DDJ Patient Article · As of March 2026 · Explained Simply
How Much Does Reported Periimplantitis Prevalence Depend on the Definition Used, and What Does That Mean for Clinical Numbers?
Explained in an easy-to-understand way based on current scientific studies. This article helps you make informed decisions with your dentist.
This topic involves a method of investigation and the question of how reliably it can detect certain problems.
Quick Summary
The most important takeaways at a glance:
- There are indications of an association, but no final certainty yet.
- The scientific basis is solid, but not all questions have been definitively answered.
- Different definitions produce very different prevalences.
- If a prevalence number seems high, the first question is not how bad it is, but how it was defined.
Why Is This Topic Important for You?
You may have heard that there are differing opinions on this topic. This is because science is often more complex than a simple yes or no answer suggests. In this article, we explain what current research actually shows—without technical jargon and without leaving out important details.
This topic is not a benefit statement, but rather a question of definition and measurement.
Why is this important for you? Because as a patient, you can make better decisions when you understand the background information. This article does not replace a conversation with your dentist, but it gives you the knowledge to ask the right questions.
In research, the most important questions revolve around the following areas: case definition, follow-up, maintenance, and population, as well as diagnostic signal versus crisis narrative. In the following sections, we will explain what the studies say about each of these areas and what that means for your daily life.
What does "case definition" mean for me as a patient?
A common patient question is how to weigh case definition. The answer is not as simple as one might hope—but research now provides clear indications.
The reported prevalence of peri-implantitis depends critically on the diagnostic thresholds a study uses. A meta-analysis of several studies by Rakic et al. (2018), which included only studies following the case definition from Sanz and Chapple (2012)—that is, bone loss of at least 2 mm, positive probing bleed, and probing depth of at least 5 mm—found a pooled prevalence of 18.5% at the patient level and 12.8% at the implant level. Even within this defined framework, the heterogeneity was high (I² regularly above 75%), meaning that even with an identical case definition, the reported numbers vary widely.
Diaz et al. (2022) expanded the data set to 57 studies and confirmed the magnitude with a weighted prevalence of 19.53% at the patient level and 12.53% at the implant level. Their stratified comparison was crucial: studies that included probing depth as a diagnostic criterion reported a patient prevalence of 24.69%, while studies without probing depth as a criterion found a rate of 17.56%. Although this difference did not exceed the significance threshold (p = 0.27), it shows how much a single parameter in the diagnostic algorithm can shift the resulting number by almost one-third.
Dreyer et al. (2018) documented the extreme range: Reported prevalences at the implant level ranged from 1.1% to 85.0%, depending on which study measured with what definition, in what population, and with what maintenance profile. The median was 9.0% for patients undergoing regular prophylaxis (weighted by sample size to 10.9%) and rose to 18.8% for patients without regular follow-up. Simply asking whether the patient participates in a structured recall program shifts the observed frequency by a factor of two.
Mahardawi et al. (2023) confirmed in their analysis on the influence of keratinized mucosa that the prevalence ranged from 6.68% to 62.3% at the patient level depending on the study—a range of variation that remains clinically uninterpretable without knowing the underlying case definition and the threshold for marginal bone loss. Studies with a threshold of more than 3 mm bone loss systematically reported lower rates than those with a threshold of 2 mm.
A meta-analysis focusing on BOP (Bleeding on Probing) from multiple studies (2024) showed that even with a seemingly simple clinical marker like probing bleed, the diagnostic assignment remains uncertain: Only 26.5% of implants with BOP and 35.1% of patients with BOP actually had peri-implantitis. The prediction intervals—5.2% to 56% at the implant level and 6.4% to 71.5% at the patient level—underscore that the diagnostic value of individual clinical parameters varies enormously between studies.
