Under pressure—mechanisms and risk factors for orthodontically induced inflammatory root resorption: a systematic review

Summary Background The application of orthodontic forces causes root resorption of variable severity with potentially severe clinical ramifications. Objective To systematically review reports on the pathophysiological mechanisms of orthodontically induced inflammatory root resorption (OIIRR) and the associated risk factors based on in vitro, experimental, and in vivo studies. Search methods We undertook an electronic search of four databases and a separate hand-search. Selection criteria Studies reporting on the effect of orthodontic forces with/without the addition of potential risk factors on OIIRR, including (1) gene expression in in-vitro studies, the incidence root resorption in (2) animal studies, and (3) human studies. Data collection and analysis Potential hits underwent a two-step selection, data extraction, quality assessment, and systematic appraisal performed by duplicate examiners. Results One hundred and eighteen articles met the eligibility criteria. Studies varied considerably in methodology, reporting of results, and variable risk of bias judgements. In summary, the variable evidence identified supports the notion that the application of orthodontic forces leads to (1) characteristic alterations of molecular expression profiles in vitro, (2) an increased rate of OIIRR in animal models, as well as (3) in human studies. Importantly, the additional presence of risk factors such as malocclusion, previous trauma, and medications like corticosteroids increased the severity of OIIRR, whilst other factors decreased its severity, including oral contraceptives, baicalin, and high caffeine. Conclusions Based on the systematically reviewed evidence, OIIRR seems to be an inevitable consequence of the application of orthodontic forces—with different risk factors modifying its severity. Our review has identified several molecular mechanisms that can help explain this link between orthodontic forces and OIIRR. Nevertheless, it must be noted that the available eligible literature was in part significantly confounded by bias and was characterized by substantial methodological heterogeneity, suggesting that the results of this systematic review should be interpreted with caution. Registration PROSPERO (CRD42021243431).


Introduction
Orthodontic tooth movement (OTM) is achieved by combining mechanical strain application and subsequent physiological adaptation of the periodontium (1).Crucially, the molecular and cellular responses of the periodontium to both tension and compression are inflammatory in nature, leading to local alterations in the surrounding environment that affect the blood supply, the release of various neurotransmitters, growth factors, and cytokines (2,3).
Periodontal ligament cells (PDLs) are necessary mechanoreceptors and primary shock absorbers that respond to constant and intermittent loading (4).In addition, PDL cells are essential for periodontal remodelling, bone resorption, and tooth movement (4,5).The activation of PDL cells leads to characteristic expression profiles, ultimately driving altered cellular responses (6,7).
Orthodontically induced inflammatory root resorption (OIIRR) is a clinically challenging, well-documented side effect of OTM (8).It is understood as a multifactorial pathological process that results in the loss of permanent root surface structure with varying severity (1,9).In advanced cases, OIIRR can have a detrimental impact on the long-term prognosis of affected teeth leading to tooth loss as a worst-case scenario-impacting patients' quality of life (10,11).
Therefore, it is of critical importance for clinicians to understand why OIIRR occurs and which patients are, based on these mechanisms, most at-risk of severe OIIRR.A plethora of research focuses either on the underlying mechanisms of OIIRR or the risk factors that can predispose to severe resorption, but typically, the biological mechanisms and the clinical factors are seldom combined (9,12).
The treatment-related or intervention-related are, for example, duration of treatment (22), direction and magnitude of force (54,19), intermittent versus continuous forces (20), the distance of movement (55), type of appliance (56) a history of tooth extraction (57), treatment mechanics (58), intentionally injected materials or medication to alter root resorption (60,59) and others.These factors can interact with each other and result in the development of the phenotype.
Therefore, we sought to systematically review reports on the pathophysiological mechanisms of OIIRR and the linked systemic and local clinical risk factors for OIIRR based on in-vitro, in-vivo animal, and human studies.

Study protocol and registration
Before undertaking this review, we created a specific protocol in line with the PRISMA guidelines and checklist (61,62).The study was registered with the International Prospective Register of Systematic Reviews (PROSPERO reference CRD42021243431).

Search sources and strategy
The following electronic databases were searched electronically for peer-reviewed publications (The searches were performed on 23 March 2021, and an updated search was conducted on 31 May 2022): • MEDLINE (Medical Literature Analysis and Retrieval System Online via PubMed).

