June 12, 2017

Which is better Invisalign v Fixed appliances: Does this study help?

Which is better Invisalign v Fixed appliances: Does this study help?

There is no doubt that the development of aligners has changed orthodontics.  However, I often wonder if we know whether aligners “work as well” as fixed appliances.  This post is about a newly published study that provides us with some information.

While aligners have become part of the standard delivery of orthodontic treatment, I have seen few studies that have evaluated their effectiveness.  I think that this lack of research may reflect the nature of this fast moving field. This is because developments in appliances may make studies, which may take five years to complete, redundant.  Nevertheless, this does not mean that new appliances should not be studied.

I came across this new study in the AJO. It is a retrospective study. I do not usually review this type of research. However, because it may be “the best we have got”, I have decided to have a look at it.


Gu J et al

Am J Orthod Dentofacial Orthop. 2017 Feb;151(2):259-266. doi: 10.1016/j.ajodo.2016.06.041.

A team from Columbus, Ohio did this study. I had close links with this department several years ago. I even went to watch the Columbus Clippers baseball team.

In their literature review they pointed out there was little evidence underpinning the use of aligners.  They wanted to ask the following question:

“Was Invisalign as effective as Fixed appliances”?

What did they do?

They carried out a retrospective study by screening the records of orthodontic patients who had been treated in the University Department at OSU.  They looked at the records of 1500 conventional and 250 Invisalign patients. The orthodontic Faculty and residents had treated the patients. They stated that they did not select the patients according to the quality of their treatment.  The main criteria for the selection were that the patients needed to be over 16 years old, treated non-extraction and had available records.

At the end of their review of records they identified 62 fixed appliance and 61 Invisalign patients. They then removed the patients who they felt had terminated their treatment early.  The final sample was 48 patients in each group.

The primary outcome was PAR scores. A secondary outcome was the duration of treatment. They carried out an appropriate statistical analysis.

What did they find?

They looked at the pre-treatment features of the two groups. They did not find any differences except the Invisalign group were older (26 years old) than the fixed appliance group (22.1 years).

At the end of treatment they found the following;

Maxilla9.6 (8.4-10.7)1.7 (1.3-2.0)2.8 (2.3-3.1)
Mandible7.4 (6.5-8.2)1.1 (0.8-1.3)2.1 (1.6-2.5)

You can see that there are no real differences in the PAR scores, apart from the CI for final PAR being wider for Invisalign and this suggests there is possibly more variation in the final treatment finish with Invisalign.

When they did their statistical analysis they also showed that fixed appliances were significantly more effective at reducing PAR scores than Invisalign. Nevertheless, there was no difference in the end of treatment result.

Their overall conclusions were that:

  • Fixed appliances were more effective than Invisalign at improving malocclusion
  • Treatment with Invisalign was faster

What did I think?

Firstly, this was a retrospective study.  While this type of study may provide us with some information we should look carefully at the methods. This is because we need to evaluate any evidence of selection bias in the sample.  When I looked at this I wondered whether the method of selection of the patients records could have caused some bias.  For example, we need to consider that the final sample included 48/1500 (3.2%)  conventional and 48/250 (19.2%) Invisalign cases from the record store. We do not know much about the large number of patient records that were not included in the final sample. Most importantly, they also rejected cases that they felt terminated early. We have to consider whether the results would have been the same if these terminated cases had been included.  While, I may be being critical here, we do have to give the authors credit for their detailed reporting that allows us to consider the risk of bias.

My next step in assessing the paper was to think about whether the bias could have influence the results.  It is important to remember that we should not make assumptions on the direction of any bias. It can go either way.  As a result, I think that we can conclude that there are limited differences between the results of Invisalign and fixed appliance treatments.  However, we need to be cautious and the authors point this out.

This study also provides us with great data that can be used for a sample size calculation for a definitive trial. As a result, it is a useful study.

I have just spotted a new systematic review on the effectiveness of aligners. In my next post, I will discuss this review.  Lets see if we can find out more..

Related Posts

Have your say!

