and when the passive oxide layer is disrupted chemically [3–5]. Titanium has been measured in jawbone adjacent to implants [6]. A systematic review concluded that titanium particles in peri-implant tissues are a common finding [7].
Titanium is also found in greater amounts where there is disease. Submucosal plaque from implants with peri-implantitis contained significantly more dissolved titanium than plaque from healthy implants after adjustment for the amount of plaque collected [8]. Tissue from peri-implantitis lesions contained a mean of 98.7 µg/g of titanium against 1.2 µg/g in control tissue [9]. Exfoliated cells from the peri-implant mucosa contained metal-like particles in patients with and without peri-implantitis, at higher concentration in the former [10]. Radiopaque foreign bodies, predominantly titanium and cement, were present in 34 of 36 peri-implantitis biopsies and were surrounded by inflammatory cells [11]. Elemental analysis found titanium in 9 of 12 samples of peri-implantitis tissue [12].
And titanium has biological effects that make a causal role plausible. Titanium ions form particles that activate the inflammasome and release interleukin-1β from primed macrophages [13]. Particles generated by ultrasonic scaling of implant surfaces induced inflammation and osteolysis in a murine calvarial model [14]. Titanium ions and particles shifted the composition of oral biofilms in vitro [15], and in patients the presence of titanium correlated with the principal axis of variation in the peri-implant microbiome and inversely with its diversity [16]. Global DNA methylation has been examined in relation to titanium and disease [17]. From these observations an account has been assembled in which implant degradation products are an abiotic danger signal that drives a non-resolving inflammation distinct from periodontitis [18–22], related to the view of marginal bone loss as a foreign body reaction [23].
The difficulty is that every one of the human observations is equally well explained in the other direction. Inflammation lowers pH and generates reactive species that attack the oxide layer; bacteria and their products promote corrosion; bone loss exposes a larger area of rough surface to the pocket; and diseased implants are instrumented more. A critical review prepared for a consensus conference stated the point directly: higher concentrations of titanium in peri-implantitis lesions could be the consequence of the presence of biofilms and inflammation and not the trigger of disease [24]. Titanium at diseased sites is what one would expect whether it causes disease, results from it, or both.
This is an established question, and it is not presented here as a new one. What the field lacks is evidence of the right kind: designs that establish temporal order and that manipulate one factor while holding the other fixed. This manuscript sets out those designs.
2. Background and Current Scientific Evidence
Sources and forms of titanium
A systematic review of 206 articles catalogued the sources: surgical (drill and implant surface wear at insertion), prosthetic (wear and fretting at connections), and maintenance-related (scaling, air abrasion, implantoplasty), together with electrochemical corrosion [3]. Insertion damages surface features and releases loose particles [1], and titanium was detected on peri-implant bone surfaces immediately after surgery in an early experimental study [2]. Implantoplasty, the deliberate removal of threads and surface texture, releases the largest iatrogenic load; an integrative review reported particles from 100 nm to 54 µm and found no consistent association with clinical outcomes [25], and a systematic review found a single case of mucosal discoloration attributed to particle deposition among several hundred treated implants [26].
Titanium exists in tissue as metallic particles of varying size, as oxide particles, and as ions that bind proteins or precipitate. These forms differ in bioavailability and biological activity, and analytical methods differ in which they detect: mass spectrometry of digested tissue measures total titanium; microscopy with elemental analysis detects particles above a size limit; and analysis of plaque or fluid measures a soluble and fine particulate fraction. Studies using different methods are not measuring the same exposure, a point made in reviews of degradation products [4].
Association with disease
The association is consistent across studies and methods [7–12]. Its limitations are equally consistent: cross-sectional sampling at one time point; small samples; controls that differ from cases in time in function, implant system, and treatment history; and no adjustment, in most, for prior instrumentation of the diseased implants. One study adjusted titanium in plaque for plaque mass [8]. None could adjust for the duration and severity of inflammation that preceded sampling, which is the confounder that matters most if titanium is a consequence.
Biological plausibility
In vitro and small-animal studies show that titanium ions and particles can activate innate immune pathways [13], that peri-implant granulation tissue fibroblasts respond to titanium dioxide particles and to Porphyromonas gingivalis [27], and that particles induce osteolysis in calvarial models in a manner dependent on particle source [14]. The magnitude of any effect depends on the amount and physicochemical properties of the degradation products [4]. The orthopedic literature on wear-particle osteolysis establishes that particulate debris can drive macrophage-mediated bone resorption around joint replacements [28, 29]; the particle loads, materials, and anatomical setting there differ from those at a dental implant.
Three cautions apply to plausibility arguments. Doses used in vitro are seldom related to concentrations measured in tissue. Particles prepared for experiments may carry adherent endotoxin, which can account for inflammatory responses attributed to the particles themselves, a recognized problem in particle research. And calvarial models place particles on periosteum, not around an osseointegrated, transmucosal implant.