Study selection
First, titles and abstracts were screened by two independent reviewers (HD, VC).Abstracts that-according to the two reviewers-met all inclusion and exclusion criteria were considered.An initial calibration exercise was undertaken.The inter-rater reliability was calculated (percentage of agreement).Any disagreement was referred to a third reviewer (AK) and discussed to reach an agreement.For the next step-the full-text search was conducted independently by HD and VC.Any difference of opinion was consulted with a third reviewer (AK) for the final decision.

Data extraction
Two reviewers (HD, VC) independently performed duplicate data extraction utilizing a pre-established and trialed spreadsheet as recommended by the Cochrane Handbook (63).In addition, data figure software (Plot digitiser) extracted the numerical data from the figures when it was not available in the text (64).

PE/ICO1a&b
The main outcome is • The genomic changes of PDL cells after mechanical force application.
• The additional outcomes are: relationship between the measured molecular and genetic expression and root resorption mechanism.

PE/ICO2a&b
The main outcome was: • Incidence of OIIRR The additional outcomes were: • Gingival index • Periodontal disease incidence • Molecular biomarkers • Histological findings

PICO3a&b
The main outcome was: • Incidence of OIIRR.

Overall Quality of Evidence Assessment (GRADE)
The quality of evidence was evaluated according to the Grading of Recommendations, Assessment, Development and Evaluations (GRADE) approach (162); Higgins et al. (161).
According to the GRADE methodology (161), the certainty level was downgraded by one level for serious concerns and two for very serious concerns in five domains: RoB, inconsistency, indirectness, imprecision, or publication bias.Vice versa, they were upgraded by one or two levels if large magnitudes of effects were recorded, there was evidence of a dose-response gradient, or plausible confounders would have reduced the demonstrated outcomes or all residual confounding and biases would have increased the effect if no effect has been observed.

Study selection
A total of 7824 studies were identified through electronic searches.However, 6027 hits remained after the removal of duplicates.After title and abstract screening, a total of 253 full texts were assessed for eligibility using full-text assessment.Overall, 113 individual studies, some applicable for more than one PICO, were included (Figure 1 PRISMA flowchart).A list of excluded articles in the fulltext stage, including reasons for exclusion, can be found in Supplementary Table 2.

Inter-rater agreement
The inter-rater raw agreement rate during the screening stage of the titles and the abstracts was 95.8%.
The Kappa correlation coefficient value was (k = 0.817) for the risk of bias analyses (72), which is considered very good.The inter-rater agreement was 89.1%.

Overall quality of evidence (GRADE)
The quality of evidence was evaluated using the GRADE approach.As no meta-analysis was undertaken, no test for publication bias was performed.Forest and funnel plots were not constructed due to high heterogeneity amongst study designs, methods used, and selected outcome measures.

Risk of bias
Thirteen in vitro studies were well-formulated and received high scores; all studies scored between 14 and 19 out of 20 points (Supplementary Table 3).

Outcomes of included studies
The main outcome was reported as increasing or decreasing in fold change of gene expression compared to the control group.The goal measurement of these studies was total RNA extract, mRNAs, miRNAs, or lncRNAs expressions and their interaction networks and pathway activation.
The summary of the included studies is listed in Supplementary Table 4. Detailed studies' descriptions and outcomes are presented in Supplementary Table 5.The most upregulated and downregulated genes are shown in Supplementary Tables 6 and 7, respectively.No secondary outcomes were found.

PE/ICO 1B Are molecular expression profiles [O] different in human periodontal ligaments cell culture models of OTM with active mechanical forces [P] with possible risk factors [E] compared to cell culture models without risk factors?
No studies were found related to this PE/ICO.

General characteristics
In this PICO, 24 articles were selected for data extraction, 1 on mice (91) 11,112).They were published between 1997 and 2021.All used a mesial tipping force on molars with close coil springs, except two utilized an expansion force (98,103).Follow-up times ranged between 14 and 35 days.

Risk of bias
The (SYRCLE) tool was used for PICO 2A and PI/ECO 2B.However, most authors rarely mentioned if allocation was randomized or if the animals were kept in random houses or commented on the blindness of the assessors, resulting in unclear certainty in many domains (Supplementary Table 8).