A particularly instructive example is the comparison between the Sanz-Chapple definition (bone loss of at least 2 mm plus BOP plus probing depth of at least 5 mm) and the World Workshop 2017 proposal (BOP and/or suppuration plus bone loss of at least 3 mm plus probing depth of at least 6 mm in the absence of prior measurements). Simply raising the bone loss threshold from 2 to 3 mm and the probing depth from 5 to 6 mm selects a more severe spectrum of disease and automatically reduces the prevalence, even if nothing had changed in the biological reality of the population. Diaz et al. (2022) grouped their 57 studies into four categories based on diagnostic criteria and showed that Group 1 (BOP plus probing depth of at least 6 mm plus bone loss of at least 3 mm) consistently yielded lower prevalences than Group 2 (BOP plus probing depth of at least 6 mm plus bone loss of at least 2 mm). The meta-analysis of multiple studies in Group 2, which had the largest specific weight with 31 articles (53.45% of the total data set), yielded a patient prevalence of 19.6% in the fixed-effects model and 20.0% in the random-effects model.
Also noteworthy is the sensitivity analysis by Diaz et al. (2022): After excluding five studies with extreme effect sizes (Gatti, Koldsland, Marrone, Romandini, and Tey), the patient prevalence in the random-effects model dropped to 18.1 %, and the heterogeneity fell from I² = 87.2 % to a moderate 44.3 %. This suggests that a core group of studies with similar definitions and populations does converge, but the global heterogeneity is driven by outlier studies with differing definitions or cohorts. The Egger test for publication bias was unremarkable (LFK index 0.30), meaning that the variation is not explained by systematically missing small studies, but rather reflects the breadth of definitions.
Rakic et al. (2018) examined 29 studies involving over 10,000 implants and found that the prevalence rate at the patient level was consistently higher than at the implant level—a mathematically expected effect because a patient with multiple implants is counted only once as a case, but every affected implant counts individually in the denominator. This level of analysis—patient versus implant—is another definitional parameter that substantially shifts the reported number and is often not explicitly stated in professional discussions.
For clinical practice, this means: A peri-implantitis prevalence rate is only as informative as the definition behind it. Anyone who tells a patient or referring provider that the peri-implantitis rate is 20% without simultaneously stating the case definition, the cohort, and the follow-up status is communicating pseudo-precision.
What this does not imply: That peri-implantitis is rare or irrelevant. The risk issue is real and consistently supported by the entire body of literature included. But the exact number is not a fixed value; it is a product of definition. Clinicians should ask every time a prevalence rate is given: What thresholds were used? What population was studied? How was the follow-up organized?
In clinical communication, high prevalence numbers tend to either dramatize the problem or—paradoxically—dismiss it as inevitable. Both approaches are professionally unsound. The most appropriate stance is to communicate the range and provide the definitional context.
What does this mean for you? Different definitions yield very different prevalences.
For you as a patient, it is important to know: No diagnostic method is perfect. Research shows under which conditions a method is most reliable and when you should ask for a second opinion.
The scientific community has intensively studied this topic in recent years. For this article, more than nine scientific papers were evaluated. It is important to understand that not every study carries the same weight of evidence. Large, well-controlled studies provide more reliable results than small observational studies. The overall picture from these various studies is what we present to you here.
💡 What does this mean for you?
Different definitions yield very different prevalences. Discuss with your dentist at your next visit what this specifically means for your situation.
What do "Follow-up, Maintenance, and Cohort" mean for me as a patient?
When it comes to follow-up, maintenance, and cohort status, the research situation is clearer than many people think. Here you will learn what current studies actually show.
The observed peri-implantitis prevalence depends not only on the case definition but also strongly on how long the implants have been functional and how well the patients have received follow-up care. Dreyer et al. (2018) showed that the median implant prevalence for a function time of at least five years was 26.0 % (sample weighted 28.8 %), increasing to 21.2 % (sample weighted 38.4 %) for at least ten years—a finding suggesting that the cumulative burden increases as expected with longer observation.
