Risk factors for orthodontic mini-implants in skeletal anchorage biological stability: a systematic literature review and meta-analysis

The reason of the biological stability loss of mini-implants is still a matter of discussion between dentistry professionals. The main objective of this systematic literature review and meta-analysis was to analyze the risk factors that prejudice this loss. A search was made in the electronic databases Pubmed, Scopus, Embase and Cochrane, in addition a manual search was made too in Grey Literature (Opengrey). No limits were set on the year of publication or language. The inclusion criteria were: studies in humans treated with fixed appliances with mini-implants, where the risk factors for secondary stability were evaluated for a minimum of 8 weeks. After eliminating duplicate studies and assessing which ones achieve the inclusion criteria, a total of 26 studies were selected for the qualitative synthesis, 18 of them were included in the quantitative synthesis. Common risk variables were compared in all of them. Analyzing the forest and funnel plots, statistically significant differences were obtained only for location, the upper maxilla having lower risk than the mandible with an odds ratio of 0.56 and confidence interval of 0.39 to 0.80. Prospective studies under controlled conditions should be required in order to obtain a correct assessment of the variables analyzed.

eligibility criteria. "Articles" and "Articles in press" were included: randomized clinical trials (RCT), longitudinal studies, cohort or case/control studies both retrospective and prospective. No restrictions were applied regarding the year of publication or language. Inclusion criteria were: human studies that evaluated risk factors for secondary stability of mini-implants used in orthodontic treatment for a period of at least 8 weeks.
outcome. Secondary, or biological stability, begins at the moment of placement and increases during the bone remodeling or healing process. Biological stability is lost when mobility, severe inflammation or infection appear.
information sources, search strategy, and study selection. To identify relevant articles, an electronic search was conducted in four databases: Pubmed, Embase, Scopus and Cochrane. A manual search was conducted in Opengrey literature. In some cases, the authors were contacted by e-mail to request additional information. The key search terms used to identify articles were: (Humans/human OR patients OR adults OR male OR female) AND (mini-screws OR micro-screws OR mini-implants OR temporary anchorage devices OR TAD* OR stability OR long-term stability OR skeletal-anchorage OR orthodontic treatment) AND (orthodontic failure-rates OR orthodontic success-rates). Two reviewers (MDC-R and CB-A), assessed the titles and abstracts of the articles identified in the electronic search; whenever disagreement occurred a third reviewer was consulted (JMM-C). If the abstract did not provide sufficient information to reach a decision to include it in analysis, the reviewers read the full text. Then, the full texts of the articles selected were read, and if they failed to meet inclusion criteria, the reasons for rejection were recorded. (Table 1).
Data items and collection. The review and meta-analysis was updated for the last time on January 4 th 2020.
The following variables were extracted from the works selected: author; year of publication; study type; sample size; demographic variables (patient age and sex); type of mini-implant (length and diameter); mini-implant location (maxillary or mandibular); position (left or right, vestibular, lingual/palatine, or crestal); and the number of mini-implants per patient. (Table 2) Risk of bias/quality assessment in individual studies. The quality of the studies was assessed by the same reviewers independently, using the Newcastle-Ottawa scale 10 . In case of any discrepancies in quality assessment, consensus was reached between the reviewers and if this was not possible a third reviewer was consulted. (Table 3) Summary measures and approach to synthesis. Means and confidence intervals were calculated for patient age, mini-implant length and diameter. Adverse events (loss of mini-implant stability) were recorded in relation to possible risk factors: mandibular or maxillary placement; vestibular, palatine/lingual, or crestal placement; placement on the left or right sides, in anterior or posterior regions; patient age and sex; mini-implant length and diameter.
Statistical analysis. For the quantitative synthesis odds ratios and 95% confidence intervals of risk factors were calculated, together with their confidence intervals. Heterogeneity was assessed by the Q test and the I2 statistic. A Q test p-value <0.1 was considered heterogeneous. The random effects method was used to calculate the odds ratios. Publication bias was analyzed using funnel plots and Egger's regression intercept method. The software employed was comprehensive meta-analysis V 3.0 Biostat.