Outcomes of included studies
All results showed that force-tested groups have a significantly higher risk of OIIRR than the no-force control groups (

Risk of bias
Please refer to section PICO 2A RoB, as the same tool was used for PE/ICO 2B.

Outcomes of included studies
Higher risks of OIIRR were found in groups with high bone turnover, acute corticosteroid intake than chronic corticosteroid intake or control groups, 4-hexylresorcinol medication intake, hypofunctional periodontium, nicotine, and strontium ranelate (118,119,126,128,108,109). Significantly less OIIRR was reported in animals with a chronic combined oral contraceptive, baicalin intake, sites with previous bone ceramic or bio osseous graft, systemic sodium fluoride, immature roots, osteopontin (OPN) deficiency, angiogenic cytokines, high caffeine, ovariectomy, and orchiectomy rats (91,97,2,120,102,123,124,105,125,127).
Studies investigating the effects of a diet with high casein phosphopeptides (CPP), high zinc in the diet, carboxymethyl cellulose and atorvastatin medication, lacunae size in amitriptyline medication, cetirizine intake, zoledronic acid, etoricoxib medicine, fluoxetine, budesonide, ovalbumin and in systemic fluoride did not report significant differences between test and control groups (92,93,113,114,95,116,117,10,121,122,106).
In terms of secondary outcomes, 19 studies reported histological and/or biological findings (93,114,95,91,118,10,119 ,2,120,102,121,122,105,125,106,126,127).The summaries of the included studies are listed in Supplementary Table 11.Detailed studies' descriptions and outcomes are presented in Supplementary Table 12.
High quality of evidence assessed via GRADE Methodology indicates that orthodontic forces in PICO 2A and/or intervention or exposure in PI/ECO 2B are associated with the incidence or increased severity of OIIRR in animal studies (Supplementary Table 13).

General characteristics
Nine articles were selected for data extraction (13,129,134 ,135,130,36,131,132,133).All articles were published between 1986 and 2021.The follow-up varied from 28 days to 6 months.

Risk of bias
Six studies were RCTs (129,130,131,133), two with low risk (131,133), and four were of some concern.One CCT was assessed to have a moderate RoB (135).The last two were case-control studies with a low RoB (13,36) (Supplementary Table 14).

Outcomes of included studies
The experimental group in all studies with force application showed an increased incidence of resorption and/or severity of OIIRR.The reported OIIRR was the resorption crater volume (129,133), the ratio of reduction in root length (135), volume of cervical RR (130), and the percentage of RR severity between the groups (131), as compared to the control group with no-force application.
Two studies reported relevant secondary outcomes (131,132): no significant association was found between OIIRR and patient hygiene level or periodontal parameters (PI, GI, BOP, PPD) The summary of the included studies is listed in Supplementary
Analysis for reporting bias across the studies revealed that nine reports were on the same cohort of the population (137,26,28,27,34,35,33,24,24).This was confirmed by emails from some authors of these publications (Linhartova p. and Iglesias-Linares A.).Therefore, each was evaluated as a different report of the same study, including the measurable outcomes.
High quality of evidence according to applied GRADE methodology indicates that orthodontic forces in PICO 3A and/or intervention or exposure in PI/ECO 3B are associated with the incidence or increased severity of OIIRR in human studies (Supplementary Table 19).

Summary of evidence
This work aimed to systematically review the existing literature on the effect of mechanical force (± patient-related risk factors) on the incidence of root resorption from in vivo studies.Further, we aimed to elucidate the underlying pathophysiological process of OIIRR by assessing cellular '-omics' changes in vitro studies.
The systematic review of current literature demonstrated that OIIRR occurs under mechanical stress in humans and animals.Further, it can be concluded that different forms of pressure application are associated with differing severities of OIIRR.

