Influence of rheumatoid arthritis on peri‐implant diseases: A longitudinal retrospective clinical and radiographic evaluation

Abstract

Background

To investigate the effect of rheumatoid arthritis (RA) on the incidence of peri‐implantitis (PI) and peri‐implant mucositis (PIM).

Methods

Radiographic and clinical chart reviews were conducted to measure the probing depth (PD), bleeding on probing, and marginal bone loss (MBL) around the implants to diagnose peri‐implant diseases based on the 2017 workshop classification. Values were recorded at the baseline (T0) to the last available chart and radiograph (T1). Maintenance compliance was evaluated. Cases were followed longitudinally to detect the incidence of PI and PIM. Various potential confounders were controlled, including the total radiographic follow‐up time, chart‐based follow‐up time, number of maintenance visits, implant dimensions, history of periodontitis, bone graft, restoration angle, emergence, smoking, and diabetes mellitus. Chi‐square and Mann–Whitney tests evaluated categorical and continuous differences. Generalized estimating equations with a Tweedie distribution were applied. Binary logistic regression ascertained the odds ratio for binary outcomes.

Results

A total of 101 patients (50 RA and 51 non‐RA) with 124 implants were investigated. The mean follow‐up period for the implants was 5.38 ± 2.22 years. Implant survival rate was high at 96%. The RA group demonstrated a significantly higher PI (p = 0.024), while the non‐RA group showed a significantly higher PIM (p < 0.001). No significant differences were observed between both groups in implant survival and MBL.

Conclusion

Compared to the non‐RA group, RA patients demonstrated a similar incidence of MBL and implant survival rates and a significantly lower incidence of PIM; however, there was a significantly higher incidence of PI.

Plain Language Summary

In this study, we investigated the condition of dental implants in 50 patients with rheumatoid arthritis (RA) compared to 51 healthy controls over 5 years. Assessments from initial treatment to the last follow‐up visit included reviewing patient records and radiographs for signs of bleeding, probing depth, and bone loss. These measures helped diagnose peri‐implantitis (PI) and peri‐implant mucositis (PIM) based on the 2017 periodontal disease classification. The findings revealed a high implant survival rate in both groups (96%) with no significant difference in bone loss. However, the RA group showed a significantly higher incidence of PI than the healthy group that demonstrated PIM.

1. INTRODUCTION

Peri‐implant diseases have emerged as challenging complications of a multiconfounding nature that impact the implant success rate. ¹ , ² The issue is increasingly becoming a concern for public health because of its widespread nature and the resulting implications, such as the loss of implants and implant‐supported prostheses. ³ This situation is further compounded by the considerable expenses associated with implant preventive, surgical or nonsurgical interventive, and supportive care. ⁴ , ⁵ As a result, conducting a comprehensive implant risk assessment that considers the disease‐contributing local and systemic factors is essential ⁶ to improve the outcomes of implant therapy. ⁷

Autoimmune diseases such as diabetes mellitus (DM), systemic lupus erythematosus (SLE), rheumatoid arthritis (RA), and others that affect the periodontal tissue complex have been of major interest for clinical research in oral medicine and periodontics due to the well‐established relationship between them and periodontal diseases, ⁸ and between periodontal and peri‐implant diseases in terms of etiology, pathophysiology, and clinical outcomes, ⁹ , ¹⁰ , ¹¹ which allows linking such immune illnesses to the cause of peri‐implantitis (PI) and peri‐implant mucositis (PIM). ¹² These illnesses significantly exacerbate the host immune response against the peri‐implant plaque biofilm bacterial toxins, causing considerable tissue damage. ⁸

