Does MARPE work in non-growing patients?
Orthodontic expansion is a current area of clinical and research interest. As we have discussed before, a relatively recent development has been using implants to anchor expansion devices. There have been several new studies on this technique. This paper is the latest from a long-established research team.
One of the stated advantages of using Mini-Implant RPE (Maxillary Skeletal Expansion) is more significant skeletal expansion than other techniques. This may be particularly relevant to older patients when splitting the suture is problematic. We have posted about this before, and there may be increasing evidence to support the technique. The amount of expansion using MSE in patients of differing ages was looked at in this retrospective study.
A large team centred in Michigan, USA, did this study. The AJO-DDO published the paper.
Craig McMullen, Najla N. Al Turkestani, Antonio C. O. Ruellas, Camila Massaro, Marcus V. N. N. Rego, Marilia S. Yatabe, Hera Kim-Berman, James A. McNamara, Jr, Fernanda Angelieri, Lorenzo Franchi, Peter Ngan, Hong He, and Lucia H. S. Cevidanes
AJO-DDO online advanced access: DOI: https://doi.org/10.1016/j.ajodo.2020.12.026
What did they ask?
They did this study to answer the following question;
“Can MSE achieve maxillary skeletal expansion in growing and non-growing patients”?
What did they do?
The authors did a secondary analysis of a study they did not reference. In effect, this was a retrospective evaluation of selected CBCT images. The team obtained these from the files of the University of West Virginia. The selection criteria were:
- Orthodontic patients who had posterior unilateral or bilateral crossbites. Treated with MSE.
- The patients had a transverse maxillary skeletal discrepancy
- They had no previous orthodontic treatment
- The records were pre-and post-expansion CBCT scans with a FOV, including the cranial base. Notably, the images had to be of adequate quality.
The investigators grouped the patients according to their growth stage (confirmed by CVM) into two groups. These were growing and non-growing. The growing group had to have a maxillomandibular differential index of >16.4mm, and for the non-growing group, this measurement was 19.6mm. I will return to this measurement later, as it is crucial to the understanding of this paper.
They did a sample size calculation that showed that they needed 11 patients in each group. Unfortunately, they did not reference any previous studies that they used for this calculation. I tried to repeat the sample size calculation and this showed that they needed to analyse 19 patients in each group.
They carried out an exact method of measuring the dental and skeletal changes of the CBCT images. The team are experts on this method.
Then they generated 26 individual measurements and ran simple “t” tests across the groups.
What did they find?
They collected the records for 11 growing and 14 non-growing patients.
They presented their results in a great deal of detail. As there were many measurements, I have decided to concentrate on what I felt were most important. I based this on the question that they asked. As a result, I focused on some dental and skeletal transverse measurements. Unfortunately, they only presented data on change. As a result, we had no accurate indication of the pre and post-expansion values. Nevertheless, this information was helpful.
This table includes the change in widths in mm. I have also calculated the 95% Confidence Intervals, as these are essential when we look at data of this type.
|Variable||Growing group||Non-growing||Diff (95% CI)||P|
|PF point||3.9 (1.3)||2.1 (1.3)||1.8 (0.71, 2.84)||0.02|
|Nasal cavity point||3.6 (1.5)||1.9 (1.2)||1.7 (0.58,2.82)||0.006|
|Molar cusp tip||5.5 (2.8)||3.6 (2.1)||1.9 (-0.12, 3.9)||0.1|
The most important things to observe from this table are that the differences between the two interventions are minor. But, more importantly, the 95% confidence intervals are very wide. Consequently, there is a high level of uncertainty in the data.
Their overall conclusions were;
“The MARPE appliance is an effective way of treating patients with transverse maxillary deficiency regardless of their sex or maturational change”.
What did I think?
This paper reported a tried and tested retrospective methodology. This method involved obtaining sets of patient records after treatment had been completed. Then carrying out a detailed analysis of the records to generate many variables and finally, analysing each variable with a simple univariate analysis. Unfortunately, this type of study has now been superseded by alternative designs, for example, prospective cohort studies and RCTs. Nevertheless, its main advantage is that it does not need extensive resources. However, this comes with a “trade-off” of bias and uncertainty in the findings.
The main issues with this study are:
The authors did not provide any information on the number of patient records they excluded from the sample. As a result, the sample must suffer from considerable selection bias.
I also could not understand or replicate the sample size calculation. Furthermore, the groups were unbalanced, and I could not find a reason for this in the text. This led to very wide confidence intervals, representing uncertainty in the data. As a result, I am not sure they could suggest greater skeletal changes in the growing group.
They also did not provide any information on the expansion they aimed for during treatment. This is important because there may be differences between the groups at the start of treatment. For example, one group may have needed more expansion. This could easily explain the differences that they detected. In my view, it was disappointing that they did not present data on the pre and post-treatment widths. This would have allowed us to establish whether the treatment had a clinically significant effect. Instead, they only published the widths changes, which was not as helpful.
The analysis method is at risk of false positives because of the multiple comparisons. This is because if you test related data, such as craniofacial measurement, you increase the chances of finding a statistically significant difference. In other words, if you look hard enough, you will find something. After all, a 5% level of statistical significance means that out of 20 tests 1 result is expected to be a false positive!
In addition, using simplistic statistics did not allow them to adjust for the baseline outcome values, account for any imbalances, and possibly produce more precise estimates. In addition, the study statistician should have considered using multivariable analysis. This would account for potential confounders. For example, there is a large imbalance in gender between the 2 arms despite the nonsignificant result in table V.
Finally, I was also not clear why they compared the two groups. Because, if we go back to the aims of the study, it was not to compare the two interventions, it was to see if expansion was possible in patients who were growing and non-growing. I was just confused by this.
Results of the treatment
There was nothing in the paper about the clinical results of treatment. For example, were all the patients successfully treated, how many of them had a mandibular displacement, and what did they feel about their treatment?
My final point is that I cannot think of a single clinical reason for taking a wide field of view CBCT on children at the end of their expansion phase of treatment.
I want to return to the aim of their study. This was to test if MSE could achieve expansion in growing and non-growing patients. The authors concluded that it can, but we cannot be very confident from this study. I am sorry to be critical, but there is a danger of this paper being quoted in support of MSE. I think that the best way to think about this paper is that it is an exploratory study, which would be very useful to help plan future trials. In this respect, it is valuable.
Emeritus Professor of Orthodontics, University of Manchester, UK.