Is there long term skeletal change with functional appliances?

Padhraig Fleming has written this post on a new study that evaluated the long term effects of functional appliances.  He provides excellent insight into this interesting area of orthodontic treatment.

Introduction

Clinicians and researchers are still interested in whether we can modify mandibular growth in the long term. This has sparked both controversy and consternation. One school of thought claims that we can modify facial growth by changing the environment.  Conversely, another viewpoint is that mandibular growth is genetically pre-ordained. As a result, orthodontic interventions may be unable to modify this permanently. Instead, they produce temporary acceleration leading to favourable occlusal changes. As such, it appears that we may ‘get the growth when we need it’. However, we lose this supplementary growth at the end of treatment.

Previous studies

This debate has prompted the publication of some landmark studies.  These trials have uniquely extended follow-up of periods of up to 10 years. Investigators in North America (North Carolina, Florida and Pennsylvania) and the U.K. did these studies. They all assessed the relative merits of early versus later intervention. They were consistent in their conclusions and suggested that early intervention is less efficient and no more effective than later treatment emerged.

The previous studies all incorporated a positive (treated) control group with the essential difference between the groups being the timing of that active phase of treatment. This was necessary because of the ethical dilemma posed by depriving children of orthodontic intervention during a period of active and useful growth. Consequently, investigators have not considered the use of a randomised design involving an untreated control to this point.

Despite the absence of this type of study from the orthodontic literature. The authors of this systematic review aimed to provide further clarity on the skeletal and soft tissue effects produced by functional appliances in treated versus untreated Class II subjects in the long term.  A research team based in Oxford, U.K., and Milan and Genoa in Italy did this review.

What did they do?

They conducted a systematic review to assess the cephalometric skeletal and soft tissue effects produced by functional appliances in treated versus untreated Class II subjects in the long-term. Importantly, they chose an end-point of either the end of active growth or three years following cessation of active treatment. They defined the end of active growth as being either age 18 years or CVM stage 5 or 6.

The PICO framing the review question was:

Participants: Class II patients (aged 16 years or under)

Intervention: Functional appliance treatment either in isolation or in combination with fixed appliance therapy

Control: Untreated Class II subjects

Outcomes: They used a range of cephalometric measurements. These included anteroposterior skeletal relationships and positions, the total mandibular length, length of mandibular ramus and body, as well as soft tissue changes of the lips and chin.

The authors hoped to include randomised controlled trials, controlled clinical trials, controlled before-after studies, and case-control or nested case-control studies. As they were unable to identify any randomised controlled trials, they used the ROBINS-I tool to evaluate the quality of evidence of the included studies.

What did they find?

They identified a final sample of 8 studies. They described these as retrospective controlled clinical trials. In six of these studies, they found that the control group was historical data.  The authors reported on the risk of bias (rather than methodological quality). They felt that this ranged from moderate to critical. They pointed out issues including selection bias, deviations from the intended interventions and the confounding effects of subsequent or concurrent fixed appliance therapy.

Despite the concerns associated with potential biases and confounding, they did a meta-analysis. They published a series of forest plots and did not include more than two primary studies in any of these. It is also notable they compared data related to both rigid and flexible Class II correctors allied to fixed and removable appliances.

The authors concluded the following:

‘Functional appliances, worn alone or in combination with multi-bracket therapy, may be effective in correcting skeletal Class II malocclusion in the long-term. The increase in the mandibular length may contribute to the improvement of the maxillo-mandibular relationship, although it brought about a negligible or non-significant improvement of the mandibular projection. We found  ‘very low’ quality of evidence for most of the outcomes at both primary time points; the clinical significance of these findings was limited. Further, randomised controlled trials evaluating clinically and essential patient outcomes are needed to confirm or reject the results of this review.’

What did I think?

I thought that this was an interesting and ambitious review. The search strategy and reporting were clear. However, I cannot help feeling that this review was always destined to raise more questions than answers. There are no randomised controlled trials or indeed, prospective longitudinal studies comparing the long-term effects of functional appliance therapy with untreated controls. As such, the quality of the supporting evidence and confidence in any conclusions has to be limited.

We can question the decision to undertake meta-analysis on several levels. In particular, as the author’s highlight, the quality of the primary studies was generally lacking. When investigators include these types of studies in meta-analysis risks this merely recycles the results from weaker studies. Besides, there was significant clinical heterogeneity between the studies and they compared rigid, removable functional appliances (e.g. FR-II and Bionator) and flexible Class II correctors (Forsus) used in conjunction with fixed appliances. Furthermore, there are problems with using historical controls. The authors mention these limitations in the Discussion and Conclusions; however, I wonder if it may have been more appropriate to highlight the available evidence using a narrative approach rather than attempting meta-analysis of such diverse and limited primary studies.

The authors suggest that further randomised controlled trials would be welcome. This is undoubtedly the case; however, we do need to be mindful of the ethical implications of this type of study. I am not sure that we can deprive adolescents of Class II correction by randomly allocating a subset to an untreated control group? It may be possible; however, I think we need to be cautious about this and mindful of the potential social and psycho-social implications of failure to treat Class II malocclusion in school-going adolescents.

What can we conclude?

While the evidence concerning the stability of occlusal correction associated with functional appliance therapy is compelling, there continues to be a lack of convincing evidence of a meaningful, permanent skeletal effect. On paper, a randomised controlled trial or even a prospective cohort study involving an untreated control would be required to answer this question with certainty. I, for one, remain unconvinced that we can justify one. The growth debate is set to rumble on.