RA is recognized as a chronic autoimmune inflammatory disease that has been strongly linked to periodontal disease and associated tooth loss due to periodontal disease, as evidenced by several cohort and case–control studies. ¹³ , ¹⁴ , ¹⁵ , ¹⁶ The potential connection between RA and periodontal disease is attributed to the similarities in pathogenesis, risk factors, and host responses. ¹⁷ An intriguing aspect of the relationship between RA and periodontal disease is the biological intersection where causative bacteria, primarily Porphyromonas gingivalis and Aggregatibacter actinomycetemcomitans, induce protein citrullination—a process that leads to the production of pro‐inflammatory mediators, proteases (such as interleukin‐1, interleukin‐6, tumor necrosis factor‐alpha, and matrix metalloproteinases), and elements that activate the osteoclastic cascade (like receptor activator of nuclear factor‐kappa Β ligand). These factors are pivotal in eliciting an immune response responsible for collagen degradation and bone resorption. ¹⁷ , ¹⁸ Although this bidirectional causal relation between RA and periodontal disease is still considered hypothetical, the compelling evidence suggests that the presence of one disease may increase the risk of developing the other. ¹⁹ , ²⁰ , ²¹ , ²² , ²³

Currently, there is an increasing interest in exploring the potential association between peri‐implant diseases and RA. ¹² , ²⁴ , ²⁵ , ²⁶ However, most of them lack comprehensive accounting for the patient‐ and implant‐related covariates that are vital for patient risk assessment and can influence clinical decision‐making. Therefore, this study aims to examine the impact of RA on the incidence of PIM, PI, and marginal bone loss (MBL) while controlling for systemic and local factors, including smoking, DM, compliance with maintenance, and the quality of implant restorations. The secondary outcome was to assess implant survival in RA compared to non‐RA patients.

2. MATERIALS AND METHODS

Data were collected from electronic charts of patients who received periodontal treatment between January 2013 and January 2023 at the University of Michigan School of Dentistry, Ann Arbor, Michigan, USA, for retrospective evaluation. The study followed the Helsinki Declaration, and the manuscript preparation followed the STROBE (Strengthening the Reporting of Observational Studies in Epidemiology) guidelines. Ethical approval was obtained by the University of Michigan School of Medicine Institutional Review Board (IRBMED: HUM00228878).

2.1. Case definitions

Peri‐implant diseases were diagnosed following the 2017 World Workshop case definition. ²⁷ These were identified through longitudinal analysis, discriminating pathologic changes such as bleeding on probing (BOP) or MBL adjacent to the implant platform over time. The surviving implant was defined as an implant and its restoration that remained in the mouth, as indicated by the investigated chart or radiograph, with no implant loss reported in the patient notes. ²⁸

2.2. Inclusion and exclusion criteria

In this study, two groups were established: a test group of patients diagnosed with RA and a control group without RA. Patients included in the test group were diagnosed with RA based on a questionnaire asking specifically if their primary care physician (PCP) diagnosed them with RA. Patients who reported no PCP exam in the last 3 years were excluded. These patients had also received a successful bone‐level dental implant at the University of Michigan School of Dentistry in Ann Arbor, Michigan, USA.

Other inclusion criteria for both groups required the availability of a radiograph (periapical or vertical bitewings) taken at the time of implant restoration, another taken 6–12 months post restoration to allow for bone remodeling, and a follow‐up radiograph obtained 12 months after the implant was restored. Additionally, patients must have had a baseline periodontal chart for the implant site 6–12 months post loading and the latest periodontal chart for the implant site, considering the clinical and radiographic follow‐up period as a confounding factor. The same inclusion criteria were applied to the non‐RA control group. Patients who were smokers, suffered from DM, or had a history of periodontitis were all included, ²⁹ and those systemic factors were considered confounding factors (see Tables S1–S3 in the online Journal of Periodontology).

Exclusion criteria for both groups initially encompassed the absence of either baseline or follow‐up radiographs, incomplete periodontal charting at baseline or during follow‐up, missing relevant medical information, or inadequate information regarding implant dimensions. Radiographs that were severely angulated or of low quality were also excluded. ³⁰ Cases with implant restorations placed later than 6–12 months post loading or where the restoration was removed during the follow‐up period were excluded from the study. Finally, due to the inability to access the complete medical profiles of patients to check for other diseases that might elevate the rheumatoid factor and potentially lead to misdiagnosis—such as SLE, liver illnesses, and Sjögren's syndrome—these patients were excluded from our sample.

