Short-term Anteroposterior Treatment Effects of Functional Appliances and Extraoral Traction on Class II Malocclusion: A Meta-analysis

In the treatment of Class II malocclusion, treatment possessing the capability to alter patients' facial growth is of particular interest, namely by means of functional appliances, extraoral traction (EOT) appliances, or a combination of both. Over the years, numerous investigations have evaluated the possibility of growth modification with these appliances. However, the results are generally equivocal, with conflicting evidence as to their effectiveness. Studies suggest that the changes, albeit present, are not very predictable and not always significantly different from those occurring either without treatment or with conventional fixed appliance systems.1 A large variation in interindividual response and small mean changes could mean that differences may be more attributable to study design than treatment.2,3 In addition, any dentofacial change may be due to normal growth that occurs irrespective of therapy.

Recent reports demonstrate that some improvement in jaw relationships can be achieved during early treatment with headgear or functional appliances.4–10 However, despite the long history of functional appliances, there continues to be much controversy related to their use, effectiveness, and mode of action.11–14 Moreover, there is a large individual variation in the results of treatment.12 Often, when there is lack of success of functional appliances, this is attributed to patient compliance as well as the inability to control the amount and direction of mandibular growth.15 EOT appliances, or headgear, on the other hand, are used to redirect or restrain maxillary growth and to distalize maxillary molars.16–19 As with functional appliances, practitioners often cite patient compliance as a critical role in treatment success with headgear, with this compliance being extremely variable among patients.20 

The aim of this study was to evaluate, based on published data, short-term anteroposterior skeletal and dental effects on Class II malocclusion in growing patients following treatment with functional appliances and/or EOT, using lateral cephalograms. As sample sizes in investigations are relatively small, this study aimed to combine data from existing literature to increase the sample considerably.

A literature search was performed using PubMed and Ovid (including OLDMEDLINE) as well as the Cochrane Library to identify orthodontic articles reporting treatment of Class II malocclusion using functional appliances or EOT. Terms used in the search were functional appliances, extraoral traction or headgear, and Class II, combined with clinical trial or randomized controlled trial. A further search, for the sake of verification that all articles had been located, was carried out using activator, bionator, twin block, and names of specific functional appliances as opposed to simply functional appliances. Other databases (Web of Science, Google Scholar Beta, Embase, Extenza, African Journals Online, Bandolier, Evidence-Based Medicine, Latin American and Caribbean Center on Health Sciences Information, Bibliografia Brasileira de Odontologia, ChinaInfo database) were also searched for completeness. The search was expanded by searching references of articles consulted. Full-text sources available on the Internet for the American Journal of Orthodontics and Dentofacial Orthopedics, European Orthodontic Journal, and the Angle Orthodontist were also searched to validate that the search had identified all relevant articles.21 

Articles were selected for inclusion and analysis if the following criteria were met.

• —Human studies

• —Pertained to removable functional appliance and/or EOT appliance use in Class II malocclusion treatment

• — In the form of prospective clinical trials

• — Treatment carried out on growing patients with age ranges mentioned

• — Duration of treatment mentioned (minimum duration of 9 months to avoid erroneous annualized values)

• — Availability of a suitable control group (untreated Class II individuals)

• — Measurable pretreatment and posttreatment cephalometric values, as well as changes during treatment, for SNA, SNB, ANB, and overjet

• — Sufficient data available for statistical calculations

In the case of more than one publication about the same patient group, the most informative and relevant article was included. For studies stating one or more but not all of the desired variables (SNA, SNB, ANB, overjet), the corresponding author22–24 was contacted, and values for these variables were also obtained from the raw data. A quality analysis was also carried out according to the methods described by Petrén et al.25 Studies were described as being of low, medium, or high quality. Only those falling into the medium- and high-quality categories were included for analysis.

The data provided in the included studies were divided into different patient groups, which were patients treated with activators, patients treated with twin block, patients treated with EOT, patients treated with combination appliances (including both functional and EOT components), and untreated Class II control groups. No distinction was made between different types of activators or EOT appliances. Herbst appliances were excluded as these often tend to be fixed and would bias the results with regard to compliance. Fränkel appliances were also excluded since they may differ from other functional appliances in their mode of action, probably by causing an increase in both apical bases and maxillary and mandibular arch widths.11,26–29 From the identified studies, the changes referring to the maxilla (expressed by SNA), mandible (expressed by SNB), intermaxillary relationship (expressed by ANB), and overjet were analyzed.