  1. Like the posts
    Would like to read more on this blog

  2. Kevin any reason you decided not to discuss treatment time based on the Table you provide us?
    Interesting they discarded patients they felt terminated early. For cases that seem to take little more than 12 months how early may be early.

    • Hi Carlos, thanks for the comments. My feeling was that this was where there was the most bias because of the early terminations etc and I felt that I should not discuss it.

  3. Hi Kevin,
    thanks for looking at this topic. It is fraught with danger as you point out, with usual research pitfalls in addition to those associated with delayed publication involving a rapidly evolving technology. There are so many to point out, I’m sure that you will ID them in your review of the review!
    I have not yet seen a prospective paper studying completed cases using the current material ( if discussing Invisalign). We have studied it at Melbourne university, but only for isolated movements such as expansion. When Invisalign was first launched,not only was the material different but the manufacturing, attachments, aligner features as well as programmed velocity have changed, confounding data used to make conclusions relevant today (perhaps conveniently those cynics out there would say).
    I do think it is important to remember that it is not fixed appliances nor aligners treating patients and reducing PAR scores, but the clinician behind the appliance; so when you wonder above “if we know whether aligners “work as well” as fixed appliances”, we should consider the clinician and their experience with the appliance they are using. As the great Sheldon Baumrind points out, to evaluate the ability of appliance systems, ideally we would study those systems in the hands of experienced clinicians with those specific appliances to eliminate clinician experience as a significant variable in “appliance ability”. It may be a case of fixed appliances produce better, same or worse results when used by the novice than do aligner systems. ( Note I’m not suggesting the clinicians in this study were inexperienced. I’m alluding to the fact that these systems have a different skill set and they may be influenced by clinician experience differently, whatever level that may be).
    In published papers to date I have not seen any that refer to who prepared the Invisalign ClinChecks – analogous to who actually was bending the wires or bonding the brackets.) Doesn’t matter who inserts the plastic, its the programming that involves (should) the majority of orthodontic skills. Also, the true test of Invisalign, I feel, is not only the resulting PAR score clinically, but to what degree did the clinical result reflect the programmed one? Both are necessary to evaluate the appliance system ability. (eg. you can have a crappy clinical result but if the ClinCheck program is the” same crappiness” (and I see this often), the aligners actually expressed close to what the clinician programmed!). We use Geomagic software to compare the clinical result with the programmed one. Would be neat to have the PAR scores of the ClinCheck as well as the clinical result in the aforementioned study.
    So I guess my point is that when we look at this question, we should ask the same of both appliance systems, be sure to evaluate the systems and not just the clinicians, and additionally get rid of that perception blindness that is so tricky to shake. …Taylor Swift song coming on…VV

    • “Would be neat to have the PAR scores of the ClinCheck as well as the clinical result in the aforementioned study.”

      I am not sure if the readers of this blog are aware of this development but I guess somebody was listening at Align Technologies since the new Itero Element Scanner with upgrade 1.5 software (just releases USA last week) gives that information when using the Progress Assessment Utility.

      Par Score: Initial, Current, Programmed, Final. The particularly nice thing about this upgrade is that the evaluation can now be made with a PVS based ClinCheck. Prior to this, it is my understanding that the assessment software would only work with a scanned based ClinChecks.

      Incidentally the progress assessment utility gives a treasure trove of data: vertical, sagittal, transverse, arch length etc. This certainly gladdens the heart of techies like myself.

  4. One factor that might explain the difference in PAR score and treatment time is that fixed patients are in braces until the orthodontist thinks they are ready and Invisalign patients have distinct end points when they finish a series of aligners. If we forced aligner patients to have the same treatment time as fixed patients I suspect their PAR scores would be lower although probably not as low as the fixed cases since compliance would falter as they get closer to ideal results. Excellent review, I do enjoy the blog very much.

    • Thanks very much, I think that the duration data must have been influenced by the factors that you have listed, however, I was not concerned that this is where the bias mostly lay.