2.3. Radiographic evaluation

The radiographic evaluation included images taken at baseline (suprastructure in place) that clearly allowed for the identification of the implant's platform as a reference point and distinct visualization of implant threads for the best assessment of mesial and distal bone levels in relation to the platform. ³¹ To reduce the inherent radiographic dimensional errors, the selected radiographs were the ones taken by the parallel technique using the aiming device. To overcome the vertical distortion, recalibration of the radiographic image was done depending on the real dimensions of the existing implant (Figure 1). A dedicated radiographic toolkit^* was used to measure the bone loss around the mesial and distal sides by careful detection of the implant's platform using contrasting tools and drawing a vertical measuring line from the platform edge to the deepest visible point of marginal bone in contact with the threads. The deepest bone resorption was chosen for different resorbed bone portions. To ensure reproducibility, one examiner (P.N.) evaluated the radiographs. Intraexaminer reliability was assessed. The blinded examiner reached an intraoperative k value of >80% based on a previous examination of 20% of the overall sample. ³²

FIGURE 1.

Radiographic evaluation. To overcome vertical distortion, the radiographic image was recalibrated depending on actual implant length, and radiographic length was measured from the implant apex to the visible platform. Measurements for mesial and distal bone loss were then taken from the deepest point of resorption to the platform.

2.4. Data collection

Electronic charts of patients who met the eligibility criteria were retrospectively screened by two masked examiners (H.M., E.K.). At the patient level, the following characteristics were collected: age at the time of implant placement, sex, DM status, smoking status (nonsmoker, former smoker, or current smoker, and number of cigarettes/day), and RA status (diagnosed with or without RA). At the tooth level, BOP and probing depth (PD; worst site) were collected from the periodontal chart at T0 and T1, where T0 is the baseline chart 6–12 months after implant loading and T1 is the last available chart of the implant. Radiographic data were collected at T0 and T1, and bone loss mesial and distal of the implant was measured at T0 and T1. The number of successful periodontal and peri‐implant maintenance visits was collected between the T0 and T1 for both the periodontal charts and the radiographs to evaluate the patient's compliance. The implant restoration was evaluated using radiographs to assess the satisfaction level and gauge the contour angle. ³³ An implant restoration was considered satisfactory when the radiograph showed no significant open margins, was well fitting, and had no excess cement or open contacts in addition to an emergence angle of less than 30 degrees. ³⁴ , ³⁵

2.5. Statistical analysis

Statistical analysis was completed using specialized statistics software^† to compare dental implant outcomes in patients with RA versus healthy controls. Our primary outcomes were the incidence of PI and PIM, while the secondary outcome was the implant survival rate and MBL. Data were presented in frequencies, percentages, and means with standard deviations. Chi‐square and Mann–Whitney tests were used to evaluate categorical and continuous differences.

Generalized estimating equations (GEE) were initially employed with a gamma distribution to account for the skewed distribution of bone loss variables. Potential confounders such as total radiographic follow‐up time for PI, chart‐based follow‐up time for PIM and other outcomes, number of maintenance visits, implant length, and implant diameter were adjusted for. Binary logistic regression was used to determine odds ratios (OR) for binary outcomes, with all tests conducted at a 5% significance level.

Given the presence of multiple implants in approximately 20% of patients, a within‐subject correlation was introduced to analyze the data. Additional confounding factors that could influence bone loss, PI, and PIM, including implant dimensions, bone graft use, restorative emergence profile, and mesial/distal restorative angle, were also incorporated into the analysis (Tables 1, 2, and 3). GEE with a Tweedie distribution were applied to accommodate bone loss variables that included zeros and were not normally distributed. Two unstructured covariance matrices were used: one for repeated measures of bone loss variables and another for within‐subject correlation. Potential confounders such as total radiographic follow‐up time for PI, chart‐based follow‐up time for PIM and other outcomes, number of maintenance visits, implant length, implant diameter, history of periodontitis, smoking, and diabetes were adjusted for.

