Is Invisalign effective at changing incisor torque?
Several papers published over the past year have shown that Invisalign may not achieve all intended tooth movements, including mandibular advancement, overbite reduction, and incisor movement. One of the most critical tooth movements is changing incisor angulation. Importantly, this is one of Andrew’s six keys. Achieving the correct torque is important for aesthetics and correcting inter-arch relationships. Researchers have carried out limited research into the effectiveness of clear aligners in torquing incisors. A new paper from a team based in Adelaide addresses this question.
Nancy Rajan, Tony Weir and Maurice Meade
AJO-DDO advance access https://doi.org/10.1016/j.ajodo.2024.06.008
This paper is open access, so everyone can read it without being a member of the AAO!
The study is one of several that use data from the Australian Aligner Research Database. This database contains information on 12,000 Invisalign patients who had been treated with Invisalign at the start of the study. Specialist orthodontists treated these patients between 2013 and 2021. It is proving to be a valuable tool in understanding the effectiveness of Invisalign. This is a unique and valuable research resource.
What did they ask?
They used the information from the database to ask
“Is there a difference between planned and achieved lingual root torque of the maxillary central incisors in patients treated with the Invisalign appliance when a greater than 10 degree change in inclination was prescribed”?
What did they do?
The team identified patients from the Research Database using the following inclusion criteria.
- Prescription of a change in incisal angulation of greater than 10 degrees
- Patients aged >18 years old with a complete dentition treated with Invisalign
- They were compliant with prescribed wear protocols as determined by the treating orthodontists
- All records were complete
- The operator did not use IPR or intermaxillary elastics.
They exported the .stl files and compared the predicted to the actual tooth movements. From this information, they calculated the percentage accuracy of the torque change achieved.
The team used the appropriate statistical analysis of the data. Importantly, they used regression to identify predictors of any discrepancy between the final and predicted tooth movements.
What did they find?
They identified 63 patients (126 incisors). This sample comprised 42 females ((66.7%).
Their main finding was that the initial angular root torque to a reproducible vector was 76.7 (SD=7.9) degrees, the planned was 60.4 (+/-6.2) degrees, and the achieved was 69(+/—7.8) degrees. These figures reveal that the treatment only achieved 41.9% of the planned torque.
Importantly, 78% of the incisors had a clinically significant shortfall of greater than 5 degrees. Significantly, most of these had an underexpression of 10-15 degrees.
Finally, the regression analysis revealed that weekly or biweekly wear protocols or the use of power ridged did not influence the accuracy of planned torque expression.
Their overall conclusion was.
“The Invisalign appliance achieved 41.9% of the planned root torque changes. Almost 80% of maxillary incisors had underexpressed torque by at least 5 degrees”.
What did I think?
This study is another interesting and clinically relevant piece of work from this team, utilising a large database of patient records. It’s great to see this kind of research.
The authors also acknowledged some of the limitations of their methodology, with the most significant being the study’s retrospective nature. Despite selecting cases from an extensive database, selection bias may influence their findings. Notably, the direction of this bias is unknown.
They also pointed out that operators need to over-correct to compensate for the shortfall in planned tooth movements.
The message from this research is clear. However, some authorities (Invisalign KOLs) have criticised this methodology because the predicted tooth movement represents a force system and not the final tooth position. Dr. Moshiri highlighted this in a paper in the AJO Clinical Companion.
Authors comment
I asked Maurice Mead, the corresponding author to comment on this and other issues with their research issue, and he stated.
- Importantly, we are not stating “CAT fails”. we have no idea what the intended outcome of the doctor is in their minds.
- We know that the digital treatment plan is supposed to be analogous to an “archwire design” and not an ABOOGS final outcome. However, it is apparent that for most orthodontists, including those (about 50% of whom are or were KOLs for Align Technology) who have contributed their databases to our research, the use of torque and other overcorrections is minimal.
- What we wanted to establish in this project was:
- How much of a given degree of torque prescribed is clinically delivered by an initial series of aligners? That way, we can plan for research-informed overcorrection amounts. Therefore, whether cases examined had “overcorrection” or were not programmed is irrelevant—a given amount of torque prescribed will result in a certain percentage of the outcome.