Data were subsequently entered into the meta-analysis program of the Cochrane Collaboration's Review Manager Software (RevMan 4.2.8, released July 8, 2005). Descriptive statistics were calculated for the different variables in the different groups including arithmetic mean and standard deviation. Treatment time between studies varied from 9 to 24 months, and therefore, annualizing results standardized this variation.

Using the random effects model, forest plots were drawn and significance tests carried out (calculating P values). Heterogeneity tests were also performed. If confidence intervals for the results of individual studies (depicted graphically using horizontal lines) have poor overlap, this generally indicates the presence of statistical heterogeneity. More formally, statistical tests for heterogeneity are available. These include χ² tests as well as calculation of I². The χ² test is based on the sum of the squared difference between the treatment effect of each individual trial and overall treatment effect, weighted by the inverse of the variance in each trial. It assesses whether observed differences in results are compatible with chance alone. A low P value (or a large χ² statistic relative to its degree of freedom) provides evidence of heterogeneity of treatment effects (variation in effect estimates beyond chance).30 

A more useful statistic for quantifying inconsistency, and present in the forest plots, is I² = [(Q − df)/Q] × 100%, where Q is the χ² statistic and df is its degrees of freedom.31,32 This describes the percentage of the variability in effect estimates that is due to heterogeneity rather than sampling error (chance). A value greater than 50% may be considered substantial heterogeneity.30 

The original search located 171 articles. From these, 168 were human studies. A total of 159 of these were carried out on growing patients. The articles were further narrowed down to 56 that contained relevant data for this study. These were read, and a total of 9 suitable studies2,5,22–24,33–36 were identified following consideration of all inclusion criteria (Figure 1 and Table 1). The results described below are summarized in Table 2 and Figures 2 through 5.

The twin block group showed a mean decrease of 1.03 degrees when compared to the control group (P = .02), with low homogeneity (I² = 81.2%), while the EOT group showed a mean decrease in SNA of 1.01 degrees when compared to the control group (P < .00001), but with high homogeneity (I² = 0%). No significant change in SNA was seen in the other groups when compared to controls.

With regard to SNB, activators revealed a mean increase of 0.66 degrees when compared to the control group (P = .04), but with a heterogeneous response (I² = 83.4%). A homogeneous (I² = 0%) mean increase of 1.53 degrees was seen for the twin block group (P < .0001), as well as a homogeneous (I² = 22.2%) mean increase of 1.05 degrees for the combination group (P < .00001). For the EOT, no significant change was found in SNB.

Activators displayed a mean decrease in ANB of 0.92 degrees in comparison to the control group (P < .00001), with mild heterogeneity (I² = 54.5%). The twin block group showed a heterogeneous (I² = 77.9%) mean decrease of 2.61 degrees in ANB (P < .00001). A mean decrease in ANB of 1.38 degrees is seen for the EOT group when compared to the control group (P < .00001), with mild heterogeneity (I² = 53.6). Finally, the combination group also showed a mean decrease in ANB of 1.8 degrees (P < .00001) but was more homogeneous (I² = 44.8%).

Concerning overjet, activators exhibited a large mean reduction of 3.88 mm when compared to the control group (P < .00001). This change, however, is highly heterogeneous (I² = 93.7). The twin block group showed a heterogeneous (I² = 83.3) 6.45-mm mean decrease in overjet when compared to the control group (P < .00001), while the combination group also showed a homogeneous (I² = 5.6) decrease in overjet, with a mean of 4.37 mm (P < .00001). The EOT group showed no significant change in overjet.

Functional appliances (activators and twin block) improve the sagittal intermaxillary relationship mainly by their effect on the mandible and show an important dental effect by overjet reduction. Twin block appliances also show a significant effect on the maxilla. The skeletal changes are brought about by stimulation of condylar growth37–42 as well as a contribution by a certain amount of fossa advancement.43–46 They also seem to display a growth-restraining effect on the maxilla.44,47–50 Besides the small sagittal skeletal base improvement influencing overjet, the dentoalveolar effect on overjet is brought about by palatal tipping of maxillary and labial tipping of mandibular incisors, respectively.44,48,51,52 

Differences were found for the different groups of functional appliances, namely, activators and twin block. To a large extent, these differences could be due to the amount of hours per day that the patients are instructed to wear their appliances, this being up to 24 hours per day for twin blocks.23 

EOT appliances also improve the sagittal intermaxillary relationship, demonstrating a large effect on the maxillary skeleton. They appear to achieve this growth modification by means of a sutural response.53–56 These appliances show very little clinical improvement in overjet.