TABLE 1.

Patient‐level characteristics.

				Arthritis
		Total patients (N = 101)		RA (n = 50)		Non‐RA (n = 51)
Variables	Categories	n	%	n	%	n	%	p value
Sex	Male	49	48.5	29	58.0	20	39.2	0.059
	Female	52	51.5	21	42.0	31	60.8
Age (years)	≤60	27	26.7	15	30.0	12	23.5	0.536
	61–70	46	45.5	20	40.0	26	51.0
	71+	28	27.7	15	30.0	13	25.5
Current smoker a	Yes	14	13.9	8	16.0	6	11.8	0.538
	No	87	86.1	42	84.0	45	88.2
Former smoker a	Yes	31	30.7	17	34.0	14	27.5	0.522 b
	No	70	69.3	33	66.0	37	72.5
Diabetes a	Yes	6	5.9	1	2.0	5	9.8	0.362 b
	No	95	94.1	49	98.0	46	90.2
Other systemic conditions	Yes	63	62.4	33	66.0	30	58.8	0.457
	No	37	37.6	16	34.0	21	41.2
History of periodontitis b , c	Yes	8	7.9	2	4.0	6	11.8	0.269 b
	No	93	92.1	48	96.0	45	88.2

Abbreviation: RA, rheumatoid arthritis.

^a Confounding factors.

^b Fisher's exact test.

^c History of periodontitis investigated at the time of the implant placement.

TABLE 2.

Implant‐level characteristics.

				Arthritis
		Total patients (N = 124)		RA		Non‐RA
Variables	Categories	n	%	n	%	n	%	p value
Implant location	Maxilla	58	46.8	28	50.0	30	44.1	0.514
	Mandible	66	53.2	28	50.0	38	55.9
Implant brand	Nobel	35	28.2	21	37.5	14	20.6	0.094
	Zimmer	79	63.7	32	57.1	47	69.1
	Others	10	8.1	3	5.4	7	10.3
Implant site	Molar	62	50	26	43.3	36	56.2	0.151
	Nonmolar	62	50	34	56.7	28	43.8
Graft with implant a	No	87	72.5	46	85.2	41	62.1	0.005
	Yes	33	27.5	8	14.8	25	37.9
Satisfactory restoration	Satisfactory	114	91.9	53	94.6	61	89.7	0.348 b
	Unsatisfactory	10	8.1	3	5.4	7	10.3
Mesial emergence profile a	Convex	55	48.7	26	53.1	29	45.3	0.406
	Concave	40	35.4	14	28.6	26	40.6
	Straight	18	15.9	9	18.4	9	14.1
Distal emergence profile a	Convex	59	52.2	33	67.3	26	40.6	0.011
	Concave	30	26.5	7	14.3	23	35.9
	Straight	24	21.2	9	18.4	15	23.4
Mesial restoration angle a	Mean (SD)	31.7	(10.5)	32.1	(11.0)	31.5	(10.2)	0.750 c
Distal restoration angle a	Mean (SD)	29.7	(11.7)	31.5	(11.7)	28.3	(11.6)	0.148 c

Abbreviation: RA, rheumatoid arthritis.

^a Confounding factors.

^b Fisher's exact test.

^c p value calculated based on t test (data normally distributed).

Bold values are statistically significant.

TABLE 3.

Confounding factors.