- However, we also looked at, and are looking at, the types of orthodontic tooth movement that occur – ie torque, translation, controlled and uncontrolled tipping. The findings from this study, the corresponding study of the lower arch, shortly to be published in the Angle, and Jiang et al.’s 2021 investigation indicate that when torque, translation or controlled tipping are prescribed, most cases get uncontrolled tipping. We cannot see how this finding can be dismissed by those who claim that ClinCheck predictions represent a force system.
Final comments
Several papers have now demonstrated consistent results indicating a difference between the planned and actual tooth movements. This leaves us in a dilemma. Should we accept the findings suggesting that aligners may not be as effective as hoped, or should we trust Invisalign’s dismissal of these findings due to measurement methods?
My understanding and teaching in Invisalign is that you have to compensate for the ineffective movement – for example increase the amount of deep bite correction so your set shows and AOB
Over rotate teeth
Over procline etc
That seems to be the only way to get what you plan for – an expect lots of refinement!!!
Oh and be prepared to switch to fixed!
Better still start with fixed ti make a difficult case easier and then finish in Invisalign!
Overcorrection of all aligner tooth movements has, or should have been, routine for years. Thankfully, Drs Rajan, Weir and Meade have given us additional scientific underpinning for that practice. Dr P R Begg, the Australian orthodontic pioneer, was a great believer in overcorrection, mainly in anticipation of the relapse tendency towards the original malocclusion. Aligners not delivering 100% of the force is a further reason to overcorrect. I have a meme that I teach postgraduate orthodontic students: If your final ClinCheck looks good, it’s bad.
Agree to your observations. A tool can ‘t deliver predicted results,when even human intelligence may falter. Aligners are just like NSAID,available over the counter.,for larger population.
I am left with several thoughts.
The first is that I suspect that the research may have been too generous with their inclusion criteria. My suspicion is that they over estimated the maxillary incisor torque that Invisalign achieved.
Specifically, I was unable to find if there was a differentiation between “true” incisor torque increase and “relative” torque increase.
A true torque increase would result in the incisal edge not moving facial or apical while the measured torque increases.
A “relative” torque increase is what we see when the maxillary incisor crowns are protruded anteriorly or intruded. In both of these instances the perceived “torque increase” actually results from the incisors rotating about their CR. The crowns move facial or apical while the apex moves lingual. This is only a relative torque increase and not the result of torqueing forces applied.
For research purposes, the former “true” torque increase might be thought of as real torque increase while the “relative” torque increase might be thought of as a false torque increase. I am not sure that a differentiation between true torque increase and relative torque increase was made. If not, this would tend to over estimate how much actual real torque increase Invisalign achieved, if any was achieved.
Indeed, even in the articles illustration of their method of measurement, a Class II division II appearing arch is used. As the incisors in their example are protruded, an increase in measured torque would occur. However, this is a “relative” torque increase resulting from rotation about the CR, not a true one.
That 50% of the data comes from current and former Invisalign KOLs is consistently with this as well. It is not hard to see posted cases online where torque is claimed to have been achieved, yet careful review reveals that what actually occurred was only a relative torque increase as the maxillary incisor crowns were either protruded or intruded.
Indeed, this is exactly why we see such a torque decrease in extracting cases treated with Invisalign, where incisor retraction is performed. Invisalign cannot achieve the moment to force ratios necessary to even maintain incisor torque as the incisors are retracted.
The next interesting point is the authors preemptive rebuttal to a classic Invisalign defense, which is that it does not achieve the shown ClinCheck movements because it is a “force system” and not a representation of the end result. The authors point is that there is frequently not as much over correction as is claimed, and the intended correction cannot be ascertained nor accurately compared to as a percentage achieved, as we do not know what was in the treating doctors mind as a goal.
That said I do believe that argument can be further strengthened. To that end I will borrow a phrase from former US Supreme Court Justice Potter Stewart, when attempting to define pornography. His famous definition was “I know it when I see it”.