Titanium and the microbiome
Titanium products altered the composition of multispecies biofilms in vitro [15], and titanium in plaque correlated with microbiome structure in a cross-sectional clinical sample [16]. It has been argued that the physicochemistry of a titanium oxide surface makes bacterial adhesion fundamentally different from adhesion to hydroxyapatite and that peri-implant dysbiosis should be understood as a microbiome–biomaterial interaction [19, 20]. The same correlation is expected if a dysbiotic, acidogenic, inflammophilic community corrodes the surface more. Direction is again undetermined.
What the existing reviews conclude
Reviews fall into two camps. One regards a causal role as an emerging concept deserving investigation [18, 20]. The other regards the evidence as insufficient and consequence as at least as likely [24]. Both agree that prospective data are absent. A registry search identified one registered study concerning the effect of metallic particles on progression of peri-implantitis [30]; its design was not reviewed.
Animal models available
Ligature-induced peri-implantitis has been established in mice [31], and a systematic review found that rodent models produce circumferential marginal bone loss resembling the clinical presentation [32]. The canine ligature model is the historical standard [33], and defects it produces have been compared with naturally occurring human defects [34]. These models allow biofilm challenge to be imposed or withheld and particles to be added or not.
3. Research Gap
The gap is methodological. No human study has measured titanium burden at implants before disease and related it to subsequent onset. No animal study, as far as the search could determine, has crossed particle exposure with biofilm challenge at osseointegrated transmucosal implants with control of endotoxin and with a microbiota-depleted arm. No study has measured whether experimentally induced peri-implant inflammation increases titanium release in the absence of instrumentation. Without these, the cross-sectional association cannot be interpreted.
4. Research Objective and Central Hypothesis
The objective is to determine the direction and nature of the relationship between titanium release and peri-implant inflammation.
Four models are distinguished. Under the initiator model, particles at clinically relevant doses are sufficient to start inflammation and bone loss without a biofilm challenge. Under the amplifier model, particles do not initiate disease but increase its severity or rate when biofilm-induced inflammation is present. Under the consequence model, inflammation and biofilm increase titanium release, and titanium has no material effect on the course of disease. Under the common-cause model, some third factor, such as connection micromotion, surface type, or deep placement, increases both particle release and bacterial challenge.
The central hypothesis, chosen as the position best supported by current evidence and stated so that it can be refuted, combines the amplifier and consequence models: (H1) particles alone, at doses within the range measured in human peri-implant tissue, do not produce marginal bone loss at osseointegrated implants in an animal model; (H2) the same particles increase bone loss when combined with ligature-induced biofilm; (H3) induced inflammation increases titanium release without instrumentation; and (H4) in a prospective human cohort, titanium burden measured while the implant is healthy does not predict incident peri-implantitis after adjustment for plaque, bleeding, and established risk indicators.
Refutation of H1 would support the initiator model. Refutation of H2 together with support for H3 would support pure consequence. Refutation of H4, a positive prospective association robust to adjustment, would be the strongest human evidence yet for a causal contribution.
5. Scientific and Theoretical Framework
Causal structure
Figure 1 shows the assumed structure as a directed graph [35]. Titanium burden T and inflammation I are connected by arrows in both directions, corresponding to the causal and consequence pathways, and both are affected by biofilm B. Upstream of T are sources independent of inflammation (insertion wear, connection fretting, instrumentation) and sources dependent on it (corrosion in an inflamed, acidic environment; exposure of surface by bone loss). Upstream of B are hygiene, prosthesis design, and host factors. Instrumentation is a consequence of diagnosed disease and a cause of titanium release, which closes a loop through clinical care.
The graph makes two things plain. Any cross-sectional association between T and I has at least four sources: the forward path, the reverse path, common causes, and the treatment loop. And the forward path can be isolated only by an intervention on T that does not operate through B or I, or by observing T before I changes.
A minimal kinetic model
To see what cross-sectional data can and cannot show, consider linear kinetics for titanium burden and inflammation at a site:
dT/dt=r_0+r_1I−λT, dI/dt=aB+c T−d I (1)
Here r_0 is the inflammation-independent release rate, r_1 the additional release per unit inflammation (the consequence pathway), λ the clearance rate of titanium from the tissue, a the inflammatory effect of biofilm, c the inflammatory effect of titanium (the causal pathway), and d the resolution rate. Units are arbitrary. The model is linear and omits the microbiome pathway and thresholds; it is used only to reason about identification.
At steady state,
T^*=(r_0+r_1I^*)/λ, I^*=(aB+c T^*)/d (2)
Both increase with B and with r_0 whenever c and r_1 are non-negative. Titanium and inflammation are positively correlated across sites if either pathway operates, and also if neither operates but r_0 and B happen to be correlated through a common cause. Section 7 develops the consequences.
