May 10, 2021

Interproximal Reduction: Splitting hairs?

Many of us use interproximal reduction (IPR) daily as a conservative means of space creation. We all seem to have our way of doing this. However, few can claim to know how precise we are at undertaking the procedure.

Inter-proximal reduction is also often prescribed as part of aligner therapy producing conservative amounts of space to promote the achievement of occlusal objectives. But how accurately do we mirror our planning in respect of space creation with IPR? Do we generate excessive or inadequate space? In which regions of the mouth or tooth might we be more accurate? Are any of the bewildering array of IPR options more or less accurate?

A team from Milan looked at these questions. The Angle Orthodontist published this study. It is therefore freely available online.

Authors:Zamira Kalemaj and Luca Levrini

Angle Orthod. 2021. doi:10.2319/040920-272.1 

What did they do? 

They conducted a prospective cohort study involving 50 participants.


Patients undergoing Invisalign therapy in one or two arches. They were periodontally stable, treated non-extraction, and not undergoing dental restorative work. 


Six orthodontists treating between 5 and 10 participants each were involved. Two of these were very experienced, with four having ‘moderate’ experience. They used a range of mechanical approaches to interproximal reduction, including A combination of manual strips and burs, burs only, and contra-angle mounted strips. Four of the practitioners used measuring gauges. They obtained impressions at the beginning of the treatment and after the first set of aligners. The latter impressions may have corresponded either with the end of active treatment or the refinement stage.

Primary outcome: 

Correlation between the interproximal planned and performed based on the mesiodistal widths of the teeth on the consecutive digital models

Secondary outcomes:

The effect of the approach to IPR, use of gauge for clinical measurement of IPR, type of tooth, and IPR timing on the accuracy of the procedure

They did a sample size calculation, which showed that 50 participants were required overall to demonstrate a difference of 0.15mm between planned and performed IPR at a single location.

What did they find?

They planned IPR in the maxillary arch in 43 patients (Mean: 0.25 mm) and in the mandibular arch in 38 patients (Mean: 0.28 mm). They did the IPR both before (24 cases) and following (26 cases) alignment.

The practitioners were conservative in their removal of enamel. On average, the mean difference between planned and implemented IPR was 0.14mm. This difference represents around 50% of the planned amount. The use of burs was associated with a minor discrepancy while measuring gauges also helped somewhat. They did even less removal on the distal surfaces of the teeth than on the mesial. Finally, prior alignment followed by IPR (round-tripping) was associated with more precise IPR than undertaking IPR at the outset.

What did I think?

I thought that this was a fascinating study. I particularly liked it because it was simple and ‘didn’t try to do too much. The authors posed a fundamental question that is undoubtedly worthy of answering- how accurately do we perform IPR? The results are interesting and instructive- we may be less good than we think, and some approaches might help us to improve. The authors wrote the paper very nicely, with a clear description of the study’s rationale and a detailed interpretation of the results.

In terms of design, the authors carried out a prospective cohort study. This allowed for longitudinal assessment of planned versus performed IPR. Moreover, they included several practitioners and a range of techniques. As such, there are some confounding factors, including operator, the mechanical approach used, the timing of IPR, and the use of gauges. Arguably, however, this is also a strength of the study. It gives us an indication of the relative merits of each approach and indicates that experience may not be the discriminator that we reassure ourselves it is.

The magnitude of IPR overall was minor, and one would have to question how sensitive ClinCheck is to these minor differences. Nevertheless, they found that the repeatability of the measurements was acceptable.

The key message is that we may perform less IPR than we initially plan. This, of course, may be a lesser evil than performing too much. The use of burs and gauges may help to minimise the discrepancy between planned and executed IPR. It would be interesting to see how this shortcoming influences the progression of aligner therapy and, indeed, the outcome of treatment. This could be assessed in greater detail. However, I can’t help feeling that doing so would have muddied the waters a little and detracted from the simple message.

The finding that IPR is more accurate following the prior alignment of contact points is intuitive. It is worth highlighting that this finding is based on a limited sample. We should counterbalance this possible benefit with the relative merits of postponing IPR from clinical and patient management viewpoints. 

What can we conclude?