I will say the same for Invisalign, when incisors need to have been torqued and they were not, I know it and I expect we all do. When meaningful Class II molar correction was not achieved, rotation of canines and premolars was not corrected, and deep bites not ideally opened through means other than relative opening through lower incisor rotation/protrusion, I know it and I expect we all do.
To this point, I have not seen a case where significant and meaningful true maxillary incisor torque increase was achieved with Invisalign, if not lost during incisor retraction in an extraction case. More, near every case I have seen where meaningful maxillary incisor torque increase has been claimed, the real cause was rotation about the CR as the incisor crown was protruded or intruded, and only a relative torque increase was achieved.
Though, as a suggestion, let’s start looking at our cases as to whether the end result was achieved as ideally possible or not. Are the incisor properly torqued, are the roots parallel, are the marginal ridges level, has class II or class III been corrected, is the overbite ideal, etc.
More, maybe we should look at the finished case and ask, were torque and other items in the case compromised because it was treated with Invisalign? If these have not been achieved with Invisalign, I don’t think it matters what the force system did or did not say, if the end result is not as it could have been or should have been, like Supreme Court Justice Potter Stewart, we know it when we see it.
nothin to add…
42 % of the planned root torque is a good number since a few years back we were told that it was not possible to torque at all with Invisalign in contrast to what I/we experienced clinically.
An aligner is a force system and I never expect it to have full expression in the form of tooth movements. It is the same thing when using fixed appliance and why I have take out and reactivate my arch wires, and that is why we always have to do additional aligners in complex cases.
As long as we can reach the desired torque, in a complex case, during 1-3 sets of aligners, both the end result and the treatment time will be comparable to the treatment time for a similar fixed appliance case.
The final comment, that “several papers have now demonstrated consistent results indicating a difference between the planned and actual tooth movements” does not leave the responsible clinician in any dilemma as long as the final outcome of an Invisalign treatment is in accordance with the final treatment goal. I would expect today, that most experienced clinicians are informing their patients before treatment that to reach the final treatment goal, we have to do more than one set of aligners to reach the planned treatment goal.
Excelente
There is a difference between ClinCheck and Reality Check!
It’s crucial to distinguish between a ClinCheck simulation and the reality of patient outcomes. Overengineering should be seen as an essential part of treatment planning with aligners.
For those who are surprised or even offended by these findings, consider this: When you bend wires, do you always achieve the exact tip or torque you intended? Haven’t you ever deliberately “misbracketed” to ensure that the movement you envisioned in your mind would actually occur? Do you believe that a 21×25 wire in a 22×28 slot will express the precise “programmed” torque, relying on some magical bracket to do everything for you? Are the bracket companies that pre-plan and bend your wires producing superior results? (if so, you should publish, fast!!! We want to see it!) If you still believe that what you see on the screen will be mirrored in real life, you’re setting yourself up for disappointment.
Treat your ClinCheck stages as you would view orthodontic wires outside the mouth—they are tools, not guarantees. Rather than trying to prove that aligners don’t work, focus on understanding how to make them work effectively. Teeth don’t differentiate between wires or ClinChecks; they respond to the forces applied over time, provided there’s sufficient space. If you could control peanut butter, you could move teeth with peanut butter; it would still come down to mastering the application of force.
Understanding and mastering your instruments and appliances—whether they are wires, brackets, or aligners—is what truly makes the difference in patient outcomes. Articles like this one are invaluable in helping dental professionals grasp these nuances and improve their treatment approaches.
Torque is really a meaningless term when it comes to Biomechanics. People use torque to describe one or more of four things – A twist in the wire, 3rd order inclination of the slot relative to the occlusal plane, 3rd order Incisor inclination or Root movement. To obtain root movement, the crowns have to be restrained. Once that occurs, the crown is stationary and the roots will move under the effect of the applied moment. So, in this case, it appears that the authors simply measured inclination change, which can occur as a consequence of simple labial tipping of the crowns. This is not root movement, which would require attachments and large M:F ratios. Thus far, there is no evidence to show that aligners can actually predictably generate root movement. Superimpositions are the only realistic method to evaluate this.