Predictions of the four models
Table 1 lists what each model predicts for the proposed studies. No single observation discriminates among all four; the set does.
| Observation | Initiator | Amplifier | Consequence | Common cause |
|---|---|---|---|---|
| Particles alone cause bone loss in animals | Yes | No | No | No |
| Particles add to biofilm-induced bone loss | Yes | Yes | No | No |
| Particle effect persists with microbiota depleted | Yes | No | No | No |
| Induced inflammation raises titanium release | Not required | Not required | Yes | Not required |
| Baseline titanium predicts incident disease, adjusted | Yes | Possibly, with interaction | No | Attenuated by adjustment for the common cause |
| Inflammasome blockade abolishes particle effect | Yes | Yes | Not applicable | Not applicable |
6. Methodology
Study A: factorial animal experiment (forward direction)
Model. Osseointegrated transmucosal titanium implants in a rodent maxilla [31] for the primary experiment, with a confirmatory experiment in a larger species if results warrant. Implants heal to integration before any exposure.
Design. Two-by-two factorial: particle exposure (none, particles) by biofilm challenge (none, ligature), with particle dose at three levels in the particle arms. Implants are randomized to cells with balanced allocation of side.
Particles. The validity of the experiment rests on the particles. They are generated from the same grade of titanium and surface as the test implants by a defined method, characterized for size distribution, shape, composition, and surface oxide, and matched to the size range reported in human peri-implant tissue and after implantoplasty [9, 25]. They are depyrogenated, and endotoxin is verified to be below detection by a quantitative assay before each use. A control arm receives endotoxin-contaminated particles at a known level to show the size of the artifact that depyrogenation removes. Doses are set as multiples of the tissue concentration reported in human lesions [9], with the conversion from tissue concentration to administered mass stated and its uncertainty acknowledged. Particles are delivered to the peri-implant sulcus in a carrier at defined intervals; a carrier-only control is included.
Additional arms. A microbiota-depleted arm (broad-spectrum antibiotics, or germ-free animals if available) receives particles, to test whether any particle effect requires bacteria. An inflammasome-inhibited arm receives particles with ligature, to test the pathway suggested by in vitro work [13]. Titanium ions in solution are tested in a separate arm, since ions and particles may differ in effect [15].
Outcomes. Primary: marginal bone loss by micro-computed tomography at a prespecified time after exposure begins. Secondary: histological area of inflammatory infiltrate; osteoclast number; tissue cytokines; peri-implant microbiome by sequencing, to test the titanium-to-dysbiosis path; and tissue titanium by mass spectrometry, to confirm delivered dose.
Analysis. A linear mixed model with animal as a random effect estimates the main effects of particles and ligature and their interaction. H1 is the test of the particle main effect in the absence of ligature, framed as an equivalence test against a prespecified margin, because the hypothesis is one of no meaningful effect. H2 is the particle effect in the presence of ligature. Dose–response is tested by trend.
Sample size. The standard deviation of marginal bone loss in the chosen model is taken from a pilot; group sizes are then set so that the equivalence test for H1 has 90% power to exclude a particle-alone effect half as large as the ligature effect. No number is asserted before the pilot. Reporting follows ARRIVE 2.0 [36].
Study B: reverse-direction experiment
Aim. To determine whether inflammation increases titanium release.
In vivo. Integrated implants are assigned to ligature or no ligature. No instrumentation of the implant surface is performed at any time. At serial time points, peri-implant fluid is sampled non-invasively and tissue is harvested in subgroups for total titanium by mass spectrometry and for particle imaging. Ligature material is verified titanium-free. The outcome is titanium concentration over time by group. A rise in the ligature group that follows the rise in inflammation supports the consequence pathway.
In vitro. Implant surfaces are exposed to defined biofilms, to acidic and oxidative conditions representative of an inflamed pocket, and to combinations, with titanium release measured over time. This identifies which features of the inflammatory environment drive release.
Study C: prospective human cohort with nested case-control analysis
Design. Prospective cohort of patients receiving implants, with annual standardized assessment and biobanking, and a nested case-control analysis when sufficient cases have accrued; reported to STROBE [37].
Baseline and repeated measures. At delivery of the prosthesis and annually: submucosal plaque and peri-implant crevicular fluid for titanium by mass spectrometry, normalized to sample mass or volume as in prior work [8]; exfoliative cytology of the peri-implant mucosa [10]; crevicular fluid inflammatory markers; plaque and bleeding indices; probing depths; radiographs at intervals consistent with standard care. Every instrumentation of the implant is recorded with date and method, since instrumentation releases titanium.
Cases and controls. Cases are implants meeting the 2017 World Workshop definition of peri-implantitis during follow-up [38, 39]. For each case, two controls are selected from implants still healthy at the time the case was diagnosed, matched on time in function and center.
Primary analysis. Conditional logistic regression of case status on titanium burden measured at the last visit at which the case implant was still healthy and before any instrumentation beyond routine prophylaxis, adjusted for plaque and bleeding at that visit, history of periodontitis, smoking, diabetes, implant surface and connection type, and keratinized mucosa. H4 is that the adjusted odds ratio does not differ from one. Because the hypothesis is null, the analysis reports the confidence interval and the smallest effect that can be excluded.