PICO1
All in vitro studies reported a stimulating effect of mechanical force on human PDL cells in an artificial environment.However, these investigations primarily focused on the interaction at the cellular level whilst not considering the interaction with the environment.These effects are expressed either as increased or decreased expression of genes and pathways, which differed between studies due to the magnitude of applied force, application method, extraction devices, and sequencing methodology.The results of the expression profiles were reposted using fold change compared to controls.Only the significant expression data were collected.

mRNAs expression profiles
It was found that the CF elevated interleukin 6(IL-6), interleukin-1b, ADAMTS1, cyclooxygenase (COX)-2, and osteocalcin (76,78,6).Both constant and intermittent CF increases SOST, TGFB1, and HEY1 mRNA expression.In contrast, the HES1 mRNA and TGF-β signalling, which affect osteogenic differentiation, are increased in intermittent CF but not in constant CF.Other experiments found a reduction in alkaline phosphatase (ALP), -a marker for new bone formation under different forces.Reduction in the Expression of FOXM1 under CF enhanced the osteoclastic differentiation and the receptor activator nuclear factor-kB ligand (RANKL)/OPG ratio.An increase in RANKL was also found after COX-2 activation and prostaglandin E2 (PGE2) induction.OPG mRNA was found to be upregulated by cyclic tension, which has a negative effect on osteoclastogenesis and transferring this gene to the periodontal area results in the reduction of OTM (75,76,78,79,80,6,82,83,86,88).These findings might shed light on the role of PDL cells in the pathophysiological mechanisms of OIIRR.

lncRNAs expression profile
Crucial evidence was discovered regarding the lncRNAs' involvement in cell differentiation and development and the osteogenesis pathways and force effect on their expression levels (3,89,90,151).The lncRNA has been reported to act as sponges for microRNAs (miRNAs) and may interact with miRNAs to control their downstream destinations (153).Under pressure, the lncRNAs promote osteogenesis through the focal adhesion kinase pathway (FAK), the mitogen-activated protein kinase pathway, the TGF-β1/ Smad3/HDAC signalling pathway and the downregulation of ANCR expression.It can also regulate the mRNAs' expression by enhancing the binding ability of miRNAs.Wang et al. novelty reported on the network of the lncRNAs-miRNAs-mRNAs relationship of the PDL stem cells under tension (3,7).These studies provide insight into the role of the lncRNA and its association with the mRNAs and miRNAs expression in osteogenic differentiation and bone and root resorption.

PICO 2 and 3
Evidence in both human and animal studies Many risk factors were identified in the current review and can be divided into systemic and local factors.As stated earlier, root resorption has an inflammatory origin; increasing the impact of inflammation leads to more severe OIIRR.This may explain the increase in OIIRR in many studies with low to moderate risks of bias in human studies, including roots with periapical pathology (142), hypofunctional teeth and periodontium in human (49) and animal studies (126), and previous trauma of teeth and the surrounding periodontium in human reports (13,23).
The overall certainty of the evidence using the GRADE tool for each PI/ECO was determined to be highly variable, ranging from low to very low, with few studies being graded at moderate and high certainty (Supplementary Tables 13 and  19).
Immature roots have less OIIRR volume than mature teeth, which may be due to the decreased expression of Jagged1, Notch2, IL-6, and RANKL signalling in both low-risk human studies but higher-risk animal one (2,11).
Unspecific allergic reactions were amongst the prominent systemic risk factors of OIIRR, with human studies of moderate evidence (19,23,146), insufficient evidences (50), and unclear evidences of an animal study (92).
Furthermore, systematic fluoride is a protective factor against RR in two high-quality RCT and CCT human studies (45,150) and two animal studies (97,120).Only one animal report stated that this effect is insignificant (117).

PI/ECO 2 animal studies
There is moderate evidence for a reduction in the volume of OIIRR of teeth adjacent to an extraction site of rats grafted with bone-ceramics, which might be explained by a better osteoinductive potential than naturally recovered sites or natural bovine cancellous bone particles graft (123).
There is moderate evidence for a reduction in the volume of OIIRR of teeth adjacent to an extraction site of rats grafted with bone-ceramics, which might be explained by a better osteoinductive potential than naturally recovered sites or natural bovine cancellous bone particles graft (123).
Medication intakes pose an additional risk factor in the development of OIIRR in animal studies, including ones that increase the severity of OIIRR with strong evidence like strontium ranelate intake (119) or with moderate to low evidence like Corticosteroids (109) and 4-hexylresorcinol medication (118).
Oral contraceptive intake, on the other hand, decreases the severity of OIIRR due to the effect on the hormonal level that affects periodontal tissue and cementum metabolism (102,127) suggests that assessment of individual bone metabolism pre-orthodontic treatment may give insight into the rate and severity of OIIRR (108).
Similar associations were shown in rats with decreased steroid-sex hormone plasma levels due to bone turnover reduction (124).
Another study with moderate evidence found that increasing caffeine intake might also reduce OIIRR occurrence, which might be due to the inhibition of the formation of osteoclasts that leads to decreased bone turnover and resorption (125).