Results
Study selection and characteristics. The electronic search in four databases identified 1182 articles: 226 in Pubmed, 547 in Scopus, 209 in Embase, and 200 in Cochrane; the grey literature manual search located four further works. Duplicates were eliminated leaving a total of 791 studies for analysis. After reading the titles and abstracts of the selected papers, 731 articles were discarded leaving 60 studies for thorough assessment. Full texts of these 60 studies were revised and 34 articles were eliminated as these were literature reviews or meta-analyses or because they failed to answer the research question or did not meet inclusion criteria ( Table 1) Results of individual studies, meta-analysis and additional analyses. Patient sample sizes in the articles selected ranged from a minimum of 34 11 to a maximum of 570 12 , although most had sample sizes averaging around 150 patients. As for distribution by sex, there were more women than men. For example, one study 13 included 295 women and 64 men, while another study 14 included 39 women and 5 men. In most of the studies mean patient age was around 25 years, samples generally consisting of predominantly young adults. Of the 26 articles, 7 were prospective studies 15-21 and 19 retrospective [11][12][13][14][22][23][24][25][26][27][28][29][30][31][32][33][34][35][36] . Follow-up periods ranged from approximately 18 months to 2 years. The shortest follow-up period was 4 months 30 and the longest 10 years 26 .
Analyzing quality with the Newcastle-Ottawa scale, scores of 6/9 or 7/9 were obtained, which indicated moderate or moderate-high quality respectively. ( Table 2) Qualitative synthesis. With regard to the number of mini-implants, high variability was found, one of the study using a total of 1,356 mini-implants 12 (study with the greatest number of mini-implants within this review) contrarily to another study that used 79 mini-implants, being the study with the lowest number of mini-implants 28 . Most works used an average of around 350 mini-implants. The analysis of mini-implant location found more common placement in the maxilla than in the mandible, excepting 3 studies 21,25,26 . Length and diameter were also evaluated, finding that the most widely used mini-implants were of 8 mm in length and 1.4 mm diameter. Mini-implants of different diameters were used ranging from 1.0 mm 27 to a maximum of 2.3 mm 27 . As for length, the minimum length used was 5 mm 12 and the maximum 21 mm 13 . Among studies that provided information about the side on which mini-implants were placed (left or right), only one study placed the same number of mini-implants on both sides 18 ; in the other studies, the right side received higher number of www.nature.com/scientificreports www.nature.com/scientificreports/ mini-implants 17,22,31,34,35 . When mini-implant positions (vestibular, lingual/palatine, or crestal) were compared, the most usual position for mini-implant placement was the vestibular area 33 , with the exception of one article where a higher number of mini-implants were placed in lingual areas 11 . Quantitative synthesis. The failure rate calculated from the total of mini-implants placed was analyzed for the risk factors studied. The highest values were observed for ages under 30 years with 19.7%, the mandibular location 18.8% and length higher than 8 mm 18.6% (Table 4).

RECORDS EXCLUDED ARTICLES EXCLUDED, WITH REASONS
Mini-implant location in the maxilla or mandible showed a significant relation with secondary stability; the odds ratio obtained in the forest plot of this meta-analysis was 0.56 (95% confidence interval (CI) between 0.39 and 0.80), which indicates that mandibular mini-implant placement is a protective factor. Moreover, the I2 value (I2 = 78%) indicated heterogeneity (Q = 63.72; p < 0.001). Analysis of the age variable found that ages of 30 or lower were also a protective factor for mini-implant secondary stability although this did not reach statistical significance, with an odds ratio of 1.59 and 95% CI of 1.01 to 2.50. The I2 value in this case was 63.67%, indicating moderate heterogeneity (Q = 22.02; p = 0.005).
Gender did not show a statistically significant relation with secondary stability, obtaining an odds ratio of 1.

