Predicting success rate of orthodontic traction with gold chain bonding: Evaluation of factors and a novel prediction model

Study design

This retrospective cohort study analyzed orthodontic patients undergoing GCB surgery at the National Dental Centre Singapore (NDCS) between 1^st January 2014 and 31^st December 2021. Inclusion criteria required at least one impacted tooth treated with GCB and fixed orthodontic appliance. The GCB surgery was performed by oral and maxillofacial surgeons after referral from the attending orthodontists. Surgical exposure of the impacted tooth was performed, along with the bonding of a gold chain for attachment during subsequent orthodontic traction. Traction was applied together with fixed appliances. Exclusion criteria included unclear documentation, congenital craniofacial syndromes, or loss to follow-up.

Data collection

Treatment notes and the relevant radiographs (orthopantomograms [OPG], anterior maxillary occlusal, and periapical films) from the center’s digital records were evaluated. The recorded patient and radiographic predictors are outlined in [Table 1]. [Figures 1-5] illustrate the categorization of each radiographic predictor, and [Tables 2-5] describe the categorization of each radiographic predictor.

Close

Figure 1:

Horizontal crown position (sectors).

Export to PPT

Close

Figure 2:

Horizontal root apex position (sectors).

Export to PPT

Close

Figure 3:

Vertical height (d = distance from occlusal plane).

Export to PPT

Close

Figure 4:

Vertical height (grades).

Export to PPT

Close

Figure 5:

Angulation of long axis to midline (α).

Export to PPT

Definitions of treatment outcome measures

1. Treatment success: Eruption of the impacted tooth into its functional position in the dental arch, documented by an orthodontist or supported by radiographic evidence.

2. Treatment failure: Surgical excision or termination of traction.

3. Re-intervention: Additional surgery required to facilitate eventual eruption, including soft tissue exposure with or without additional removal of bone, gold chain repositioning, or rebonding.

4. Duration to eruption: Number of days from the GCB surgery to successful eruption of the tooth as documented by the orthodontist, or sufficient eruption such that a bracket or a button could be bonded.

Two clinicians evaluated the radiographs (C.T.N. and Q.L.), and measurements for 30 randomly selected teeth were repeated. Intra-examiner and inter-examiner reliability were assessed, with intra and interclass correlation coefficient (ICC) estimates calculated using a two-way mixed-effects model.

Following the collection of patient and radiographic variables [Table 1] and treatment outcome measures, the GCB traction complexity score was calculated for each tooth, using the GCB traction complexity matrix [Table 6]. Scores of 0, 1, or 2 were assigned for each of the six complexity factors (simple, moderate, and complex, respectively), with the sum representing the GCB traction complexity score for each tooth. The complexity factors were chosen based on existing literature on impacted maxillary canines to evaluate whether these factors can be generalized to other tooth types.

Statistical analysis

Success rates were calculated as the proportion of teeth erupting into functional occlusion. Baseline demographics and clinical features were compared between the two groups (success vs. failure) using a two-sample t-test or Mann-Whitney U-test (depending on normality assumption) and Chi-square or Fisher’s exact test (where appropriate) for continuous and categorical variables. The association of baseline characteristics and success rate was investigated using univariate logistic regression analysis, with results expressed as odds ratio (OR) and 95% confidence interval (CI). Kaplan–Meier analysis was performed to calculate the median survival time of duration to eruption and to compare various subgroups using the log-rank test. Statistical analysis system (SAS) software (V 9.4) was used for analyses, with significance set at P <0.05.

Patient and tooth characteristics

A total of 530 patients with 616 impacted teeth underwent GCB between 1^st January 2014 and 31^st December 2021. After exclusions, the cohort included 421 patients (mean age: 13.6 years, 53% female). Among the 495 analyzed teeth, 462 erupted successfully (93.3% success rate). The patient and radiographic variables in relation to success rate are presented in [Table 7]. Crown position, apex position, vertical height, and angulation significantly influenced success [Table 8].

Intraclass correlation coefficients for both observers were 0.99, indicating a high level of agreement within each examiner. ICC was 1, indicating a perfect level of agreement between examiners.

Predictive factors of success

The patient and radiographic variables, along with their respective ORs, are presented in [Table 8]. Patient age was found to be significant, with every 1-year increase in age reducing the odds of success (P = 0.0099) [Table 8].