  5. Thanks for reviewing this paper. I do quite a bit of Invisalign so I am particularly interested in these studies. The first thing that came to my mind when I read the paper is that it was done at a school setting. From my experience particularly with university/school settings the treatments many times are not set up for success. Also, this study looked at patients from 2009-14. the aligner material was changed during that time frame. the new material is quite a bit more flexible and I have found it to fit better. Also, scanning became more prevalent which quite possibly makes the aligners fit and work better. One of the findings was that the aligner cases finished 5 months faster. Why? If the treatment was not up to standards why not make more aligners? Due to these concerns I found the study not to be useful for me. There is very little reason on the type of cases that were studied excellent results were not obtained by either method.

    • Hi David, I agree with you, to a degree, this is because the PAR scores showed an excellent standard of treatment. This means that the treatment was effective. Your comments about the changes in material are very true and this is why orthodontic research becomes outdated with rapid developments.

  6. Thanks for this interesting review.

    I have used fixed appliances (SW and Tip Edge) for 27 years or so, and Invisalign for about 12-14. I have over 200 Invisalign cases under my belt so far. I have noticed:
    1: Invisalign is definitely faster as a general rule, for mild to moderate cases. There is little or no round tripping of teeth.
    2: Invisalign can do 80% of cases to a good standard, as they used to claim. Some it can`t do, usually the more extreme cases as you`d expect.
    3: Invisalign is useless at derotating teeth well, and very poor at shifting lower canines.

    This is all subjective I understand, but just my two pennyworth.

    Best wishes

    Andrew Adey

  7. Thanks for reviewing this article Kevin,

    In my practice one of the main reasons for patients not succeeding with invisalign is a lack of compliance. It’s not surprising that the invisalign treatment times are shorter because 14 months is about the maximum attention span for many invisalign patients. I have had a handful of invisalign patients who haven’t even achieved basic alignment before losing interest or going into braces. For this reason I feel that eliminating patients who terminated treatment early is a selection bias. I don’t think this study is a true reflection of my clinical practice. The lack of inclusion of extraction cases is also not a reflection of my practice. I’d like to see a study that compares invisalign extraction cases with braces extraction cases.

    With invisalign, smile direct club and dental monitoring joining forces, for how much longer will we be able to say that the clinician is behind the treatment result? The appliance has been around long enough now that the suppliers think that they can do it without us so I feel that it is time to really see some research that shows just how effective invisalign can be.


    • here is 1 I am aware of Lachlan.

      Int J Clin Exp Med. 2015; 8(5): 8276–8282.
      Published online 2015 May 15.
      PMCID: PMC4509355
      The effectiveness of the Invisalign appliance in extraction cases using the the ABO model grading system: a multicenter randomized controlled trial
      Weihong Li, Shimei Wang, and Yanzhen Zhang

      As far as I know, Invisalign (Align Technology) and DM (dental monitoring) have discussed but are not joining forces. I suspect Align may have their own version – Progress Assessment Utility – , as Barry pointed out above – thx Barry!. VV

      • Thanks a lot for the link to this study. Much better conducted than the one which is the subject of this blog .

      • Thanks Vicki,
        I’ll check out the article.
        Regards Lachlan

      • Hi Vicki, I checked out this paper as I was going to post about it. But the Journal that published it is something called Beall’s list. This is a list of journals that are classified as questionable, scholarly open-access publishers. As a result, of this I do not think that we should give much credence to this paper?

  8. I have serious doubt about any study aiming at differentiating two modalities of treatment and using the PAR for doing so. The PAR is more an epidemiologic index than a clinical one IMO. No evaluation of the root position, no evaluation of the vertical dimension and general esthetic. Plus doing a comparative study on invisalign which is notoriously weak in controlling the root and choosing an index which do not look at root speak of a serious flaw in design.
    If I was doing this type of studies and I wanted the result to be ” the clinical results are more or less the same ” I would choose the PAR .

    For evaluating difficulties of a case and finish of the case you do not weigh each parameter the same way, hence you need two distinct index. One for grading difficulty of the case and one for grading results. The way the american board of orthodontics do it.

Leave a Reply

Your email address will not be published. Required fields are marked *