Dental arch size and shape after maxillary expansion in bilateral complete cleft palate: A comparison of three expander designs

Cleft lip and palate is the craniofacial anomaly with the highest prevalence in the human population.¹  Bilateral complete cleft lip and palate (BCLP) is the most severe cleft lip and palate type, maintaining the embryological maxillary division and challenging the multidisciplinary treatment team.² 

In untreated patients, the premaxilla maintains its anterior projection and the posterior segments tend to approximate, resulting in a narrow dental arch.³  Lip and palate repair surgeries often have a negative influence on maxillary growth and development, resulting in maxillary dental arch constriction and maxillary sagittal deficiency.⁴  As a consequence, anterior and posterior crossbites and maxillary arch perimeter reduction are frequently observed, and maxillary expansion is often required.⁵  Many appliances have been used for maxillary expansion in cleft lip and palate patients, including the Quad-helix appliance (QH) and the Hyrax expander (H). Both appliances have been shown to be effective for maxillary dimension improvement, with the QH achieving a greater intercanine expansion.^6–8 

Transverse maxillary constriction in BCLP results in a triangular-shaped dental arch.³  Intercanine distance decreases more markedly during growth than does the intertuberosity distance, requiring greater expansion in the anterior region.⁹  For this reason, the expander with differential opening (EDO) was introduced for achieving different anterior and posterior amounts of expansion in BCLP.¹⁰  EDO promoted a greater expansion in the anterior region of the maxillary dental arch, with no skeletal differences from the traditional Hyrax.¹¹  No previous morphological analysis of the expander outcomes was performed in BCLP. The purpose of this study was to compare the effects of H, QH, and EDO on maxillary dental arch size and shape in patients with BCLP.

This retrospective study was approved by the ethics committee (protocol 1.991.298) of the Hospital for Rehabilitation of Craniofacial Anomalies at the University of São Paulo, Brazil. All patients with BCLP were treated in a single center from 2011 to 2013. The inclusion criteria for the study were age between 7 and 10 years; lip and palate repair, respectively, performed at 3 and 24 months of age; presence of first permanent molar; presence of maxillary arch constriction; and need for maxillary expansion. Exclusion criteria were associated syndromes, history of previous orthodontic treatment, and absence of teeth to support the expansion appliance. After searching patients' records, 75 patients with BCLP were initially selected (55 males and 20 females), with a mean age of 8.7 years.

The expanders (Figure 1) were supported by bands adapted to the first permanent molars or second deciduous molars. When bands were placed on second deciduous molars, a distal arm was present in order to include the permanent first molar. H and EDO also had C-shaped clasps bonded to the deciduous canines. In the H appliances, the screw was activated two quarter turns in the morning and two quarter turns in the evening until the posterior maxillary teeth palatal cusps were aligned with the mandibular posterior teeth buccal cusps. After the active phase, the appliance was kept for retention for 6 months.

QH appliances were constructed using 0.036-inch stainless-steel wire and activated 6 mm every 2 months until overexpansion of the molars was reached, using the same protocol as for the H group. QH anterior extensions were activated against the premolar and canine palatal surfaces, distally rotating the banded molars, until an approximately 2-mm overcorrection was reached in the canine region. After the expansion phase, the QH was also maintained for retention purposes for 6 months.

Both anterior and posterior screws of the EDO were activated two quarter turns in the morning and two quarter turns in the evening until the maxillary molar palatal cusps were aligned with the mandibular molar buccal cusps. On the following days, only the anterior expander screw was activated until an approximately 2-mm overcorrection was achieved in the canine region. The amount of anterior region expansion depended on the severity of the maxillary constriction of each patient. The appliances were maintained as retainers for 6 months.

Dental models were obtained before expansion (T1) and 6 months after the active phase when the appliances were removed (T2). Two dental models from the group H were lost. Ten patients in each group were excluded during the analysis because of tooth absence due to the transitional dentition or because the expander was supported on the second deciduous molars.

The dental models were scanned using a 3Shape R700 three-dimensional scanner (3Shape A/S, Copenhagen, Denmark). After scanning, the digital models were imported into the software Stratovan Checkpoint (Stratovan Corporation, Davis, Calif). Twelve homologous landmarks, described in to a previous study,¹²  were placed by one investigator on the T1 and T2 maxillary dental models (Figure 2). The x,y coordinates for each landmark were collected. The raw landmark coordinates were imported into the software MorphoJ (Klingenberg Lab, Manchester, UK).

For arch size analysis, the centroid size of each dental arch in T1 and T2 was calculated from the raw coordinates and used as a maxillary dental arch size measurement. For arch shape analysis, a Generalized Procrustes Analysis¹³  was performed using the software MorphoJ. With this method, non-shape variation was removed from the raw data, expressing pure shape differences between groups. A mean shape dental arch was determined for each group and time point.

The sample size was determined by the arch size analysis. For a minimum difference of 1.31 and a standard deviation of 0.98, each group required 12 participants for an alpha error of 5% and a test power of 80%.

To determine the method error, the same investigator replaced the 12 landmarks in 100% of the sample after 3 weeks. The random error was calculated using intraclass correlation coefficient (ICC) for both x and y coordinate values. SPSS Statistics, version 23 (IBM, Armonk, NY), was used.