Crowns positioned more mesially (in sectors 3, 4, or 5) had significantly lower odds of success compared to those in sectors 1 or 2 (P = 0.0003). Similar trends were observed when crowns in sectors 3 or 4 were compared to those in sectors 1 or 2 (P = 0.0004), and when crowns in sector 5 were compared to those in sectors 1 or 2 (P = 0.0381), with consistently lower odds of success in higher sectors [Table 8].

Teeth with root apices more horizontally displaced (sector 3) had significantly lower odds of success compared to those in their intended apical position (sector 1) (P = 0.0406), but the difference between sectors 2 and 1 was not significant (P = 0.3334) [Table 8].

Cases classified as grades 3 or 4 vertical height exhibited significantly lower odds of success than grades 1 or 2 (P = 0.0147). A higher vertical grade indicates a greater distance from the occlusal plane. Greater statistical significance was observed when comparing grade 4 alone against grades 1 or 2 (P = 0.0003), whereas grade 3 showed no significant difference when compared to grades 1 and 2 (P = 0.1219). Analyzing teeth more than 12 mm versus less than 12 mm from the occlusal plane revealed lower odds of success in the former group (P = 0.0047). Incremental increases in vertical height also affected the success rate, with every 1 mm (P = 0.0020), 5 mm (P = 0.0020), and 10 mm increase (P = 0.0020) reducing the odds of success [Table 8].

Angulation significantly affected the success rate, with larger angulations leading to lower odds of success (P = 0.0029). Significant differences were also observed when comparing the 0–15° group with those >30° (P = 0.0125) but not for the 0–15° group with the 16–30° group (P = 0.2141) [Table 8].

Apex formation and root form did not significantly impact success rates (P = 0.1060 and P = 0.1026, respectively).

Duration to eruption

The median duration to eruption for the successful cases was 274 days (9.0 months), ranging from 25 to 1596 days (0.8–52.5 months). [Table 9] represents the patient and radiographic variables in relation to duration.

Teeth >12 mm from the occlusal plane required 138 days (4.5 months) more to erupt compared to those <12 mm away, and this difference was statistically significant (P <0.0001) [Table 9].

Cases with less vertical displacement (grades 1 and 2) required significantly shorter duration compared to those displaced further vertically (grade 3 or 4) (P < 0.0001), with grades 1 or 2 taking 141 days (4.6 months) fewer to erupt. Grades 1 or 2 also exhibited significantly shorter duration compared to grade 3 alone (P < 0.0001), and when compared to grade 4 alone (P = 0.039), but no significant difference was found between grades 3 and 4 (P = 0.884) [Table 9].

The GCB traction complexity score significantly affected the duration. Teeth with a complexity score of 0 required significantly less time to erupt compared to scores 1–3 (P = 0.0112) and scores 4–8 (P = 0.0006), requiring 54 days (1.8 months) and 78 days (2.6 months) fewer, respectively [Table 9].

Other patient and radiographic factors did not show significant differences in terms of duration to eruption.

Complexity matrix validation

The GCB traction complexity scores of this sample ranged from 0 to 8, with the largest proportion of teeth having a score of 4 (18.2%) [Figure 6]. Higher complexity scores were associated with lower success (P < 0.0001) [Table 8] and longer treatment durations (median: 274 days for successful cases). The matrix’s area under the receiver operating characteristic curve was 73.7%, 95% CI (64.2%, 83.2%), reflecting moderate predictive accuracy.

Close

Figure 6:

Distribution of gold chain bonding traction complexity scores. (x axis: GCB traction complexity score)

Export to PPT

Additional subgroup analysis was performed according to tooth types, with specific analysis on maxillary canines and maxillary central incisors, which make up the majority of our sample at 39.0% and 31.3%, respectively. It showed that the complexity score was a significant predictor of success for maxillary canines (P = 0.0024) but not for maxillary central incisors (P = 0.1211) [Table 8].

Horizontal crown position

Concerning horizontal position, this study revealed significantly higher success rates for crowns in lower sectors (less mesial displacement). Existing literature primarily focused on interceptive treatment, with few analyzing its effect on the success rate of orthodontic traction. Notably, Grisar et al. and Chaushu et al., who categorized horizontal position similarly, found no difference in treatment prognosis.^[24,25] The disparity may stem from our larger sample size, a higher proportion of cases in sector 1, and the decision to group sectors for comparison. However, this study’s findings aligned with Motamedi et al., who demonstrated reduced success for impacted maxillary canines with overlap of more than half the adjacent lateral incisor root.^[26] These results emphasize the prognostic importance of precise radiographic assessment for mesial crown displacement in treatment planning.