Excellent topic to carry out a research on, kudos to the team in Adelaide and thanks to you Kevin for finding and discussing it. Talking about fixed appliances, torque is one of the most difficult movements to achieve in orthodontics, difficult to introduce to the correct degree and to have the patience (a rare trait) to wait for it to express fully. On many an occasion you take the 19X25 st steel AW out, add a bit more or tone it back down a bit to achieve the right amount of it. If you add too little torque there is no change and if you add too much it could lead to root resorption. May I say many a post graduate students have not figured it out by the time they graduate. PR Begg himself has said, stage three of the Begg technique (torquing stage), takes double the time that stage 1 and 2 would take, one needs to be patient.
What make aligners more complex than fixed appliances is you cannot take an archwire out and introduce torque into it as you treat, every movement must be pre-planned and so however good a clinician you are, you would still end up with additional aligners or fixed appliances at the end. Added to this show me one person who wears the Aligners for 22hrs/day week in and week out without defaulting.
If we look at the history of orthodontics first came removable appliances, Why did they lose favour? Because patients were not compliant ! What do we do ? Stick them on their teeth and guess what we also have better control ! Fixed appliances were invented, 18-24 months later straight teeth, parents happy + orthodontist happy. However history has a habit of repeating itself, removable appliances are BACK, Aligners, which promise straight teeth without the hassel of poking wires or broken brackets, but with a difference we advertise straight to the consumer. The clinicians know what is the best appliance for the patients, but do the patients know, let us get to them before the clinicians do !
Paraphrasing Karl popper, every observation for a student of science is interpreted in the light of a theory, which only strengthens it. If we as clinicians have such a favourite theory, our aim must be to make sure that it is backed by research. Aligners definitely have a place in orthodontics are they good for torquing anterior teeth, I doubt it (like Thor says 42% is really good, I would take that), until i see good evidence to the contrary.
The posts above all make reference to overcorrection for tooth movement with aligners. All make reference to not viewing the Clincheck as the final tooth outcome, but as a force system, and making use of refinement aligners. The recent publication and others have shown that aligners fall short of the desired or planned tooth movement, whether torque control, intrusion/overbite reduction, or rotation are considered in the first set of aligners.
From my read, it also appears that all believe that over-engineering is required to achieve the final outcome and that refinement is essential.
My question is, how much over-engineering is required, and is that over-engineering expressed at the same success rate/ratio as that now affirmed by research into tooth movement by initial set of aligners ( 42% for maxillary incisor torque)? Do we know if this is a linear expression problem of an exponential one when even more desired tooth movement is attempted? Acknowledging that we are working with plastic that is deformable and has force decay properties that are not considered here.
In other words, if we attempt to improve maxillary incisors lingual root torque by 10 degrees and can only express 42% movement, does it mean that if we doubled the planned Clincheck torque (over-engineer) that we will express 42% of the over-engineered 20 degrees, or some other percentage, if this is not a linear force expression/decay problem? In addition, will the refinement set of aligners express the 42% success rate when applied to the next increased increment of torque correction required?
The real implications are what do we aim for when planning individual tooth movement, when should meaningful refinement be planned if lack of a linear movement model of 48% is assumed, and what does this mean with regards to treatment time and success in aligners?
Tooth movement is about M/F ratios..
In a system like Invisalign to get the necessary M/F ratio may need like 100-120 sets .
For complicated cases ,class 11 div 11 with extraction,and apical base correction with proper biomechanics it can be done in half the time .
Whoever can control the Cres location will prevail:
When orthodontic publications describe tooth movements based on definitions not grounded in the most basic mechanical laws, the results can mislead clinicians. Orthodontic physics is based on a minimal number of mechanical laws, among which are:
1. The paramount importance of accurately locating the center of resistance (Cres) in restraint by the periodontal ligament (PDL) and the alveolar bone tooth, which is essentially the equivalent of the center of mass (CM) of a free body.
2. Determining the type of tooth movement based on the relationship between the various vectors acting on the tooth and its Cres.