PI/ECO 3 human studies
Occlusal abnormalities such as an increased overjet and Angle's class II malocclusion are of substantial evidences to be associated with more severe OIIRR; this may be caused by increased treatment times or the amount of applied force (23,48).
In addition, increased incisal inclination and the presence of an obstacle obstructing the path of root movement (e.g.impacted teeth, increase in cortical thickness) may lead to more stress on the root and subsequently to an increase in odontoclastic effects, with high-quality human studies (138,54,24,15).
Biological sex seems to have a controversial effect on the OIIRR among studies; Four studies with strong evidence reported increased OIIRR in male compared to female participants (32,23,28,27,148).In comparison, others with medium evidence do not find an effect of a sex difference (29,142,14).
Higher age was associated with increased OIIRR in human observational studies, 4 of which are of high quality (25,36,140,41,23,15), whilst three reported no association with low to medium risks (18,142,14).Age may increase OIIRR due to the increased risk of iatrogenic damage by orthodontic forces (154).Consequently, special consideration and measures should be contemplated to manage root resorption risk in adults, such as increasing the gap between visits, less force, and maintaining oral hygiene during treatment (155,156).High-quality RCT studies are needed to solve the controversy.
Data from low risks observational studies conclude that the association between asthma and OIIRR is inconclusive; one shows a link to increase OIIRR risk (40), and the other shows no link (41).
The initial size of the root, initial root length, shape, and width are significant factors in some observational variable qualities studies (25,36,145,23,15).In contrast, other lowevidence retrospective studies reported no effect (50,147).These conflicting outcomes may be linked to the different methodologies, treatment time, and measurement techniques.
Genetic considerations may be an important perspective for the incidence and severity of OIIRR.These genetic variations might be associated with increasing or decreasing the incidence of OIIRR.It is important to note that all of the included genetic studies were only observational.
Conversely, according to other low risks studies, some genes may have a protective effect, such as variations in the OPN gene and the P2RX7 gene, osteopontin (OPN) deficient subjects, polymorphisms of osteoprotegerin and IRAK1 (31,33,28,27).
Many articles discussing the genetic effect, both included and not included in our work, supported the belief of a possible relationship between genetic variations and OIIRR, which explains the difference in the amount of root resorption between individuals with other factors held constant.