Secondary analysis: temporal ordering. With repeated measures of titanium T_t and an inflammatory marker I_t at each visit, a cross-lagged panel model estimates the association of T_t with I_(t+1) conditional on I_t, and of I_t with T_(t+1) conditional on T_t. The first coefficient reflects the forward path and the second the reverse. Random intercepts separate stable between-implant differences from within-implant change.
Sensitivity analyses. E-values quantify how strong an unmeasured confounder would need to be to explain an observed association [40]. Negative-control exposures (an element not derived from the implant, measured in the same samples) test for artifacts of sampling. Analyses stratified by implant surface and by connection type address the common-cause model. The analysis is framed as an emulation of a target trial in which titanium exposure is assigned [41], and the ways in which the observational design departs from that trial are stated.
Sample size. For an exposure defined as the highest third of baseline titanium, prevalence 33% among controls, a target odds ratio of 2.0 (an assumption about the smallest effect of interest, not an estimate), two controls per case, 80% power, and two-sided α of 0.05, approximately 100 cases and 200 controls are required. The cohort size needed to accrue 100 incident cases depends on incidence and follow-up and will require a multicenter design over several years.
Study D: use of existing natural experiments
Two existing situations provide partial evidence at low cost. Implantoplasty delivers a large particle load at a known time; if particles were a strong initiator or amplifier, treated sites should fare worse than sites treated without it, and reviews have not found this [25, 26], with the caveats that implantoplasty also removes contaminated surface and that its particles are comparatively large. Non-titanium implants provide a setting in which peri-implant inflammation occurs without titanium release from the fixture; a systematic comparison of lesion characteristics would bear on whether titanium is necessary. Both are confounded and are proposed as supporting analyses only.
Bias and confounding
In Study A, endotoxin on particles and mechanical irritation from particle delivery are the main threats and are addressed by depyrogenation, the contaminated-particle control, and the carrier control. In Study C, the principal threats are reverse causation from subclinical inflammation already present at the baseline measurement, addressed by adjusting for inflammatory markers at that visit and by the cross-lagged analysis; differential instrumentation, addressed by recording and by restricting the exposure window; and measurement error in titanium, which biases toward the null and is quantified by duplicate sampling. Laboratory analyses are blinded to case status.
Ethics
Animal studies require institutional animal care and use committee approval. The cohort requires institutional review board approval and consent for biobanking and for elemental and microbial analyses; sampling is non-invasive.
7. Analytical Development
Why cross-sectional association is uninformative about direction
Solving equations (2) simultaneously gives
I^*=(λ aB+c r_0)/(λd−c r_1), T^*=(d r_0+r_1 aB)/(λd−c r_1) (3)
provided λd>c r_1, the condition for a stable steady state. Suppose sites differ in biofilm load B and are otherwise alike. Then across sites, from the first of equations (2),
(dT^*)/(dI^*)|_(B varies)=r_1/λ (4)
The slope of titanium on inflammation equals the consequence parameter divided by clearance and contains no information about c. Now suppose sites differ in inflammation-independent release r_0 and are alike in B. From the second of equations (2),
(dI^*)/(dT^*)|_(r_0 varies)=c/d (5)
The slope of inflammation on titanium equals the causal parameter divided by the resolution rate and contains no information about r_1.
The same scatter of points has a different meaning depending on what generated the variation. In a clinical sample, sites differ mostly in biofilm, hygiene, and host response, so the observed association mostly reflects r_1, the consequence pathway, as equation (4) indicates. To learn about c, variation must come from r_0: that is, from differences in titanium release that are unrelated to biofilm and inflammation. Randomized particle delivery in Study A is such a source by construction. In humans, candidate sources are design features that affect release mechanically, provided they do not also affect bacterial challenge, an assumption that is difficult to defend for connection type and somewhat easier for one-time events such as documented insertion damage.
What temporal data add
In the dynamic model, a change in T produces a change in I only after a delay of order 1/d, and a change in I produces a change in T after a delay of order 1/λ. If sampling intervals are short relative to these times, the cross-lagged coefficients separate the two paths: T_t→I_(t+1) estimates a quantity proportional to c and I_t→T_(t+1) a quantity proportional to r_1. If sampling is annual and both time constants are weeks, each visit observes a near-steady state, the lagged coefficients inherit the ambiguity of the cross-sectional slope, and only the baseline-before-disease comparison retains discriminating power. The time constants are unknown. Study B estimates the delay from inflammation to titanium rise in animals, and a subset of the cohort should be sampled more frequently to check the assumption in humans.
The amplification condition
Equation (3) shows that when both pathways operate, their product appears in the denominator: steady-state inflammation for a given biofilm load is multiplied by 1/(1−c r_1/(λd)) relative to the case with no feedback. The dimensionless loop gain
G=(c r_1)/(λ d) (6)
measures how much the titanium loop amplifies biofilm-driven inflammation. If G is small, titanium is a bystander in practical terms even if c is not zero. If G approaches one, a small biofilm challenge produces a large and self-sustaining response. Whether particles matter clinically is therefore a question about a product of four quantities. Studies A and B together estimate the two numerator terms in an animal model; the denominators, clearance and resolution, are properties of the tissue. Framing the debate as cause versus consequence obscures the fact that the two are factors of one loop, and that the loop’s gain is the quantity of interest.