Given the relatively small sample and range of variables, we cannot make concrete inferences. However, it should certainly help us reflect on our practice and evaluate a little more critically our performance of interproximal reduction. It appears that we may undertake much less IPR than we plan to. Our precision may be influenced by the technique that we deploy, and gauges may help us carry out IPR more accurately. Simple conclusions from a simple and relevant study.

Have your say!

  1. Avatar

    Pad, there was a previous relatively similar attempt (DOI: 10.1186/s40510-020-00329-1) published in Progress in Orthodontics last year that found similar results. The authors concluded that less IPR was done than planned (-0.45mm/45% maxilla & -0.53mm/37% mandible). Differences in mm not likely clinically relevant. And just last month another Italian group published another similar attempt in BMC Oral Health (DOI: 10.1186/s12903-021-01487-2). This one actually argued that we are precise in the amount of IPR we do. Differences likely in the methodology of their measurement approach. Nevertheless, the main point is that we do not seem to overdo it. This is an important consideration.

  2. Avatar

    Great share Kevin! Perhaps I am a slow learner — but I have realized that a contributing factor to achieving less IPR than planned is that I/we do not account for the subtle movement of the tooth in the pdl space with the pressure created by whatever means of IPR we use. Such that when we are reducing one inter-proximal surface the abuting tooth/teeth move into the IPR space we created on the opposite side of the tooth. Whether measuring or not — evaluate how the IPR space you created adjusts as you move through the segment — and either be a bit more aggressive to account for that subtle shifting and/or go back through those IPR space you thought you made.

  3. Avatar

    I have to admit to being saddened and disappointed that IPR still happens! I also admit that I used IPR several times many years ago before I learnt there is a better way.

    IPR is a crude dental solution to a fundamentally skeletal problem, therefore by definition an erroneous way of addressing the problem. Crowding and the resultant imbrication of the teeth is NOT generally due to macrodontia (except in 0.03% of the population).

    Dental crowding is due to insufficient space due to inadequate growth for all the teeth to fit in correctly. The intelligent answer is to create space by increasing arch length/size by encouraging the arches to grow, as opposed to the destruction of healthy tooth tissue by enamel stripping (or extractions).

    • Avatar

      Utter nonsense, especially in the mandible. Bone can’t grow interstitially.

      • Avatar

        Waiting for a reply from Noel Stimson leaves me nearly, but not quite wild with anticipation

      • Avatar

        Why is it that when growth and development of the jaws are mentioned most orthodontists respond with “you cannot grow the mandible”? I am aware that that may be true, and I never suggested in my blog that we can make mandibles grow, although it is possible in young children. Adults are a different matter with maybe a few millimetres of change if we get lucky.
        However, we do know that there are parts of the mandible, other than the basal bone, that can change. For a start, the alveolar process is relatively plastic and can be moved radially forwards and buccally over the basal bone, to provide a broader and more expansive base for the lower teeth with an increase in arch length. This effect is often criticised by the ‘nay sayers’ as “all you have achieved is dento-alveolar change”. That is probably correct, but maybe that problem malocclusion was the result of the reverse, caused by imbalances in the lingual, oral and facial muscles? In any case, it works!
        The other mandibular components that can be grossly underestimated are the condyles. Under the effects of malocclusion and dysfunction, the condyles can be distorted by bending and wear and tear. The condylar discs may even be displaced. Appropriate functional therapy can decompress the joints and allow the condyles to straighten up, thereby helping to move the mandible downwards and forwards slightly towards a Class I position. These two changes combined can bring about considerable improvement. However, one must bear in mind that a displaced disc(s) should be addressed separately and before any orthodontic treatment. If successful, this treatment alone can cause considerable dimensional changes in the joint(s) with concomitant effects on the occlusion.
        The poor maxilla is so often is left out of the argument when it is patently obvious to those of us who work this way that it is a bone that, within certain limits, can be readily and positively changed in size and shape, such that when adequate upper arch form and width are achieved, an entrapped Class II mandible will often spontaneously drop forward into Class I without any further treatment.
        Obviously, appropriate maxillary growth and development will also enhance the nasal airway by widening the nasal floor.
        In my view, the point to bear in mind is that the teeth are merely passengers in dentofacial growth; they have no ability to change or move in themselves, unlike bone. There are 2 main factors involved in the final positions of the teeth – facial and oral muscles including the tongue, and cranio-facial development. Occlusion, and malocclusion, are a result of these 2 factors. If they are in harmony and correctly developed, good occlusion follows. If not, treating the malocclusion purely dentally is missing the point.