Variables	Arthritis	Valid n	Median	Q1	Q3	p value
Number of maintenance visits (survival) a	Yes	50	4.00	2.00	7.00	0.105
	No	48	5.00	2.00	9.00
Number of maintenance visits (radiographic) a	Yes	50	4.00	2.00	6.00	0.058
	No	48	6.00	2.00	9.50
Radiographic follow‐up (years) a	Yes	50	3.39	1.83	5.48	0.997
	No	51	3.19	1.93	5.32
Chart‐based follow‐up (years) a	Yes	50	4.94	3.91	6.88	0.744
	No	51	4.72	3.45	7.13
Pocket depth ≥ 4 mm	Yes	48	8.0	3.5	17.5	0.550
	No	48	10.0	3.5	22.5
Total bleeding score (%)	Yes	48	10.0	4.5	20.0	0.866
	No	48	11.0	4.5	18.5
Implant diameter	Yes	56	4.3	3.9	4.7	0.372
	No	68	4.1	3.7	4.7
Implant length	Yes	56	11.5	10.0	11.5	0.569
	No	68	11.3	10.0	11.5

Note: p value using Mann–Whiney test (non‐normal data).

Abbreviations: Q1, first quartile; Q3, third quartile.

^a Confounding factors.

Using an unstructured covariance matrix for within‐subject correlation, GEE models with binary logistic regression were utilized to ascertain the OR for binary outcomes. All tests were conducted at 5% significance. To ensure the necessary statistical power to detect a clinically meaningful difference in the chance of PIM between arthritis and nonarthritis groups, a power analysis was conducted.^‡ Assuming a significant level (alpha) of 5%, a power of 80%, and an effect size OR of 2.25 (moderate), the required sample size was 52 participants per group for a total sample size of 104 patients.

3. RESULTS

3.1. Characteristics of study population

The presented cohort included 101 patients with a total of 124 implants. Of these, 50 had RA and 51 were non‐RA. The patients' mean age was 65.6 ± 10.2 years, and almost half of them, 52 (51.5%), were female. Fourteen percent were current smokers, and 31% reported as former smokers. Only 5% were diabetic, and 8% had periodontitis. The mean follow‐up period for the implants was 5.38 ± 2.22 years. Regarding group homogeneity, there were no significant differences in patients' characteristics between the two groups, except that the RA group had more males. Tables 1 and 2 show the patient‐level and implant‐level characteristics.

3.2. Data analysis

3.2.1. PIM and PI

There was a significant difference between the RA and non‐RA groups regarding the incidence of developing PIM. PIM was more prevalent in the non‐RA group (60.3% vs. 16.1%; p < 0.001). The RA group exhibited a significantly lower risk with an OR of 0.12 and a p value of less than 0.001 compared to the non‐RA group. However, patients with periodontitis had a marginally significantly higher chance than those without periodontitis (aOR = 1.63; p = 0.059) (see Tables S4 and S5 in the online Journal of Periodontology).

When looking at the PI outcome, the initial statistical model demonstrated that none of the independent variables reached statistical significance, suggesting that there were no differences between the two groups in this regard. However, the inclusion of other variables, in addition to considering the within‐subject correlation being analyzed by the GEE with a Tweedie distribution, showed that the RA group had a significantly higher incidence of PI than the non‐RA group (aOR = 3.18; p < 0.024). Moreover, patients who had a graft with the implant had a significantly higher incidence of PI than those without a graft (aOR = 7.8; p = 0.005). With every additional year of radiographic follow‐up, the risk of developing PI increases by 83% (aOR = 1.83; p = 0.010). A positive correlation was observed between the frequency of maintenance visits and the likelihood of PI (aOR = 1.30) (see Tables S6 and S7 in the online Journal of Periodontology).

3.2.2. Marginal bone loss

After adjusting for the confounding factors, no significant difference was observed in the MBL (calculated based on the mesial and distal MBL) between the two studied groups. However, current smokers demonstrated significantly higher MBL compared to nonsmokers (B = 0.67; p = 0.035) (see Tables S8–S13 in the online Journal of Periodontology).