Interpreting null results
H1 and H4 are hypotheses of no effect. Absence of evidence is weak unless the study could have detected an effect of relevant size. Both analyses are therefore specified as equivalence or exclusion tests with stated margins, and the report of each will give the largest effect compatible with the data. A particle-alone effect smaller than half the ligature effect, or a baseline-titanium odds ratio whose confidence interval lies within 0.67 to 1.5, would be taken as evidence against a clinically important initiating role. These margins are proposals to be fixed with clinical input before data collection.
8. Discussion
The debate over titanium particles has continued for a decade on evidence that cannot settle it. The framework here does not favor one side. It adopts as its working hypothesis the position that particles amplify and result from inflammation without initiating it, because that position is compatible with the in vitro evidence for biological activity, with the ubiquity of titanium at healthy implants, and with the lack of excess harm after implantoplasty. Each of the proposed studies can overturn it.
The strongest argument for an initiating role is mechanistic: particles activate the inflammasome, and the peri-implant lesion is larger and less contained than the periodontal lesion. The strongest argument against is epidemiological: nearly all implants shed titanium and most do not develop peri-implantitis, and plaque control prevents much of the disease. Neither argument addresses dose. It is possible that typical loads are harmless, that high loads from specific events are not, and that the distribution of exposure is highly skewed. The dose levels in Study A and the tertile analysis in Study C are designed with that possibility in mind, and a threshold dose–response would be a coherent outcome that neither camp currently predicts.
Several features of the design are open to criticism. Rodent implants are small, the ligature model is acute, and repeated delivery of particles to a sulcus is itself a disturbance. A particle-alone effect in such a model could reflect delivery trauma; the carrier control addresses this imperfectly. Translating a tissue concentration in micrograms per gram to an administered dose involves assumptions about distribution and clearance that could be wrong by an order of magnitude. And an antibiotic-treated animal is not free of microbial products.
In the human cohort, the central difficulty is that an implant sampled while clinically healthy may already have subclinical inflammation that has raised titanium release. A positive prospective association would then still be compatible with consequence. Adjustment for inflammatory markers at the sampling visit mitigates this and may over-adjust if titanium acts through those markers. The cross-lagged analysis helps only if the sampling interval is short enough, as Section 7.2 explains. No observational design fully escapes the problem. The cohort provides the best achievable human evidence short of an intervention on exposure, and there is no ethical way to randomize people to titanium particles. Randomizing a reduction in exposure, for example to maintenance instruments that do or do not abrade the surface, is feasible and is the logical trial if the cohort suggests an effect.
The microbiome pathway is under-represented in the kinetic model. If titanium shifts community composition toward a more inflammatory state [15, 16], its effect on inflammation is mediated by B and would appear in Study A as a change in the microbiome preceding a change in bone, absent in the microbiota-depleted arm. The sequencing endpoints are included for this reason.
Finally, a clear answer may have limited practical consequence. If particles amplify disease, the response is to reduce avoidable release, which is already reasonable practice. If they are purely a consequence, titanium in plaque or fluid might still serve as a marker of active corrosion in an inflamed site. The value of resolving the question lies mainly in directing therapeutic research: host-modulatory treatment aimed at titanium-induced pathways [20] is worth developing only if the loop gain is appreciable.
9. Clinical Significance
Resolution of the question would inform several decisions that are currently made on preference. Choice of maintenance instruments and of decontamination methods involves a trade-off between cleaning efficacy and surface damage whose terms depend on whether released particles matter. The same applies to implantoplasty, to connection designs that differ in fretting, and to the growing interest in non-titanium implants. If titanium burden proves to be a consequence that rises with inflammation, its measurement in crevicular fluid could be explored as an indicator of disease activity. If it proves to be an amplifier with meaningful loop gain, minimizing iatrogenic release becomes a specific objective, and anti-inflammatory strategies directed at particle-induced pathways acquire a rationale.
At present the evidence does not justify changing practice to avoid titanium release, and it does not justify dismissing the possibility that release contributes to disease.
10. Limitations
The kinetic model is linear and deterministic and is intended for reasoning about identification, not for quantitative prediction. The animal experiments use models that differ from human disease in scale, time course, and microbial ecology. Particle preparation, however careful, may not reproduce the form of titanium that matters in vivo. The cohort requires many years and centers to accrue cases, and its primary hypothesis is null, which demands precision. Titanium assays in small plaque and fluid samples are near detection limits and vulnerable to contamination from instruments. The common-cause model cannot be excluded by adjustment if the common cause is unmeasured. The sample-size calculation uses an assumed odds ratio. The literature was assessed from abstracts and index records, and the orthopedic wear-particle literature was cited only for general context.