Vertical crown height

Vertical height was found to be a significant factor affecting success, consistent with Grisar et al.’s study employing a similar grading system.^[24] Koutzoglou and Kostaki also reported that canines buried at a level deeper than half the root length of the adjacent incisor were more likely to be ankylosed,^[27] thereby increasing the odds of failure if traction was attempted. In contrast, Chaushu et al. found no significant association, possibly due to the exclusion of extreme cases, representing the upper and lower ends of prognosis.^[25]

Crown angulation

Regarding angulation, our results mirrored Motamedi et al.’s findings, associating larger angulations, particularly above 45°, with lower success.^[26] This highlights the need for early intervention in cases with severe angulation to mitigate risks and improve outcomes.

Root development and morphology

Root development and dilaceration showed no significant effect on success, aligning with findings by Motamedi et al.^[26]

GCB traction complexity matrix

The introduced GCB traction complexity score, amalgamating six predictors, demonstrated significant correlation with success rate. This validates it as one of the first prognostic indices supported by clinical data on treatment outcome measures. Most variables within the complexity matrix (age, crown position, root apex position, and vertical height) were identified as good predictors of success. When assessed according to tooth types, the complexity score proved to be a significant predictor of success for maxillary canines but not for maxillary central incisors. This could be because most of the complexity factors were based on existing literature that is predominantly focused on maxillary canines, and the common presentation of impacted upper central incisors can vary from that of canines, usually associated with more severe dilaceration, greater vertical displacement of the crown, and oftentimes inverted. These radiographic findings are not easily evaluated on OPGs, given the lack of buccolingual view, which likely further explains this discrepancy.

Treatment duration

Examining traction duration, two significant factors emerged – vertical height and the GCB traction complexity score. These results align with earlier studies indicating that an impacted tooth positioned further from the occlusal plane requires increased duration.^{[16,24,25,28,29]} However, Arriola-Guillen et al. reported no impact of initial height on duration,^[30] which could be attributed to different traction modalities and a smaller sample size.

The GCB traction complexity score significantly influenced duration, establishing it as a valuable tool for predicting both success rate and treatment duration. Utilizing this matrix in treatment planning enhances the patient selection process. Thus far, most existing indices available have not been evaluated for their correlation with treatment duration,^{[15,18,20,22]} except for the KPG index,^[19] which was shown to correlate well with the 2D measurements from Stewart et al. that were observed to affect duration.^[31] As such, very few of them have been proven to serve as reliable indicators of duration, which is a critical deciding factor during treatment planning.

Despite the remaining factors not reaching statistical significance in this study, existing literature suggests that patient age,^{[16,28,29,32]} angulation,^[28,24,33] horizontal crown position,^{[24,28,30,33]} and root dilaceration^{[25,29,32,34]} can impact duration. These discrepancies may arise due to non-standardized definitions and measurement methods. Authors reported duration as the surgery to the end of fixed appliance,^[28,24] the start of traction to good alignment^[29] or to emergence of the cusp tip,^[33] and the start to end of fixed appliance.^[16] Furthermore, duration was mostly measured as the number of months,^{[16,25,29,33]} but one study also used the number of visits.^[28]

Study limitations

Study limitations include its retrospective nature, potentially introducing errors and selection bias. Alternative traction methods such as cantilever springs,^[35] ligature wires,^[36] nickel–titanium coil springs,^[30] or open surgical approach^[37] were not considered, potentially affecting generalizability. OPGs were used for 2D measurements, which limited accuracy but can serve as a low-dose, multi-purpose radiograph that is commonly taken regardless of the treatment planned, compared to CBCTs. Although 2D imaging was used, its widespread availability ensures that the findings remain applicable to routine clinical settings. Treatment was performed by multiple orthodontists and surgeons; thus, the varied levels of skill and experience could have affected the success rate. This cannot be accounted for by the matrix but improves generalizability. The GCB traction complexity matrix’s inherent lack of specificity to a particular tooth type may reduce accuracy, but expands its utility as a generic risk assessment tool with wider applications.

Future research could explore 3D measurements using CBCTs for improved accuracy. The GCB traction complexity matrix could also be evaluated for its correlation with other treatment outcomes, such as function and esthetics after traction, periodontal health, or the prevalence of negative sequelae like root resorption.