3.2.3. Implant survival and compliance

The overall survival rate for all implants was 96%; only five implants did not survive, which makes the comparison unreliable. Based on the new statistical model, patients who received bone grafts with implants had a lower chance of survival than those without grafts (aOR = 0.07; p 0.022). A significant association between restoration angle and implant survival was found. A less obtuse restoration angle was associated with a higher likelihood of implant survival (aOR = 1.18; p < 0.001) (Table 4; Tables S14 and S15 in the online Journal of Periodontology).

TABLE 4.

Primary and secondary outcome—clinical measurements.

				Arthritis
		Total patients		RA		Non‐RA
Variables	Categories	n	%	n	%	n	%	p value
PIM	Yes	50	40.3	9	16.1	41	60.3	<0.001
	No	74	59.7	47	83.9	27	39.7
PI	Yes	26	21.0	14	25.0	12	17.6	0.317
	No	98	79.0	42	75.0	56	82.4
Implant survival	Yes	119	96.0	54	96.4	65	95.6	1.000 a
	No	5	4.0	2	3.6	3	4.4

Abbreviations: PI, peri‐implantitis; PIM, peri‐implant mucositis; RA, rheumatoid arthritis.

^a Fisher's exact test.

Bold value is statistically significant.

Although there was no significant difference in the follow‐up period between the two groups, the non‐RA group had higher compliance to maintenance than the RA group (mean of chart‐based compliance was 6.97 ± 6 years in non‐RA group and 4.16 ± 3.3 years in RA group; mean of radiographic‐based compliance was 6.86 ± 5.7 years in non‐RA group and 4.25 ± 3.4 years in RA group).

4. DISCUSSION

This study investigated the relationship between RA and peri‐implant diseases by evaluating the peri‐implant status of a group of patients diagnosed with RA, compared with non‐RA controls, over a mean follow‐up period of about 5 years. Patients in the RA group demonstrated a significantly higher likelihood of PI compared to those without RA, with an OR of 3.18. This effect was amplified in patients who received bone grafts in conjunction with implants. While there were no substantial differences in implant survival rates or MBL between the two groups, current smokers exhibited higher levels of MBL. The survival rate of implants was positively correlated with the restoration angle. However, the incidence of PI was lower in the RA group.

While the relationship between peri‐implant diseases and RA remains unconfirmed in the literature, ³⁶ , ³⁷ RA has shown similarities to periodontitis in previous research. ³⁸ , ³⁹ Previous studies have associated RA with severe periodontitis as patients with RA often exhibit greater alveolar bone and attachment loss. ⁴⁰ Gonzalez et al. found that increased alveolar bone loss correlated with elevated levels of anticitrullinated protein antibodies, which are commonly found in RA patients. ⁴¹ To assess the association between implant‐related adverse events (RA and PI) and patient characteristics, we initially employed a GEE analysis with a gamma distribution. However, this analysis failed to yield statistically significant results, likely due to confounding factors and within‐subject correlation arising from patients having multiple implants. To address these limitations, we implemented GEE with a Tweedie distribution. ⁴² This is particularly well suited for analyzing repeated measurements within individuals, such as in cases where multiple implants are present. By accounting for within‐subject correlation and adjusting for confounding variables, this GEE model provides a more robust and powerful approach to detecting significant associations. ⁴³ Consequently, our GEE analysis revealed a significant relationship between RA and PI in the study population. As demonstrated in a previous study by Lipsitz et al., GEE is an effective method for analyzing repeated categorical data. In their arthritis clinical trial, ignoring within‐patient correlations led to an underestimation of standard errors and potentially misleading inferences. By incorporating the appropriate correlation structure, Lipsitz and colleagues were able to obtain reliable estimates of treatment effects. Similarly, our GEE analysis provided a more accurate assessment of the association between RA and PI in the context of multiple implants and confounding factors. ⁴³