Male and female values were compared using a t-test (SigmaPlot 12.0) to assess sex dimorphism. The sample distribution was tested using the Shapiro-Wilk normality test. Intergroup comparisons for arch size were performed using analysis of variance (ANOVA). SigmaPlot 12.0 (Systat Software, San Jose, Calif) was used. The mean shapes of the maxillary dental arch were compared using ANOVA (Procrustes ANOVA)^13–15  in the software MorphoJ. Results were regarded as significant for P < .05.

ICCs showed a high degree of reliability for repeated landmark placement (0.999 for the x coordinates and 0.987 for the y coordinates). There were no significant dental arch size differences between males and females at T1 (Table 1), so the data for both sexes were combined.

There were no significant dental arch size differences among expanders, either at the pre- or postexpansion phase (Table 2).

Before expansion (T1), differences in arch shape was found between the QH and EDO groups (Table 3). After expansion (T2), differences in arch shape were found between all groups (Table 3). Both the QH and EDO expanders changed the maxillary arch shape between T1 and T2, while the H expander showed no change in arch shape between T1 and T2 (Table 4).

Geometric morphometrics has been used in biology and anthropology to observe species variation in evolutionary and biological processes.¹⁶  Landmark-based geometric morphometrics involves summarizing shape regarding a landmark configuration, providing the intuitive visualization of the shape and the spatial localization of shape variation easily with graphical representation.¹⁷  Morphometrics has been previously used as an alternative cephalometric analysis,^18,19  overcoming limitations of the traditional method that does not separate size from shape.^20,21  For the current study, centroid size was used as the size measurement. Centroid size is an isometric estimator of size, calculated as the square root of the sum of the squared distances between each landmark and the centroid of the form.²²  Generalized Procrustes Analysis was performed for shape analysis. With this method, non-shape variation was removed from the raw data by translating all dental arches to a common location (same centroid), rescaling all dental arches to unit centroid size and rotating all dental arches into an optimal least-squares alignment (Figure 3). All differences in location, size, and orientation were removed, expressing pure shape differences between groups.²³  Procrustes Analysis has been used in growth and maturation studies,^24,25  facial profile analysis,^26,27  skeletal shape evaluation,²⁸  and even in facial attractiveness.²⁹ 

The main limitation of the method was the need for 12 homologous points to be placed on digital models for running the Procrustes superimposition. In the case of a missing tooth, a corresponding point was not placed and the subject was excluded from the sample.

There was no statistically significant difference in dental arch size between all groups at T1 and T2. Results at T2 were in agreement with those of Almeida et al.,⁸  in which no differences were found between the orthopedic changes of H and QH expansions in BCLP patients. The greater intercanine expansion observed in groups QH and EDO may have been compensated for by a greater posterior movement of the maxillary incisor in these groups, which might explain the similarity in the dental arch size with group H at T2. Maxillary expansion achieved by the three appliances produced effective dental arch size changes in BCLP patients.

There were no sex differences found in dental arch size (Table 1). These results agree with those of Silva Filho et al.,³  who showed that sex had no influence on the upper dental arch dimensions of cleft patients. It seems that the presence of the cleft in itself played a more important role in the determination of the dental arch dimensions.

A difference was found in the comparison between QH T1 and EDO T1 dental arch shapes. However, these findings may be related to the premaxilla deviation to the left or right side, a striking feature of BCLP patients,³⁰  as observed in Figure 4. There was a statistically significant dental arch shape difference between all groups at T2 (Figure 5). However, the intragroup comparison showed no dental arch shape difference between H T1 and H T2, indicating that only the QH and EDO promoted changes in the shape of the dental arch after maxillary expansion (Figure 6). The greater expansion in the anterior region of the maxillary dental arch described by Dalessandri et al.³¹  and Tindlund et al.³²  using QH and by Garib et al.¹¹  using EDO in BCLP patients appeared to be responsible for these shape differences. These changes in dental arch shape are positive for BCLP patients as a result of the triangular morphology of the dental arch described previously.⁵ 

Another change observed for QH and EDO with shape superimpositions (Figure 6) was the posterior movement of the central incisor, indicating that this posterior movement was enhanced by the greater intercanine expansion. This result was previously reported in cleft³³  and noncleft³⁴  patients. At the same time that a more parabolic dental arch shape was achieved with greater intercanine expansion using QH and EDO, a Class III interarch relationship can be impaired by the posterior movement of the premaxilla. The sagittal hypoplasia of the midface is already expected in BLCP patients as a side effect of the lip and palate repair.⁴ 

An interesting difference among groups was the first molar rotation after expansion with the QH, which was not observed in the H and EDO groups (Figure 6). QH anterior extensions were activated against the premolar and canine palatal surfaces, distally rotating the banded molars. The ability to rotate molars with the QH during maxillary expansion was previously reported by Vasant et al.³⁵ 

The orthopedic effect of the maxillary expansion in BCLP patients was expressed as a lateral displacement of the posterior segments, leading to an alignment of these segments with the premaxilla. The maxillary expansion is important to prepare the maxilla to receive the secondary bone graft. The alignment of the maxillary segments, better achieved in the QH and EDO groups (Figure 6), provides lateral walls for performing the alveolar bone graft.³⁶  Maxillary expansion with the H appliance may lead to posterior overexpansion in order to align the maxillary segments in the cleft area.¹⁰  Further studies will be needed to analyze this clinical effect.

• H, QH, and EDO promoted similar changes in maxillary dental arch size in patients with bilateral complete cleft lip and palate. H did not produce changes in the maxillary dental arch shape.

• When dental arch shape changes are desired, with greater intercanine than intermolar increase, either the QH or EDO may be used.