      • Avatar

        You took the words right out of my mouth. But then again, you put them there.

    • Avatar

      Destruction is a bit overkill. I agree with everything else although let’s also make the distinction between growing and non-growing individuals.

    • Avatar

      If someone can tell me how to get my adult patients to grow I will be delighted and excited! My oldest patient at the moment is 72.

  4. Avatar

    I agree, Dr Pickard. Unlike astrophysicists, unfortunately, orthodontists cannot “create” space; a phenomenon that only occurs at the edge of an expanding universe. Mere mortals can simply displace space, as noted by Archimedes. Moreover, enamel thickness has evolved over millennia, is unique in modern humans, and is often associated with systemic health detriments when it is lost prematurely (e.g. sleep bruxism).

  5. Avatar

    How soon we forget our professional history. The Angle – Case debate must have been over 100 years ago. The 2 arguments lacking what we might consider academic rigour but divisive nonetheless. Begg’s treatise from the 1950’s suggests interproximal enamel wear was a fact of living and his samples suggest much greater loss than today’s IPR might harvest. Little’s work suggests the need for continuing discussion. But let’s not get bogged down in the rubbish espoused by the ‘mandible growers’.

  6. Avatar

    Thx Kevin. I wonder if the clinicians were blinded? I may have missed a mention of this in the paper. If clinicians were aware that the accuracy of their clinical procedure was being assessed, this may influence the results. I am impressed by the accuracy reported. Considering that 12 of the 50 subjects had measurements based on manual models rather than digital models, the accuracy is impressive. Personally would like to see either all digital models or all manual, not a combination as the results may in fact reflect the source of the model. Also, did the clinicians program any IPR with the intention of not performing , or underperforming it clinically? I often do this in my Invisalign ClinCheck programming (extraction site closure, open posterior contacts), and additionally u aim to underperform clinically by 0.1mm, as I believe that I do not have the manual ability to reproduce 0.1mm exactly each time, and as these IPR numbers are generated on the false assumption that we will achieve clinically 100% of every movement programmed. All of these factors will also influence interpretation of results. The IPR numbers reported via our ClinCheck Software are also rounded from 3 decimal places of TREAT software measurements. These small but inevitable engineering plus significant patient variables – biologic and physiologic – tell me that space will likely be lacking in these areas; however better to err on the side of slightly under performing and continually assess?
    Another curious detail you mentioned was that more IPR was performed on mesial interproximal surface; I was under the impression that generally enamel was thicker on distal interproximal areas. As a result, when I program and perform IPR (sorry Dr Stimson), I favor the distal side. Interestingly the authors made another assumption that IPR was performed 50:50 on each tooth surface. I rarely do this, for the reason above, that I favor what I believe to be the larger tooth, or surface with more enamel, and take into consideration anatomy of the surfaces. Wonder if the clinicains were instructed to do this, otherwise a significant potential source of error.
    As to your question Kevin regarding sensitivity of ClinCheck (really TREAT) to IPR programmed, in my mind it would be 99% as it is a digital model generated with high accuracy (when accepted scanner is used) and we can see the effect that adding and reducing IPR has on the tooth position (actually fun to play with, you can see requirement for space change using tooth movements table as you intrude or move teeth, thinking about Germane, Woods et. al.). I would also love to see the next step, to determine the clinical tolerance range for discrepancy between dental vs aligner arch length; when IPR is underperformed or to me more specifically when space is lacking due to failure to expand or procline or distalize at the 1:1 ratio programmed but not clinically expressed. At what point do teeth begin to intrude or fail to align and occlude as fully as they may have when space was available? It is not easy to answer, as all programmed adjacent movements will impact the result.
    Ultimately it is worth reminding ourselves that enamel only comes once, and that clinicians , not default algorithms, should be programming amount and location of IPR.

Leave a Reply

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