AcceleDent has no effect upon orthodontic root resorption; the trials are coming thick and fast!
AcceleDent has no effect upon root resorption; the trials are coming thick and fast!
We have another report on a trial of AcceleDent. This time the investigators looked at orthodontic root resorption. The results are interesting.
This paper reports on the analysis of data on orthodontic root resorption from a trial that I have previously discussed. This, in effect, is a subset of the data from this large study. A team from the South of England, mostly in Kent, wrote this paper. This is in the beautiful South and is very different from the North, where I live.
Effect of supplemental vibrational force on orthodontically induced inﬂammatory root resorption: A multicenter randomized clinical trial
Andrew DiBiase et al. Am J Orthod Dentofacial Orthop 2016;150:918-27. http://dx.doi.org/10.1016/j.ajodo.2016.06.025
Firstly, I would like to point out that AcceleDent have never claimed that using their device will result in a reduction in orthodontic root resorption. The authors based their study question on the premise that there is an increased risk of root resorption with prolonged orthodontic treatment. As a result, shortening treatment with Acceledent may reduce root resorption. This is what they tested in their study.
What did they do?
They aimed to investigate root resorption that occurred during tooth alignment with fixed appliances, supplemented with vibrational force provided by the AcceleDent device.
The null hypothesis was that the supplemental vibrational force did not affect the levels of resorption during treatment.
Their PICO was;
Participants; orthodontic patients under 20 years old with mandibular crowding requiring extraction of mandibular premolars and fixed appliance treatment.
Intervention; AcceleDent or sham AcceleDent
Control; treatment as usual (no accident)
Outcome; amount of orthodontic induced idiopathic root resorption
They randomly allocated a sample of 81 patients to the interventions. They measured root resorption using long cone periapical radiographs which they took at the start of treatment and at the end of alignment.
The method of randomisation and allocation were good. An examiner blinded to the intervention allocation measured the radiographs and the standard way.
They carried out a sample size calculation using the primary outcome of the main study. This was the rate of tooth alignment. This is not the outcome reported in this paper and is important. I will come back to this later.
They carried out a relevant and complex statistical analysis which used multivariate techniques that took relevant confounders into account.
What did they find?
They found that the amount of root resorption at T3 ranged from 0.00 – 3.6 mm, with a mean of 1.08 mm for the whole sample. When they looked at the effect of the intervention, they found the following:
|Duration of treatment (months)||17.4 (16.85-17.75)||20.74 (20.21-21.27)|
|Change ANB||-0.75 (-0.07-1.43)||-0.12 (-0.49-0.25))|
|Change OJ (mm)||-3.36 (-3.89, -2.83)||-2.43 (-2.93, -1.93)|
In effect there was no difference between the groups. It is important to note that the differences between the groups were very small and the 95% confidence intervals were large.
In the discussion, they pointed out that their results were similar to other investigations of root resorption. They also justified the use of traditional radiographs as opposed to CBCT because of radiation concerns.
What did I think?
I have posted about this well designed study before. While I thought the original study was very good, I do have several concerns with this paper.
Sample size and power
My greatest is that the sample size of this study was based on tooth alignment. This means that the study was powered to measure this variable. As a result, when we analyse a secondary outcome the study may lack power to detect a difference. I have made this mistake before!
The authors address this problem in their discussion and carried out a power calculation for root resorption using their data. This revealed that the study had a power of 25 to 30%. This means that it is unlikely that this study would have detected a difference between the interventions because it did not have enough statistical power. However, we should also look at the size of the difference between the groups. You can see that this is small. This also means that if the power was increased there is a chance that a similar difference would be detected. We need to bear this in mind when we consider the findings from this study.
Timing of data collection
I think that is also important to consider the time points of data collection. They collected their data at the end of alignment. I feel that it would have been more relevant if they had collected this data at the end of treatment. This is because that when we consider root resorption we need to take into account the use of rectangular wires and Class II elastics.
I am also aware that one criticism of the study design is that the investigators did not take patient cooperation into account. They could not do this because the timers in the AcceleDent devices failed. I disagree with this suggestion. This is because we should consider that this is a “real world” study. In this case we must accept that all the patients may not comply with the treatment. This means that the investigators have studied the effect of giving (or selling) the device to the patients. This is a robust methodology
In summary, I feel that this study only provides us with an indication of whether root resorption is influenced by AcceleDent. It does not provide us with a robust answer because of the lack of power. As I have mentioned before, much more work needs to be done on the new interventions to make teeth move faster before we can recommend these to our patients.
Emeritus Professor of Orthodontics, University of Manchester, UK.
Have your say!
Given that power is an apparent problem, I think the sham group is a meaningless waste of subjects. It deserves a bit of criticism, despite the fact that it’s part of an RCT.
Calculating the power after the analysis is considered to be meaningless and not the best approach.
Sample size calculation and power assumptions should be considered at the design stage.
After the analysis, precision is assessed by looking at the confidence intervals for the between group difference.
I have no comments about the study – but I was born in north England before I relocated to the US. My sister lived in London (south England). The contrast in the you pics you selected isn’t a true reflection if taken on the same day, according to my experience. Why do northerners get such a bad press? Maybe there’s a joke I’m missing! Happy Holidays everyone 🙂