11. Future Validation
If Study A shows a particle effect that requires biofilm and is abolished by inflammasome blockade, replication in a large-animal model with clinically realistic implants would be warranted before any human intervention. If Study C shows a prospective association, a randomized trial of a maintenance protocol that reduces surface abrasion, with titanium burden and marginal bone as outcomes, would test causality by lowering exposure. If both are null, further work on particle-targeted therapy would be hard to justify. Harmonization of titanium measurement across laboratories, with shared reference materials, is a prerequisite for any multicenter study and should be undertaken first.
12. Conclusion
Titanium is more abundant at diseased implants, and that observation does not show whether titanium contributes to the disease. A simple kinetic analysis demonstrates that the cross-sectional association reflects the consequence pathway when sites differ in biofilm and the causal pathway only when they differ in inflammation-independent release, and that the practical importance of particles depends on the gain of a loop in which cause and consequence are both factors. A factorial animal experiment with endotoxin-controlled particles, a reverse-direction experiment without instrumentation, and a prospective cohort with baseline sampling before disease would together distinguish the competing models. The working hypothesis, that particles amplify and follow inflammation without initiating it, may be wrong, and the studies are designed to show it if so.
Cite this paper
Jain S, Gupta A. Cause or Consequence? A Translational Research Framework for Establishing the Temporal and Causal Relationship Between Titanium Particle Release, Microbial Dysbiosis, and Peri-Implant Inflammation. Working paper. Center for Implant Dentistry; 2026. Available from: https://bayareaimplantdentistry.com/clinician-research-and-publications/titanium-debris-cause-or-consequence
References
- Senna P, Antoninha Del Bel Cury A, Kates S, Meirelles L. Surface Damage on Dental Implants with Release of Loose Particles after Insertion into Bone. Clin Implant Dent Relat Res. 2015;17(4):681-92. doi:10.1111/cid.12167 (opens in a new tab). PMID: 24283455 (opens in a new tab).
- Schliephake H, Reiss G, Urban R, Neukam FW, Guckel S. Metal release from titanium fixtures during placement in the mandible: an experimental study. Int J Oral Maxillofac Implants. 1993;8(5):502-11. PMID: 8112789 (opens in a new tab).
- Delgado-Ruiz R, Romanos G. Potential Causes of Titanium Particle and Ion Release in Implant Dentistry: A Systematic Review. Int J Mol Sci. 2018;19(11). doi:10.3390/ijms19113585 (opens in a new tab). PMID: 30428596 (opens in a new tab).
- Noronha Oliveira M, Schunemann WVH, Mathew MT, Henriques B, Magini RS, Teughels W, et al. Can degradation products released from dental implants affect peri-implant tissues?. J Periodontal Res. 2018;53(1):1-11. doi:10.1111/jre.12479 (opens in a new tab). PMID: 28766712 (opens in a new tab).
- Mouhyi J, Dohan Ehrenfest DM, Albrektsson T. The peri-implantitis: implant surfaces, microstructure, and physicochemical aspects. Clin Implant Dent Relat Res. 2012;14(2):170-83. doi:10.1111/j.1708-8208.2009.00244.x (opens in a new tab). PMID: 19843108 (opens in a new tab).
- He X, Reichl FX, Wang Y, Michalke B, Milz S, Yang Y, et al. Analysis of titanium and other metals in human jawbones with dental implants - A case series study. Dent Mater. 2016;32(8):1042-51. doi:10.1016/j.dental.2016.05.012 (opens in a new tab). PMID: 27298240 (opens in a new tab).
- Suárez-López Del Amo F, Garaicoa-Pazmiño C, Fretwurst T, Castilho RM, Squarize CH. Dental implants-associated release of titanium particles: A systematic review. Clin Oral Implants Res. 2018;29(11):1085-1100. doi:10.1111/clr.13372 (opens in a new tab). PMID: 30280418 (opens in a new tab).
- Safioti LM, Kotsakis GA, Pozhitkov AE, Chung WO, Daubert DM. Increased Levels of Dissolved Titanium Are Associated With Peri-Implantitis - A Cross-Sectional Study. J Periodontol. 2017;88(5):436-442. doi:10.1902/jop.2016.160524 (opens in a new tab). PMID: 27858551 (opens in a new tab).
- Pettersson M, Pettersson J, Johansson A, Molin Thorén M. Titanium release in peri-implantitis. J Oral Rehabil. 2019;46(2):179-188. doi:10.1111/joor.12735 (opens in a new tab). PMID: 30325523 (opens in a new tab).
- Olmedo DG, Nalli G, Verdú S, Paparella ML, Cabrini RL. Exfoliative cytology and titanium dental implants: a pilot study. J Periodontol. 2013;84(1):78-83. doi:10.1902/jop.2012.110757 (opens in a new tab). PMID: 22414261 (opens in a new tab).
