Fixed retainers do not prevent relapse: A long term study.
One of the greatest challenges in orthodontic treatment is our ability to retain our treatment. Even though orthodontic retainers have been researched for years, our level of knowledge about retention is surprisingly low. This new paper provides us with interesting, but perhaps, disappointing information.
Jenette Steinnes et al
Am J Orthod Dentofacial Orthop 2017;151:1027-33
A team from Tromso in Norway did this interesting study. In their introduction, they pointed out that fixed orthodontic retainers are the most commonly used form of retention in Norway. However, there are few long term studies of their effectiveness. They set out to answer the following question.
“What is the stability of orthodontic treatment outcome and retention status 7 or more years after placing fixed retainers”?
What did they do?
They identified a sample of patients who they had treated in the public orthodontic service from 2000-2007. All the patients had originally presented with more than 4mm crowding in the mandible or maxilla. They also had to have full records. They found 105 eligible patients and sent an invitation to them asking if they could attend a follow-up examination. When they came to this appointment the authors took impressions for dental casts and collected data on the process of treatment, the type of retainer, compliance and satisfaction with treatment. They then scored the study casts with the PAR and Little’s Irregularity (LII) Indices.
What did they find?
67 (64%) patients attended for the appointments. The average time between the end of treatment and the second data collection was 8.5 years. 90% of the patients had been fitted with a mandibular fixed retainer. In the maxilla 54% had a fixed and removable retainer, 39% had a removable only and the remainder had no retention (4%) or fixed only (3%). At the long term recall visit some of the patients had stopped wearing their retainer or it had debonded. This meant that they had a sample of patients with and without retainers at the second visit. As a result, they could make comparisons between these groups to identify the effect of retention.
They provided a large amount of detail on the dental changes in retention and I have summarised them in this table.
Little's Irregularity Index (means and 95% CIs)
|Maxilla||9.6 (8.4-10.7)||1.7 (1.3-2.0)||2.8 (2.3-3.1)|
|Mandible||7.4 (6.5-8.2)||1.1 (0.8-1.3)||2.1 (1.6-2.5)|
Increase in LII (mm) in maxilla and mandible
They also found that in the maxilla there was no difference in the amount of relapse between the patients who did or did not have retainers. That is the retainers did not prevent relapse. I presume that this was because of failure in bonding etc, but they did provide any additional data on this.
When they looked at patient satisfaction, most of the patients were satisfied with the treatment outcome. However, more than half noticed that their teeth had moved after treatment.
What did they conclude?
They concluded that
- Occlusal relapse occurs irrespective of the use of long-term use of retainers.
- A fixed retainer is effective in the mandible, but it did not appear to make any difference in the maxilla.
What did I think?
I thought that this was an interesting paper that outlined an ambitious project. When I looked at the results it was interesting to see that despite most of the patients wearing their orthodontic retainers, the average relapse in PAR index scores was 14%. This is clinically significant, but it may be composed of a number of minor factors, for example, mild lower incisor crowding and increase in overbite.
Importantly, the use of retainers in the maxilla did not have an influence on changes of maxillary incisor irregularities. However, in the mandible the patients without a retainer had significantly more relapse.
When we read a study like this we need to interpret the results by thinking about several important points. These are:
Effect size and confidence intervals
The differences that they detected between the groups were rather small. For example, the amount of relapse, according to Little’s Index was only 1mm. I am not sure if this is clinically significant. The confidence intervals are not too wide and this shows that we may have moderate certainty in the data.
This was an ambitious study and they recalled a high number of patients who had various levels of success in wearing their retainers. Their choice of outcome measures was logical. The PAR index measured all components of relapse and the LLI was directed at visible incisal relapse. They also used patient satisfaction as an outcome that is relevant to our patients. Interestingly, the results reflected clinical experience as there was more relapse in the lower incisors.
This was a retrospective study and one of their selection criteria was that the patients had good records. This introduces selection bias because we are not sure why some of the patients had complete records and other potential patients did not.
While the response rate is high. We must assume that those who attended for the follow-up appointment were different from those who did not attend. This is response bias and is important. For example, the patients who had relapsed may be more likely to attend. Alternatively, the patients who attended may be pleased with their long term result and want to share this with the orthodontists!
Orthodontic retainers; Putting all this together..
My feeling is that this is a study that may be considered to be “best that we can get”. I think that it provides useful information and reinforces that relapse is inevitable, particularly in the lower incisors. It also shows that the concept of long-term retention may not be the answer to preventing relapse. However, there is a degree of uncertainty present because of the methods and response rate. So it is up to you to decide whether these findings are robust.
Let’s see if we can have a good discussion on this in the comments to this post.