Discussion
Summary of evidence. The null hypothesis of the present study was rejected. When analyzing failure rates, significant differences were found when comparing different mini-implant locations. The maxillary location showed significantly lower failure rates than the mandibular, while the rest of the analyzed variables did not show significant effects.
Bone anchorage in orthodontic treatments is becoming ever more widespread. As the literature states repeatedly, the advantages offered by mini-implants are numerous, and the technique shows a success rate of over 80% 37 . But when mini-implants fail they do so during the first 8 weeks after placement, the period when implant stability changes from mechanical to biological stability 6 . The reasons for failure (mobility, displacement or infection of the surrounding soft tissues) were not assessed in the present literature review, as its objective was to determine which variables predispose mini-implants to failure.
Longitudinal studies, cohort and case/control studies both prospective and retrospective were included in the present review. Although RCTs were eligible to be included, none met the inclusion criteria. Cohort, case/ control and RCT studies are the ones that better evaluate the risk factors associated with mini-implants stability. Including different study designs could lead to broad variations in outcome and quality terms. However, when analysing the quality of all of them we found that most of them present values between 6 and 7 in the Newcastle-Ottawa scale indicating that they have moderate quality ( Table 3).
Since weaker studies can influence the outcome of the present study, the sensitivity of the estimation of the 19 conducted meta-analyzes was analyzed and none of the studies was found to significantly influence the results, except one (Figs. 2 and 3).
The variables analyzed can be divided into three groups: variables deriving from the patient, others related to the mini-implant used (length and diameter), and a third set related to location and technique. Among the variables deriving from the patient, the present study identified some controversy surrounding the possible association www.nature.com/scientificreports www.nature.com/scientificreports/ between mini-implant failure and age. Some studies found no association between age and failure 14,15,32 , while others found that the patient's age could influence mini-implant failure because younger patients present finer cortical bone and lower bone density 19,[30][31][32][33][34] . (Fig. 2) The variable sex was also subject to controversy. While some studies [12][13][14]30,31 claim that the patients' gender does not influence the success or failure of treatment with mini-implants, other works 33,35 found a higher mini-implant success rate among men, which they attributed due to their higher bone density. However, the present review did not generally find a significant association between sex and loss of secondary stability.
Several other factors may influence failure/success rates but the present review was unable to analyze them due to the lack of published articles. These include poor hygiene, inflammation of surround tissues and the individual patient's bone density 36 , all thought to play a role in treatment failure 13 .
When mini-implant characteristics were analyzed, those articles that analyzed different implant sizes 16,27,28 , agreed that diameters smaller than 1 mm show a tendency to failure. These findings make sense since smaller diameters provide less attachment surface and also, they are weaker so prone to fracture. But other studies 14,31 found that length and diameter were not variables with any significant influence on success/failure rates. Meanwhile, two studies 11,24 found that length was a factor that influences stability. The present meta-analysis did not show that mini-implants longer than 8 mm or with diameters over 1.4 mm undergo fewer failures than those with smaller dimensions.
Regarding mini-implant location, two studies 14,16 stated that mini-implants presented a worse prognosis in posterior regions and in the alveolar mucosa. Between vestibular, lingual/palatine and crestal placement, two studies 16,26 agreed that there is a higher risk of loss of mini-implant stability in lingual/palatine areas. But another study 24 found that placement in crestal areas shows greater risk of failure. Meanwhile, two studies 31,34 affirmed that the right side presents a higher risk of failure than the left. This could be due to patient's oral hygiene maintenance capability or the dexterity of the clinician. The present review did not find significant relations between mini-implant failure and vestibular, lingual, or crestal placement or between placement on the left or right side. Only two articles 14,26 differentiate jaw and maxilla together with the anterior and posterior position obtaining higher success rates in the anterior maxillary area (including the palate), with no statistically significant differences.
However, in agreement with some studies 33,35 placement of mini-implants in the mandible could be a protective factor due to the mandible's higher bone density. (Fig. 3) Among the technical characteristics of mini-implant placement, there is also some controversy surrounding the intensity and duration of the force applied. One study 16 claimed that the correct force is between 100 and 200 grams, and concluded that lower forces produced higher success rates, as did applying force for the shortest possible time. As for surgical technique, one work 28 concluded that mini-implant placement without flap raising produces less damage and so higher success rates. Mini-implant placement using a motor also favored   Table 4. Risk factors and event rate. a successful outcome, moreover the surgical protocol with or without predrilling has only been assessed in a study that showed no significant differences. One study 33 obtained a success rate of 88.38% for no-predrilling mini-implant insertion and 89.09% for predrilling, with an odds ratio of 0.93 (95% CI 0.36 to 2.42). For some authors, surgical technique is one of the most influential variables affecting stability 18 .
Knowledge of the factors involved in the failure of mini-implants can help the clinicians to improve their clinical practice. Within the analyzed factors there are some that could be adapted for a higher success. Only location seems to have a significant effect, having the upper jaw lower failure rate than the lower. However, this cannot be clinically modified since it depends on the biomechanics required in each case. The other analyzed factors did not show differences and their failure rates are very similar. (Table 4) Limitations. The limited number of studies that have investigated the variables analyzed in the present review could lead to estimation bias or even failure to identify their significance. Further prospective studies under