- Wilson TG Jr, Valderrama P, Burbano M, Blansett J, Levine R, Kessler H, et al. Foreign bodies associated with peri-implantitis human biopsies. J Periodontol. 2015;86(1):9-15. doi:10.1902/jop.2014.140363 (opens in a new tab). PMID: 25224174 (opens in a new tab).
- Fretwurst T, Buzanich G, Nahles S, Woelber JP, Riesemeier H, Nelson K. Metal elements in tissue with dental peri-implantitis: a pilot study. Clin Oral Implants Res. 2016;27(9):1178-86. doi:10.1111/clr.12718 (opens in a new tab). PMID: 26508041 (opens in a new tab).
- Pettersson M, Kelk P, Belibasakis GN, Bylund D, Molin Thorén M, Johansson A. Titanium ions form particles that activate and execute interleukin-1β release from lipopolysaccharide-primed macrophages. J Periodontal Res. 2017;52(1):21-32. doi:10.1111/jre.12364 (opens in a new tab). PMID: 26987886 (opens in a new tab).
- Eger M, Sterer N, Liron T, Kohavi D, Gabet Y. Scaling of titanium implants entrains inflammation-induced osteolysis. Sci Rep. 2017;7:39612. doi:10.1038/srep39612 (opens in a new tab). PMID: 28059080 (opens in a new tab).
- Souza JGS, Costa Oliveira BE, Bertolini M, Lima CV, Retamal-Valdes B, de Faveri M, et al. Titanium particles and ions favor dysbiosis in oral biofilms. J Periodontal Res. 2020;55(2):258-266. doi:10.1111/jre.12711 (opens in a new tab). PMID: 31762055 (opens in a new tab).
- Daubert D, Pozhitkov A, McLean J, Kotsakis G. Titanium as a modifier of the peri-implant microbiome structure. Clin Implant Dent Relat Res. 2018;20(6):945-953. doi:10.1111/cid.12676 (opens in a new tab). PMID: 30255621 (opens in a new tab).
- Daubert DM, Pozhitkov AE, Safioti LM, Kotsakis GA. Association of Global DNA Methylation to Titanium and Peri-Implantitis: A Case-Control Study. JDR Clin Trans Res. 2019;4(3):284-291. doi:10.1177/2380084418822831 (opens in a new tab). PMID: 30931716 (opens in a new tab).
- Fretwurst T, Nelson K, Tarnow DP, Wang HL, Giannobile WV. Is Metal Particle Release Associated with Peri-implant Bone Destruction? An Emerging Concept. J Dent Res. 2018;97(3):259-265. doi:10.1177/0022034517740560 (opens in a new tab). PMID: 29130804 (opens in a new tab).
- Kotsakis GA, Olmedo DG. Peri-implantitis is not periodontitis: Scientific discoveries shed light on microbiome-biomaterial interactions that may determine disease phenotype. Periodontol 2000. 2021;86(1):231-240. doi:10.1111/prd.12372 (opens in a new tab). PMID: 33690947 (opens in a new tab).
- Kotsakis GA, Ganesan SM. Microbial Dysbiosis, Titanium Release, and Peri-implantitis. J Dent Res. 2025;104(5):473-480. doi:10.1177/00220345241307939 (opens in a new tab). PMID: 39953673 (opens in a new tab).
- Asa’ad F, Thomsen P, Kunrath MF. The Role of Titanium Particles and Ions in the Pathogenesis of Peri-Implantitis. J Bone Metab. 2022;29(3):145-154. doi:10.11005/jbm.2022.29.3.145 (opens in a new tab). PMID: 36153850 (opens in a new tab).
- Chen L, Tong Z, Luo H, Qu Y, Gu X, Si M. Titanium particles in peri-implantitis: distribution, pathogenesis and prospects. Int J Oral Sci. 2023;15(1):49. doi:10.1038/s41368-023-00256-x (opens in a new tab). PMID: 37996420 (opens in a new tab).
- Albrektsson T, Dahlin C, Jemt T, Sennerby L, Turri A, Wennerberg A. Is marginal bone loss around oral implants the result of a provoked foreign body reaction?. Clin Implant Dent Relat Res. 2014;16(2):155-65. doi:10.1111/cid.12142 (opens in a new tab). PMID: 24004092 (opens in a new tab).
- Mombelli A, Hashim D, Cionca N. What is the impact of titanium particles and biocorrosion on implant survival and complications? A critical review. Clin Oral Implants Res. 2018;29 Suppl 18:37-53. doi:10.1111/clr.13305 (opens in a new tab). PMID: 30306693 (opens in a new tab).
- Rodríguez Alvarez MB, Padullés-Roig E, Cabanes-Gumbau G, Callejas-Cano JA, Gil J. Release of Titanium Particles After Implantoplasty in the Treatment of Peri-Implantitis: Local and Systemic Implications-An Integrative Systematic Review. J Clin Med. 2025;14(24). doi:10.3390/jcm14248661 (opens in a new tab). PMID: 41464564 (opens in a new tab).