Limitations
Multiple aspects of this systematic review accentuate the limitations of the currently available literature dealing with the question posed.
Results of in vitro studies are generally questionable regarding their generalizability and applicability to the natural environment.For example, in some studies with the force application model device, direct contact between the glass and cells may generate an electrical charge and gas exchange.In addition, using different devices and software may also lead to variability in outcomes.Furthermore, using a two-dimensional or three-dimensional medium is questionable; some studies used both, showing different cellular and molecular responses (77).In addition, different methodologies, homemade devices, and software not verified by a standardized technology cause high variability in outcomes.
Most animal studies were unclear regarding the randomization and blinding of the assessors.Another big issue for the in vivo studies is the selection of radiographic techniques for reporting the OIIRR.As tooth resorption is a three-dimensional process, it may affect not only the length of the root, which is assigned as apical resorption (which is the reported outcome of most 2D radiographic studies) but may also exhibit itself as the reduction in thickness, cervical resorption, creation of coves, lacunae, and mini cracks.These increase with biomechanics application and cannot be detected by 2D radiographs (periapical (PA), panoramic, or cephalometric).Another drawback of the none standardized radiograph includes different magnification levels, angulation and superimposition of other structures, limiting the usefulness of this data (157), and PA results may be overestimated (158).In contrast, the 3D imaging (CBCT) and scanning electron microscope studies provide more relevant data on root resorption (159).Since these methods are becoming easier to access nowadays, this systematic review suggests that future in vivo observations should consider utilizing them.
Also, histological studies using a light microscope offer less information than scanning electron microscopes.Nevertheless, the main limitation of both methodologies is the fact that they can only be utilized with extracted teeth.Additionally, some studies report results by grouping outcomes, leading to generalizing results and data loss.
For human studies, it is well known that randomized clinical trials deliver the most evidence-based results: Due to our inclusion and exclusion criteria, we included only five randomized-controlled trials for this systematic review (129,130,131,45,23).Three additional papers did not comment on randomization methodology in their text and were considered nonrandomised clinical interventions (135,143,47).All other studies were either extended cohort or case-control studies; some exhibited a significant delay between conducting experiments and publication of results.And some used a four-point grading system (Malmgren Levander Index) or other indecies to classify the severity of OIIRR, without an exact numerical measurement, which decreased the ability for comparison.Nevertheless, more RCTs and observational studies in a controlled environment are needed to support this hypothesis.
Although these are exciting results of the genetic studies, the findings remain controversial.Newer evidence suggests that not only one SNP within a gene can act as an absolute determinant for the disease or increase the severity of the OIIRR, but other genetic factors such as (gene-gene interactions) and environmental and lifestyle risk factors (gene-lifestyle, geneenvironment, and lifestyle interactions-environmental) may play a role (160).
Only a tiny portion of genes could be identified in genetic studies, and genetic variance may have structural variations, leading to difficulties in discovering the missing SNP risk factor unless the main contributing factors are fully understood.According to Schäfer et al., genetic-association studies have four main problems First, careful case selection is essential to minimise phenotypic heterogeneity.Also, sufficient participants should contribute to the study's statistical power.Further, replication, which can be considered the gold standard, and the complete capturing of the genetic information is essential.If all these factors were applied, this should result in a significant finding, but unfortunately, many OIIRR genetic studies miss one or more (160).The small and underpowered SNPs' studies available resulted in controversial and questionable findings.
Variability in methodology, sample sizes, and baseline characteristics may contribute to the observed high heterogeneity of included studies.Missing or unclear information around the randomization process, sample size calculation, blinding of the patient and/or assessors, and methodology on outcome measurements impact the risk of bias assessments and reduce the reproducibility of included studies.
We have generalized our research question to include mechanical forces.Therefore, we have not compared different treatment mechanics or interventions, even though the choice of the appliance with different force direction and strength may have a relevant impact on OIIRR incidence and severity; this can be considered a limitation of our work.Last, we refrained from performing a meta-analysis due to the high heterogeneity of the studies included regarding methodology, outcomes measures, and reporting strategies.

Conclusions
1. Available evidence uniformly demonstrates a linkage between varying mechanical orthodontic forces and the occurrence of OIIRR.Overall, this work highlights the need for genetic studies with an '-omics# basis for PDL cells under pressure.Further, this systematic review has also highlighted the shortcomings of currently available literature.Eligible studies were of high heterogeneity in methodology and often had a different risk of bias and small sample sizes, and therefore need to be interpreted with caution.More profound research is needed to understand the pathophysiological mechanisms underlying OIIRR, especially concerning the host response.

Table 15 .
Detailed studies' descriptions and outcomes are shown in Supplementary

Table 16 .
PE/ICO 3B Is the incidence of root resorption [O] in a population of patients undergoing orthodontic treatment [P] higher in patients/situations with potential systemic or local risk factors [E] compared to patients/conditions without such risk factors [C] after a minimum follow-up of 4 weeks [T]?
2. Risks for increasing OIIRR should be viewed with caution, including occlusion, previous trauma, tooth shape, allergy, low bone turnover, and medication like corticosteroids, strontium ranelate, 4-hexylresorcinol and genetic polymorphism of IL-1B, IL-1RN, RANK, osteoprotegerin, and vitamin D receptor TaqI polymorphisms.Simultaneously, other factors seem to decrease the rate of OIIRR, such as oral contraceptives, baicalin, high caffeine, root-filled teeth, and polymorphism of osteoprotegerin and IRAK1.3. Evidence from in vitro studies on different force types and mechanical regimes on HPDL cells, which affect biological behaviour and genetic expression, provided a possible explanation for their extracellular matrix gene expression and adhesion, osteotropic cytokines, growth factors, controlling PDLC osteoblastic/cementoblasts differentiation and proliferation capacity, and autophagy which provides a novel mechanism in the regulation of the clinical OTM process and root resorption.