The influence of sidedness in unilateral cleft lip and palate on mid facial growth at 5 years of age

Registration

This study was nested within a larger ongoing investigation of midfacial growth in children born with complete UCLP (IRAS: 225282, Research Registry UIN: 4068). The purpose of the larger study is to investigate intrinsic factors leading to poor mid facial growth via the analysis of pre-surgical 3D scans of the maxillary arch.

Participants and exposure

Children were included if they were born in the United Kingdom (UK) between 2000-2014 with a complete UCLP and receiving treatment from nine UK cleft centres: West Midlands, North Thames, Evelina, South West, Spires, North West, Scotland, Leeds and Newcastle. For inclusion, children needed dental models to have been taken in infancy prior to primary surgical intervention and at the age of five years. The exposure variable was UCLP sidedness, and this was determined by two authors (L.E., B.R.) from 3-D scans of dental models of infants prior to their first surgical procedure. Patient records were used to confirm the complete UCLP cleft subtype and to identify the co-variable of patient sex.

Outcome measure

The outcome variable was maxillary growth, assessed via occlusal relationships between the maxillary and mandibular arches on dental models taken at 5 years of age and analysed by the 5-Year-Old’s occlusal index described by Attack et al.¹⁵ The 5-Year-Old’s index has been accepted as the UK national audit tool at five years of age for the assessment of occlusal relationships in the primary dentition, prior to any orthodontic intervention.¹⁶

The 5-Year-Old’s index assigns a score of 1 (good) to 5 (poor) on an ordinal scale, based on the relationship of upper and lower dental model casts. Each cleft centre sent their dental models to be assessed by two nationally calibrated orthodontists, who scored the models individually. Differences were resolved through discussion to reach a consensus. All orthodontic examiners underwent a calibration process via the UK Orthodontic Clinical Excellence Network, with inter- and intra-rater kappa scores above 0.8, in order to be accepted as a scorer in this process.

Analysis

The relationship between UCLP sidedness and maxillary growth was analysed initially using descriptive statistics and cumulative percentages. These associations were quantified by Odds Ratio (OR, 95%CI) and p values.¹⁷ Due to current contention about the best way in which to analyse ordinal data¹⁸, we took two regression approaches: First we calculated the OR of having a better outcome comparing left and right UCLP by logistic regression for each of the 4 potential cut points in the 5 Year-Old’s ordinal data. Each cut point dichotomised the 5-Year-Old’s index scores into better and worse outcome groups. If the common odds assumption was demonstrated to be satisfied, we used a proportional odds logistic regression to estimate a common odds ratio on sidedness effect.^19,20 Adjustment was made for sex in each regression model in recognition of this potential explanatory variable in orofacial cleft sidedness research.¹ Statistical analysis was performed using the R Project for Statistical Computing, version 4.3.1 (https://www.r-project.org)

Summary of key findings

There was a trend for better maxillary growth in left compared to right sided UCLP at 5 years of age (OR 1.24, 95% CI 0.85 to 1.82; p=0.27), yet the uncertainty around the effect estimates in the regression analysis of ordinal 5-Year-Old’s Index scores implies the evidence for a difference in growth between left and right sided UCLP is weak.

Strength and weaknesses

This study is strengthened by the sample of patients, which was heterogeneous in terms of the complete UCLP phenotype, and the consistent measurement of exposure and outcome variables. The inclusion of multiple UK cleft centres increases the applicability of the findings, yet introduces variability in terms of surgical techniques, timings and sequences, which are known to have an impact on maxillary growth.²¹ We adjusted for the co-variable of sex as this was readily available in our dataset but we were not able to add additional potentially important variables to our model, such as ethnicity.

The 5-Year-Old’s Index is an important outcome as it has been accepted as the recognised national audit measure in the UK and has been recommended for use in primary dentition.¹⁶ The long-term predictive ability of the 5-Year-Old’s Index has been questioned, particularly of poor growth (scores 4 and 5 on the 5-Year-Old’d Index), which has shown poor correlation with growth outcomes at skeletal maturity.²² The analysis of ordinal data in healthcare research is notoriously challenging and there is a paucity of a gold standard in terms of analytic strategies.¹⁸ Dichotomising the ordinal scale is appealing for simplicity yet the resulting effect estimate is only applicable for the arbitrarily chosen cut point and does not account for the variation of the data within each ordinal level.¹⁹ The use of the proportional odds logistic regression analysis maintains the integrity of the 5-Year-Old’s Index ordinal data and has been encouraged in healthcare outcome studies to retain clinically relevant information.²³ The proportional odds model does rely on the assumption that the sidedness effect is constant across all cut points.²⁴ This assumption was supported by our data with unadjusted odds ratios (Table 2) falling within a similar magnitude.

Comparison to other studies

Whilst we are not aware of previous studies analyzing the relationship between UCLP sidedness and maxillary growth using the 5-Year-Old’s Index, other studies have investigated dental arch relationships using different outcome measures. Haque et al. ²⁵ in a case series from Bangladesh reported no difference in dental arch relationships between left and right UCLP at a mean age of eight years (before orthodontic or orthognathic intervention) using the GOSLON yardstick, dichotomized into a binary outcome of favorable versus unfavorable growth (OR 1.12, 95%CI 0.409 to 3.11; p=0.83). Staudt et al.,²⁶ in a case series from Switzerland reported better post-treatment occlusion at a mean age of 19 years for left sided UCLP using Modified Huddart Bodenham scores, although the prevalence of orthognathic surgery was higher for left UCLP (65% vs 20%). An Iranian study investigating cephalometric measurement of facial growth at an unspecified age, reported no difference in most of the 28 measurements reported between right and left UCLP (including vertical and sagittal skeletal, dental and soft tissue measures), but within the four measurements noted to be different there was greater mandibular length in left sided compared to right sided UCLP.²⁷

Interpretation and further work

The trend for better growth in left sided UCLP, reported in this study and previous studies, may suggest a subtle intrinsic influence of sidedness on maxillary growth outcomes. The severity of UCLP in terms of size of the alveolar cleft and circumference of the dental arch has been associated with maxillary projection.¹¹ There is increasing evidence that sidedness may also be related to severity in the UCLP phenotype; right UCLP has been associated with a higher prevalence of additional congenital malformations^28,29 and anatomical hypoplasia and asymmetries in the face.^2–4,30 If patients born with right sided UCLP truly are at higher risk of maxillary growth disturbance, this has implications on patient counselling, and treatment options that might attempt to mitigate the increased risk.

The uncertainty around the effect estimate in this study may have a number of reasons. First, it is possible that UCLP sidedness does not impact maxillary growth. Second, the sample size may not have been sufficiently large enough to differentiate subtle differences in maxillary growth by sidedness. Third, the outcome measure at five years of age may not be the most appropriate to determine definitive growth outcome. Fourth, additional variables not considered within our model, both extrinsic (such as surgical technique) and intrinsic (such as ethnicity) may have an important influence. Further work is required to analyse maxillary growth in a large heterogenous sample of patients born with complete UCLP, ideally at skeletal maturity.