- Stavropoulos A, Bertl K, Eren S, Gotfredsen K. Mechanical and biological complications after implantoplasty-A systematic review. Clin Oral Implants Res. 2019;30(9):833-848. doi:10.1111/clr.13499 (opens in a new tab). PMID: 31254417 (opens in a new tab).
- Irshad M, Scheres N, Crielaard W, Loos BG, Wismeijer D, Laine ML. Influence of titanium on in vitro fibroblast-Porphyromonas gingivalis interaction in peri-implantitis. J Clin Periodontol. 2013;40(9):841-9. doi:10.1111/jcpe.12136 (opens in a new tab). PMID: 23875835 (opens in a new tab).
- Goodman SB. Wear particles, periprosthetic osteolysis and the immune system. Biomaterials. 2007;28(34):5044-8. doi:10.1016/j.biomaterials.2007.06.035 (opens in a new tab). PMID: 17645943 (opens in a new tab).
- Ingham E, Fisher J. The role of macrophages in osteolysis of total joint replacement. Biomaterials. 2005;26(11):1271-86. doi:10.1016/j.biomaterials.2004.04.035 (opens in a new tab). PMID: 15475057 (opens in a new tab).
- ClinicalTrials.gov. Identifier NCT07792499. Effect of metallic particles on the progression of peri-implantitis. Bethesda (MD): National Library of Medicine (US).
- Pirih FQ, Hiyari S, Barroso AD, Jorge AC, Perussolo J, Atti E, et al. Ligature-induced peri-implantitis in mice. J Periodontal Res. 2015;50(4):519-24. doi:10.1111/jre.12234 (opens in a new tab). PMID: 25244403 (opens in a new tab).
- Chew RJJ, Lu JX, Sim YF, Yeo ABK. Rodent peri-implantitis models: a systematic review and meta-analysis of morphological changes. J Periodontal Implant Sci. 2022;52(6):479-495. doi:10.5051/jpis.2200900045 (opens in a new tab). PMID: 36468467 (opens in a new tab).
- Lindhe J, Berglundh T, Ericsson I, Liljenberg B, Marinello C. Experimental breakdown of peri-implant and periodontal tissues. A study in the beagle dog. Clin Oral Implants Res. 1992;3(1):9-16. doi:10.1034/j.1600-0501.1992.030102.x (opens in a new tab). PMID: 1420727 (opens in a new tab).
- Schwarz F, Herten M, Sager M, Bieling K, Sculean A, Becker J. Comparison of naturally occurring and ligature-induced peri-implantitis bone defects in humans and dogs. Clin Oral Implants Res. 2007;18(2):161-70. doi:10.1111/j.1600-0501.2006.01320.x (opens in a new tab). PMID: 17348880 (opens in a new tab).
- Greenland S, Pearl J, Robins JM. Causal diagrams for epidemiologic research. Epidemiology. 1999;10(1):37-48. PMID: 9888278 (opens in a new tab).
- Percie du Sert N, Hurst V, Ahluwalia A, Alam S, Avey MT, Baker M, et al. The ARRIVE guidelines 2.0: Updated guidelines for reporting animal research. PLoS Biol. 2020;18(7):e3000410. doi:10.1371/journal.pbio.3000410 (opens in a new tab). PMID: 32663219 (opens in a new tab).
- von Elm E, Altman DG, Egger M, Pocock SJ, Gøtzsche PC, Vandenbroucke JP. The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: guidelines for reporting observational studies. Lancet. 2007;370(9596):1453-7. doi:10.1016/S0140-6736(07)61602-X (opens in a new tab). PMID: 18064739 (opens in a new tab).
- Renvert S, Persson GR, Pirih FQ, Camargo PM. Peri-implant health, peri-implant mucositis, and peri-implantitis: Case definitions and diagnostic considerations. J Periodontol. 2018;89 Suppl 1:S304-S312. doi:10.1002/JPER.17-0588 (opens in a new tab). PMID: 29926953 (opens in a new tab).
- Berglundh T, Armitage G, Araujo MG, Avila-Ortiz G, Blanco J, Camargo PM, et al. Peri-implant diseases and conditions: Consensus report of workgroup 4 of the 2017 World Workshop on the Classification of Periodontal and Peri-Implant Diseases and Conditions. J Clin Periodontol. 2018;45 Suppl 20:S286-S291. doi:10.1111/jcpe.12957 (opens in a new tab). PMID: 29926491 (opens in a new tab).
- VanderWeele TJ, Ding P. Sensitivity Analysis in Observational Research: Introducing the E-Value. Ann Intern Med. 2017;167(4):268-274. doi:10.7326/M16-2607 (opens in a new tab). PMID: 28693043 (opens in a new tab).
- Hernán MA, Robins JM. Using Big Data to Emulate a Target Trial When a Randomized Trial Is Not Available. Am J Epidemiol. 2016;183(8):758-64. doi:10.1093/aje/kwv254 (opens in a new tab). PMID: 26994063 (opens in a new tab).