At the same time, this same meta-analysis showed a dramatic difference between managed and unmanaged cohorts: The median for regular prophylactic participation was 9.0 % (sample weighted 10.9 %), compared to 18.8 % (sample weighted 8.8 %) with no recall, where the lower weighted number reflects the small sample size of an individual study.
Diaz et al. (2022) attempted to quantify the time effect, comparing studies with five to nine years of functional time against those with more than nine years. The prevalences were not significantly different at either the patient level (17.1% vs. 18.63%, p = 0.82) or the implant level (10.98% vs. 9.76%, p = 0.8). This initially contradicts an expected time effect but can be explained by the massive heterogeneity in study populations and definitions: the time effect is overlaid by a definition effect.
The All-on-four literature provides an example of selection bias: In 24 included studies with 11,743 implants, the cumulative loss rate was 1.5% (175 out of 11,743), with most losses occurring in the first year. Periimplantitis was reported as the second most common biological complication after implant loss, but without precise definitions or detailed case numbers. The high survival rate of 99.8% after more than 24 months reflects patient selection—well-controlled cohorts in specialized centers are not representative of general care.
A summary of multiple studies on the diagnostic value of probing bleeding, (2024) underscores the collective effect from a different perspective: Across more than 30 studies with 5,826 patients and 17,198 implants, the prevalence of periimplantitis in patients with BOP-positive sites ranged from 6.4% to 71.5%. This range cannot be explained by diagnostic inaccuracy alone—it is largely due to differences in study design, maintenance protocols, and patient selection.
Dreyer et al. (2018) differentiated risk factors with remarkably granular systematicity: For 111 identified putative risk factors and indicators, Forest Plots were created for 12 selected factors. In addition to smoking and diabetes, the absence of a prophylactic program and a history of gum disease (periodontitis) proved to be consistent risk factors. Conversely, age (effect summary OR 1.0, 95% CI 0.87–1.16), gender, and upper arch positioning were not associated with periimplantitis at the medium to high level of evidence. The importance of maintenance is further emphasized by the fact that the median prevalence in patients with fixed partial dentures was 9.6% (sample-weighted 9.6%)—a value that deviates significantly from the global median and shows the interaction between treatment type and disease frequency.
Another aspect of collective selection concerns the treatment model: Implants in specialized All-on-four protocols are typically used in high-volume centers with standardized surgical protocols and close postoperative monitoring. The systematic review included 24 studies with 11,743 implants, of which 175 were lost (cumulative loss rate 1.5%). The low rate of biological complications in this context should not be misinterpreted as general implant safety—it is the result of a highly selected population and a controlled environment. A history of gum disease (periodontitis) was sometimes explicitly listed as a contraindication in these studies, which makes direct comparison with population-based prevalence data methodologically inappropriate.
Mahardawi et al. (2023) added another dimension to the collective effect: In their subgroup analysis for patients undergoing regular implant maintenance, the absence of keratinized mucosa remained a significant risk factor for periimplantitis with an odds ratio of 2.08. In studies that adjusted for other variables, the odds ratio even increased to 3.68. This means: Even within a well-maintained population, soft tissue status remains an independent influencing factor that shifts the observed frequency of periimplantitis—a finding that underscores the need to incorporate collective characteristics into every statement about prevalence.
For clinical practice, this leads to a clear actionable consequence: Anyone who wants to estimate the periimplantitis prevalence in their own practice must check the basis for comparison. Data from university cohorts with ten years of follow-up and patients with a history of periodontitis are not transferable to young, healthy patients with short functional times.
What this does not imply: That maintenance makes no difference. On the contrary—the data consistently support that structured maintenance at least halves the observed periimplantitis burden. The clinical error is not in recognizing the risk, but in communicating it without context.
In patient communication, this means: The statement that every fifth implant gets periimplantitis is misleading without context. The same patient in a structured recall program has a significantly different risk profile than a patient with no follow-up care.
What does this mean for you? Population and maintenance shift the observed burden significantly.
