Keep pushing! A new trial on expansion.

This month certainly seems to be “expansion time”. This post is about a new trial looking at the effects of MARPE and RME.

A couple of weeks ago, I published a post on Maxillary Skeletal Expansion. I wrote that there was limited evidence on the benefits and risks of these appliances.  It was, therefore, great to see this new trial on the use of Micro-implant assisted rapid palatal expander (MARPE) versus traditional RME.

Before we get into this post, I want to clarify the type of expander that I am discussing. This is because there is an almost religious zeal to follow different gods of expansion and this can be confusing.  I think that there are two main types of mini-screw implant-retained expanders.  These are:

• Micro-implant Assisted RPE (MARPE). These appliances localise the lateral forces to the mid palatal suture and minimise unwanted tooth movements from tooth-borne RME. We use these in older patients. (Carlson, Sung, McComb, MacHado, & Moon, 2016; Lee, Park, Park, & Hwang, 2010). MARPEs are either entirely or partially bone-borne expanders.
• The Maxillary Skeletal Expander (MSE). This appliance is a type of MARPE anchored by bi-cortical engagement (palatal and nasal cortical bone layers) of the implants. They are supposed to promote expansion in the nasal cavity’s posterior and superior aspects (Cantarella et al., 2017, 2018; Moon, 2018).

The main difference between the two is that the MSE screws go through the palate into the nose.

It was great to see a study that looked at the airway rather than morphological change following treatment.  This study was registered, and you can find the information here.

The registered trial’s primary outcome was a change in the angulation of the palatal root of the first molar to the palatal plane (Tor) as assessed in the sagittal view on Cone Beam CT slices.  They also based the sample size on this outcome. This was not the case for the paper, and I will look at its implications later.

What did they ask?

They did this study to find out:

“Was there a change in nasal ventilation parameters induced by RME v Hyrax v Keles expander using CBCT data”?

What did they do?

They did a parallel-sided RCT with a 1:1:1 allocation.  The PICO was:

Participants:

66 growing patients (10-16 years old) with unilateral or bilateral crossbite with a transverse deficiency of at least 5mm.

Intervention 1.

Classic RME Hyrax expander with bands fitted to maxillary permanent molars and premolars.

Intervention 2:

Keles keyless expander with bands fitted to maxillary permanent molars and premolars.

Intervention 3:

Hybrid-Hyrax with bands on the first maxillary molars and 2 Benefit mini-implants inserted bilaterally to the mid-palatal suture at the level of the third rugae.

Control:

There was no untreated control

Outcome.

Treatment-induced changes in nasal airway ventilation (pressure and velocity) using CFD from reconstructed CBCT data.  Secondary outcomes were skeletal and dental changes.

They took a CBCT at the start of treatment. The patients activated the appliances by turning the screw two turns/day until the crossbites were over-corrected. The expanders were left in place for six months after they stopped activation. At this point, they took another CBCT in a supine position.

They built a 3D model of the airway from the CBCTs and Computational Fluid Dynamics (CFD) to simulate airflow. This was relevant to the upper airway’s ventilation condition, adenoids, soft palate, and palatine tonsils.

They also made many measurements of dental and skeletal variables.

What did they find?

They analysed a large amount of data, and this wasn’t very clear. This is my interpretation of their results.

Sixty-six patients entered the trial, with 22 in each intervention group.  Unfortunately, 8 of the Keles group were lost to the study because of appliance blockage, which did not allow expansion. They also lost 2 Hyrax and one hybrid-hyrax patient because of poor quality CBCTs.They did not follow an intention to treat analysis, and they did not analyse the data for these patients. This problem is crucial because it leads to bias in the results.

I will only look at the primary outcome, which was nasal ventilation and velocity.  I am also only going to look at the post-treatment data. This is because if there are differences in the outcomes, this can only be due to the interventions. This table contains the data

There were no differences across the groups.

They carried out several regression models to take the effects of baseline severity confounders into account.  These showed that the Hybrid-Hyrax expander performed better in terms of post-expansion maximum velocity and nasal expansion resolution. However, at no point did the authors mention that these differences were statistically significant. Notably, the univariate analysis that I included in the table above showed that there were no differences.  As a result, I cannot help feeling that this data was over-analysed in the search for statistical significance.  I was somewhat confused with their analysis, and some of their statements did not agree with the conclusions.

What did they conclude?

They concluded that

• Hybrid Hyrax showed greater effects in increasing nasal width and CSA.
• A similar trend was noted in reducing pressure and maximum airflow: However, comparisons did not reach statistical significance. (THIS MEANS THAT THEY MAY HAVE OCCURRED BY CHANCE).

In effect, no appliance was better than the other in airway parameters.  Finally, they concluded that RME resulted in the improvement of nasal skeletal parameters and simulated ventilation.

What did I think?

I thought that this was a disappointing study. I could not agree with the conclusion that RME improved ventilation. The authors could not make this conclusion because there was no untreated control group. The changes that they reported may have occurred with simple facial growth.

I was also very concerned that the primary outcome presented in this paper was ventilation. However, in the trial registration documents, this was molar angulation. This issue means that they changed the primary outcome for the publication. This is poor practice and is sometimes done when a study team is searching for significant findings.

When I looked at the statistics, they reinforced this feeling.  The authors carried out a large number of statistical tests on a small sample.  For example, they did three regressions and 20 univariate comparisons. I wonder if this was a “fishing expedition” looking for statistical significance. A significant risk of this type of approach is finding spurious statistically significant findings, which can be over-emphasised. I can’t help feeling that this is what happened here.

I am not sure if there was a statistician as part of the study team. While I am not an expert in statistics, I cannot help wondering if these statistics are flawed.

Final comments

I think that this paper is somewhat like a film that starts well. Yet, when you get to the end, you are disappointed with what you have seen.  Most of the problems lie in the statistics, and I do not understand why the referees did not pick this up.  This paper could have been much more helpful.

Another critical factor is that they reported short-term changes.  While this may be useful, it would have been more helpful to have reported longer-term effects or some measure of patient-reported outcome such as disordered breathing.  This study appears to be a lost opportunity.

I, therefore, remain to be convinced that we should do this type of expansion as a predictable remedy for nasal obstruction in the short-term. I am sorry that I cannot be more positive.