The exact Cres position in a tooth cannot be known at any given moment, as the system restraining the tooth comprises living biological elements that continuously change their properties. In Newtonian mechanics, reliance is never placed on estimated or less than 100% precise critical points! That is to say, when a specific vector is applied to a tooth, the necessity of knowing its relationship to the Cres is absolute. There is no approximation. The moment the exact location of a biomechanical point is unknown, all analytical assessments, vectors, and colorful diagrams only obscure the truth and, unfortunately, become meaningless.
The same applies to movement definitions: in tipping, the force vector does not pass through the Cres and causes the tooth to rotate around the point. Likewise, a couple that can be applied anywhere on the crown will cause the tooth to move around its Cres. Since the exact position of the Cres is unknown, the possibility of achieving translation does not exist. Therefore, in orthodontics, translation cannot be achieved with just a single vector. In other words, anyone claiming this is possible is mistaken and misleading.
The article is also misleading regarding the definition of torque. While Andrews accurately defined orthodontic torque in practical terms, the description in words was not precise. Torque is not merely a movement that changes the labiolingual inclination of the tooth, as the article mentions. Torque is the movement that alters the tooth’s inclination in all dimensions, with the rotational axis located at the crown. According to Andrews’ interpretation, with the specific torque built in the brackets, the axis of tooth movement during buccolingual torque application is located on the buccal side of the tooth, at the height of the bracket (at the point of vertical tangency to the tooth’s crown). If the axis of movement is located there, the conclusion is that, in fact, due to the significant restraining forces of the fixed appliance, named ‘the artificial orthodontic restraint,’ unfortunately not accounted for by CS Burstone, which completely nullifies the natural restraint by the alveolar bone and PDL, the Cres shifts toward the point of most significant restraint, namely toward the center of the bracket. The tooth rotates, so the root moves several times more than the crown in opposite directions. This actual torque movement definition can be measured only through cephalometric imaging before and after the torque application. All other measurements do not serve the profession. The immense amount of restraint and the anchorage level required during torque application was “mentioned” in the work of CS Case as early as 1895 (root movement machine) in a famous picture that I cannot upload to the forum. Where are these restraints and anchorage ‘hiding’ while using Clear Aligners? Sorry, but are we on the moon? Why have we become blind to basic biomechanic principles?
No removable appliance, especially clear aligners made of relatively thin plastic, can develop the needed anchorage and restraining forces on any tooth. Therefore, unfortunately, the article misleads its readers.
In conclusion, it’s time to change the existing orthodontic biomechanics theories. Sadly, something went wrong in tooth movement physics analysis, which opened the door for using and abusing it.
Whoever can control the Cres location will prevail.
PS:
1. The instantaneous center of rotation is at the Cres. In contrast, the total center of rotation lags behind all the Cres positions the tooth has gone through due to different forces, including growth in young individuals.
2. In orthodontics, translation can only be the outcome of the combined movement of a couple and tipping. There exists no other way to deliver translation in orthodontics.
3. Unfortunately, Burstone’s theory needs a complete revision because it allows for establishing unfounded facts.
4. Moment-to-force ratio is genuinely a meaningless term, since it does not own any contribution to the profession. Sorry.
Very much to the point. Biomechanics seems to be mostly forgotten and the less loved part of orthodontics…
In my understanding by reading the material and methods, it’s not clear which additional method other than power ridges (PR) has been used to try to effectively translate torque modification on the screen into real life.
My everyday experience with aligners has been similar to the findings in terms of torque expression with power ridges (they make it more likely the aligner to unfit) and without power ridges + without a compensating attachment.
The study thus studies the non expression of very likely inadequate tools to correct effectively torque.
Would be interesting to study the use of attachments, as a tool to keep adherence and express the couple of force system necessary to read torque. Would also be interesting to distinguish between crown torque (typical example CLII/2 central incisors) and root torque changes. Of course the latter is more unpredictable, while the first one is a very predictable one.
Dear Kevin,
In your post you mention “These figures reveal that the treatment only achieved 41.9% of the planned torque”. Did you actually calculate this percentage? I seem to get a different number…