Our investigation represents the first study to establish a significantly elevated risk of PI in patients with RA relative to healthy controls. Previous research, such as the systematic review by Turri et al., has yielded inconclusive results regarding the association between RA and PI. While Turri et al. found no overall association, their analysis included limited studies specifically examining RA. ³⁶ Moreover, the retrospective study by Renvert et al. failed to identify a link between RA and PI, potentially due to the insufficient consideration of confounding factors (only age, smoking, and sex). ⁴⁴ In contrast, our study comprehensively addressed a broader range of potential confounders, including radiographic follow‐up, grafting procedures, implant emergence profile, and restoration angle, a recently identified risk factor for PI in bone‐level implants. ³⁵

The systemic bone loss associated with RA is multifaceted, being potentially affected by a variety of factors such as glucocorticoid therapy, reduced physical activity, and the disease's pathophysiology itself. ⁴⁵ Our analysis indicates that individuals ≤70 years old experience less MBL as compared to those >70 (p = 0.004 for both age groups ≤60 and 61–70 years).

In our cohort, the non‐RA group showed higher compliance to dental maintenance visits than the RA group, which could be linked to the fatigue commonly experienced by RA patients, affecting their physical capabilities and psychosomatic state. ⁴⁶ According to a review by Brus and colleagues, RA patients generally demonstrate low compliance rates. However, compliance was variably defined across the studies. ⁴⁷ Unlike previous research, our study explicitly investigates adherence to maintenance visits among RA patients with PI, a relatively unexamined issue. Although some studies report RA patients being less likely to attend preventive dental care appointments annually, ⁴⁸ a study by Serban et al. revealed that RA patients often found it more difficult to maintain oral hygiene routines during disease flare‐ups. Notably, older patients expressed more negative attitudes toward their oral health and dental appointments. ⁴⁸ In contrast, a study by Juan et al., conducted in 2022 on a Taiwanese cohort, observed a higher frequency of dental visits among RA patients. ⁴⁹

Interestingly, our study revealed that the odds of developing PIM were lower in the arthritis group (OR = 0.12; p < 0.00), with the likelihood increasing over the follow‐up period. This finding is somewhat unexpected given the pathogenesis of RA, which might imply a more severe impact on oral health. Supporting our observation, existing evidence by Krennmair et al. in 2010 revealed that implants placed in patients with RA showed acceptable levels of marginal bone resorption (mean 2.1 ± 0.5 mm), PD (mean 2.8 ± 3.2 mm), and healthy soft‐tissue conditions (plaque, bleeding, gingival index Grade 0 in 80%). Their study also concluded that RA patients did not experience a higher rate of implant loss compared to patients without RA. ⁵⁰

The reduced incidence of PIM in the RA group might seem contradictory at first glance. However, Tar et al. reported that patients with RA exhibited more gingivitis and significantly higher BOP than a non‐RA control group. ⁵¹ This discrepancy could be explained by the use of certain therapeutic agents in RA treatment, such as disease‐modifying antirheumatic drugs (DMARD) and glucocorticoids, which may impact oral health due to the shared inflammation pathways between the diseases and the anti‐inflammatory effects of these medications. ⁴⁹ A recent comprehensive review by Petit et al. (2024), which analyzed 35 studies, concluded that DMARD monotherapy may, to some extent, improve the periodontal condition of patients with RA who also have periodontal disease. ⁴⁹ Supporting this, Mayer et al. found that RA patients undergoing antitumor necrosis factor‐alpha therapy had lower periodontal indices (gingival recession, PD, and clinical attachment loss [CAL]). ⁵²

Underlying autoimmune diseases can lead to more pronounced marginal bone resorption and bleeding, affecting implant stability and success. ⁵⁰ While our study accounted for multiple variables, it was unable to fully assess the effect of the underlying autoimmune condition. In our cohort, only five implants failed—two from the RA group and three from the non‐RA group. This resulted in a survival rate of 96%, which aligns with the rates reported in the existing literature. ⁵² Alsaadi et al., in 2008, also found retrospectively that systemic health factors did not seem to significantly impact the etiology of late implant loss. ⁵³ Similar findings were presented in a systematic review by Guobis et al. in 2016, which reported no significant correlations between RA and implant success in the investigated studies. ³⁷ More recent literature, including a systematic review by Esimekara et al. in 2022, which assessed dental implants in patients with autoimmune diseases, found that all studies specifically analyzing RA reported an implant survival rate of 100%. ⁵⁴