For you as a patient, it is important to know: No diagnostic method is perfect. Research shows under what conditions a method is most reliable and when you should ask for a second opinion.
How do scientists arrive at these conclusions? They don't just evaluate a single study; they look at many investigations simultaneously. This allows them to determine whether a result was random or if it is consistently confirmed. In this case, the findings are based on 9 scientific papers from different countries and research groups.
💡 What does this mean for you?
Population and follow-up care can significantly shift the observed risk. Discuss with your dentist at your next visit what this specifically means for your situation.
What matters more: Diagnostic Signal or Crisis Narrative?
One point that often causes confusion is the diagnostic signal versus the crisis narrative. However, science has made important progress in recent years.
Bleeding on Probing (BOP) is considered a key clinical marker for peri-implant inflammation and is included as a necessary criterion in almost all definitions of peri-implantitis. A systematic review summarizing multiple studies on the long-term predictive value of BOP (2024) examined this assumption across over 30 studies with at least five years of follow-up. The result: Only about one quarter of implants positive for BOP (26.5 %, 95 % CI 21.2–32.1) and about one third of patients positive for BOP (35.1 %, 95 % CI 27.4–43.1) actually had peri-implantitis.
The prediction intervals were exceptionally wide: 5.2 % to 56 % at the implant level and 6.4 % to 71.5 % at the patient level. This breadth does not come from statistical weakness in a single study, but from the systematic heterogeneity in case definitions, populations, and follow-up protocols. The substantial heterogeneity (I² high in all subgroup analyses) confirms that BOP as a singular marker for peri-implantitis has a relevant false-positive rate.
Rakic et al. (2018) already argued that identifying adjuvant diagnostic markers is necessary to refine disease classification. Their meta-regression analysis showed that neither study design nor functional time significantly influenced the prevalence rate, but rather the implant surface characteristic: moderately rough surfaces were associated with a significantly lower prevalence (p = 0.011). This suggests that local factors on the implant itself can overshadow the diagnostic signal.
Mahardawi et al. (2023) added to this picture by including the influence of keratinized mucosa: In the overall analysis, the absence of keratinized mucosa increased the risk of peri-implantitis with an odds ratio of 2.78. In the subgroup with a uniform case definition (bone loss of at least 2 mm), the effect remained significant (OR 1.96, 95 % CI 1.41–2.73, p < 0.0001), but was significantly attenuated. This shows that what appears to be a clear risk factor is partially an artifact of differing definitions.
The literature on adjunctive antimicrobial photodynamic therapy in diabetics with peri-implantitis provides another piece of the puzzle: In this specific high-risk population, peri-implantitis was defined as an inclusion criterion, but the underlying diagnostics varied between studies. Thus, even in a seemingly homogeneous intervention study, definitional issues become a methodological confounder.
The clinical relevance of surface characteristics as a diagnostic confounder deserves special attention. Rakic et al. (2018) classified the implant systems used in their 29 included studies by surface roughness into three categories (minimal, moderate, and rough) and performed a meta-regression analysis. The result was clear: Moderately rough surfaces were significantly associated with a lower prevalence of peri-implantitis (p = 0.011), while neither study design nor functional time showed a significant influence. This finding has diagnostic implications: If the implant surface influences the risk of peri-implantitis, then a study that primarily includes implants with moderately rough surfaces measures a different biological reality than a study with predominantly rough or smooth surfaces—even if the clinical case definition is identical.
Diaz et al. (2022) also conducted an analysis by geographic scope and found that the reported prevalences varied among European, Asian, and North American studies without a clear systematic trend being discernible. However, their sensitivity analyses showed that after removing individual influential studies, the pooled estimates fluctuated only moderately (between 16.2% and 19.9% at the patient level), suggesting that the core finding—a periodontitis prevalence of approximately 15–20% at the patient level using criteria similar to Sanz-Chapple—is robust, even if individual studies show considerable variation. The meta-analysis summarizing multiple studies for the implant level showed similar robustness: the prevalence was 12.3% in the fixed-effects model and 11.5% in the random-effects model, with extremely high heterogeneity (I² = 97.21%).