Combination appliances (with both functional and EOT components) seem to affect the sagittal intermaxillary relationship by acting mainly on the mandibular skeleton as well as having a dentoalveolar effect on overjet. A possible reason as to why no significant change in the maxilla was seen is that dentoalveolar changes as recorded by the decrease in overjet have somewhat of a buffer effect and therefore reduce maxillary changes that may otherwise have taken place.

An accumulation of data from existing studies implies that the sample size will increase dramatically, and hence even small skeletal changes will be statistically significant. However, statistically significant changes do not necessarily correspond to noticeable therapeutic effects. Changes seen in our data analysis are often small but statistically significant, which has also been suggested by Proffit and Tulloch.57 Thus, too much weight should not be placed on statistically significant changes but rather on clinically significant ones. A cutoff value for clinical significance was set at 1 degree for angular measurements and 1 mm for linear measurements, as it is often described to be the error in lateral cephalogram measurements.58 For overjet measurements, there is also the concern of magnification. However, since studies do not usually mention this, and since the influence that this would have on measurements is minimal, this was not taken into account.

It is important to consider to what extent the results of studies are consistent. Consistency is estimated by the overlap of confidence intervals or more formally using statistical tests for heterogeneity (χ²). Some argue that since clinical and methodological diversity always occur in a meta-analysis, statistical heterogeneity is inevitable whether or not we happen to be able to detect it using a statistical test. Methods such as I² have been developed for quantifying inconsistency across studies, and these methods move the focus away from testing whether heterogeneity is present to assessing its impact on the meta-analysis and describing the percentage of variability in effect estimates that is due to heterogeneity rather than sampling error.

The variation in treatment success undoubtedly is affected in part by patient compliance. Uniformity was desired with regard to compliance; thus, trials using appliances fixed to the mouth were excluded. Other factors contributing to variability include patient age, patient maturity, growth pattern, severity and etiology of the initial condition, treatment timing, soft tissue characteristics, and amount of force applied.

Soft tissues may be a reason for variability in treatment outcomes, following functional appliance treatment. Not all individuals respond similarly to this treatment, and considerable variation is present among patients. Given a certain mandibular advancement, the forces applied through the functional appliance by the soft tissues may differ due to myotatic reflexes or variation of the soft tissue viscoelasticity.59 This is perhaps one of the reasons why there is higher heterogeneity in functional appliance treatment outcomes.

One area that has not been investigated much is that of long-term treatment changes. Appliances that have a supposed orthopedic effect may cause changes, but whether these changes are maintained long term and to what extent is a further question. Dermaut and Aelbers60 concluded that on a long-term basis, the few studies that provide scientific data report that orthopedic changes induced by Class II therapy are only temporary. This does not, however, hold true for dentoalveolar changes that were generally found to be more stable. Stability is also suggested to be dependent on the age when treatment was carried out. If treatment is delayed, the treatment results tend to be more stable.61 

The literature search in this investigation was executed mainly in PubMed and Ovid, resulting in a higher number of included articles than if only one database had been used. It should be borne in mind that there are limitations to the literature search. These limitations include the fact that some studies may not be included in Medline and others may not have been published (introducing publication bias). Medline is regarded, however, as a powerful and relatively accurate tool in retrieving orthodontic literature.21 Data collection revealed that studies differed in areas such as age, sample size, control groups, and appliances. Data were, however, grouped as best as possible, weighting studies according to number of patients and annualizing mean changes. In the inclusion criteria, the treatment duration minimum was set at 9 months since any treatment time less than that would mean that the methodological error is likely to increase and would give an overestimation when annualized. The random effects model of meta-analysis was thus used. A control group was necessary, providing a comparison consisting of untreated growing Class II individuals displaying changes due to natural growth alone.

• Appliances believed to have a growth modification effect used in the treatment of Class II malocclusion growing patients, namely, functional and EOT appliances, are associated with an improvement of the sagittal intermaxillary relationship.

• Activators and twin block appliances accomplish this mainly by acting on the mandible. Twin block appliances also show a significant change in the maxillary skeleton.

• EOT achieves this by acting on the maxilla.

• Combination appliances mainly bring about changes in the mandibular skeleton.

• In addition, activators, twin blocks, and combination appliances reveal a large decrease in overjet, which is not the case in the singular use of extraoral traction.