Our retrospective study contributes to the limited literature investigating peri‐implant disease in patients with RA. We employed the 2017 World Workshop case definitions for PI and PIM and controlled for several confounding factors. However, the retrospective design introduces limitations. Our sample size was not precalculated and included all eligible RA patients, potentially limiting generalizability. Furthermore, the analysis did not consider the severity of RA, only its presence. This is a significant limitation as RA disease activity and progression can influence peri‐implant outcomes. A follow‐up period of 5.38 ± 2.22 years may be insufficient to accurately assess the long‐term incidence of PI. Likewise, the study's underpowered design and nonrepresentative sample further restrict the generalizability of its findings.

Extended, long‐term prospective studies are needed to evaluate biological complications and implant failure rates in RA patients comprehensively. Such studies should consider RA disease severity and progression and employ standardized methodologies for assessing peri‐implant outcomes. The severity of RA is believed to influence the extent of periodontal disease. ⁵⁵ , ⁵⁶ Consequently, if RA among our patients was well controlled, the periodontal outcomes might reflect those of the general population, with the premise that advanced RA stages may have a greater impact. ⁵⁷ Additionally, our study's findings are constrained by incomplete records regarding important potential confounding factors like the specific medications RA patients were taking and the glycemic levels for diabetic patients, such as fasting sugar levels or glycated hemoglobin levels. This lack of data is exacerbated by the exclusive use of electronic health records, which are not always meticulously updated and may omit detailed information. The type of RA medication patients use is another potential confounding factor that could have influenced the results.

Future studies should aim to assess disease severity and its manifestation at implant sites by analyzing biochemical markers. They should also consider factors such as underlying diseases, osteoporosis, and immunosuppressant medication. Li in 2021 reported a potential relationship through crosstalk genes (FCGR2B and CD14). ⁵⁷ Additionally, Alenazi et al. in 2021 examined the level of rheumatoid factor in crevicular fluid. They found that elevated rheumatoid factor levels could serve as diagnostic markers for peri‐implant complications in patients with RA. ⁵⁸

5. CONCLUSION

Despite the limitations of this longitudinal retrospective study, we can conclude that patients with and without RA showed similar rates of implant survival and no significant difference in the MBL. However, the RA group showed a significantly higher incidence of PI compared to the healthy group, which demonstrated a significantly higher incidence of PIM. Further interventional longitudinal studies are necessary to comprehensively evaluate the incidence of biological complications and implant failure in individuals with RA.

AUTHOR CONTRIBUTIONS

Hamzeh Almashni: Project administration (lead); investigation (equal); formal analysis (supporting); writing—original draft (lead). Era Kakar: Investigation (equal); writing—original draft (supporting). Paolo Nava: Investigation (equal). Hom‐Lay Wang: Writing—review and editing (supporting). Muhammad H. A. Saleh: Conceptualization (lead); methodology (lead); writing—review and editing (lead). All authors read and approved the final manuscript and agreed to be accountable for all aspects of the work.

CONFLICT OF INTEREST STATEMENT

This paper was partially supported by the University of Michigan Periodontal Graduate Student Research Fund. The authors declare that there are no conflicts of interest to disclose.

Supporting information

Supporting Information

Almashni H, Kakar E, Nava P, Wang H‐L, Saleh MHA. Influence of rheumatoid arthritis on peri‐implant diseases: A longitudinal retrospective clinical and radiographic evaluation. J Periodontol. 2025;96:933–943. 10.1002/JPER.24-0376

DATA AVAILABILITY STATEMENT

The data that support the findings of this study are available from the corresponding author upon reasonable request.