The subgroup analyses by Mahardawi et al. (2023) regarding the influence of keratinized mucosa with different prosthetic types provide another instructive example of the interaction between diagnostic framework and clinical context: In patients with only fixed prostheses, the odds ratio for periodontitis when keratinized mucosa was absent was 2.82. This effect was stronger than in the overall analysis, suggesting that the type of restoration modifies the diagnostic association. For studies with mixed prosthetic collections (fixed and removable), the effect was also significant but more widely distributed. Overall, 16 cross-sectional studies included in the meta-analysis showed high heterogeneity (I² = 65%), but this was not driven by a single study—a pattern that reflects genuine biological differences between populations and settings.
For clinical practice, this means: Bleeding on probing (BOP) remains an indispensable screening marker but not a unique diagnostic feature. A positive BOP finding should always be interpreted alongside radiological evaluation and the clinical history before diagnosing periodontitis.
What this does not imply: That BOP is superfluous or unreliable. The false-positive rate is about two-thirds—meaning conversely, a negative BOP finding has relevant exclusionary value. The clinical error would be to use BOP alone as a diagnostic criterion and immediately assume periodontitis upon a positive finding.
In public and professional discussions, high periodontitis numbers are often instrumentalized as an alarm signal—following the pattern: Implants are more dangerous than thought. However, scientific evidence shows that a significant portion of these numbers is generated by the diagnostic framework, not solely by the disease reality. This is not a reason to relax caution, but a reason for diagnostic diligence rather than numerical drama.
What does this mean for you? The risk issue is real.
For you as a patient, it is important to know: No examination method is perfect. Research shows under what conditions a method is most reliable and when you should seek a second opinion.
What makes these results reliable? In medical research, the rule is: The more independent studies that arrive at the same result, the more certain the statement is. The type of study and the number of participants also play an important role in this.
💡 What does this mean for you?
The risk issue is real. Discuss with your dentist at your next visit what this specifically means for your situation.
Frequently Asked Questions
Here we answer the questions patients most often ask about this topic:
❓ What does "case definition" mean for me as a patient?
Different definitions produce very different prevalences.
❓ What does "follow-up, maintenance, and cohort" mean for me as a patient?
Population and follow-up significantly shift the observed burden.
❓ What matters more: Diagnostic signal or crisis narrative?
The risk issue is real.
❓ How certain are the results?
The scientific basis is solid, but not all questions have been definitively answered.
❓ Should I change my behavior based on this information?
Speak with your dentist before making any changes. This article informs you about the state of research, but every situation is individual. Your dentist knows your personal health status best.
❓ Where can I learn more?
You can find the full professional version of this article, with all study details, on Daily Dental Journal. For personal advice, please consult your dentist.
❓ What is the most important message of this article?
Periimplantitis is relevant, but its reported prevalence is definitionally fluid.
❓ Why are there differing opinions on this topic?
The literature debates the magnitude and definition range more than its existence.
🦷 When Should You See a Dentist?
Schedule an appointment with your dentist if:
- You have noticed something unusual and would like it checked
- You would like a second opinion on a diagnosis
- You are unsure if a recommended examination is necessary
- You have questions about the topics described in this article
- It has been more than a year since your last dental visit
Important: This article does not replace a dental visit. It helps you go into the conversation informed.
What You Can Do Yourself
Here are concrete steps you can take as a patient:
The Most Important Point in One Sentence
If a prevalence number seems high, the first question for a specialist is not how bad it is, but how it is defined.
Note on Source Material
This article is based on current scientific evidence and the DDJ editorial interpretation. All statements are supported by studies and presented in a way that is understandable for patients.
The content was prepared by the DDJ editorial team for patients. Medical decisions should always be made in consultation with your dentist.
Date: March 2026 · Language: American English (en-US) · Target Audience: Patients and interested laypeople