|
A Randomized Prospective Trial Comparing Two Retraction techniques
Introduction:
In many malocclusions, the goals of therapy can be achieved without extraction in the permanent dentition. But in some consequential percentage of cases, almost all orthodontists now agree that the goals of treatment cannot be achieved satisfactorily without the extraction of some permanent teeth. Extraction therapy is frequently indicated to correct severe crowding, to retract the anterior teeth, to correct molar mal-relationships, and/or to modify the facial profile [1-3]. In a large percentage of these cases, maxillary anchorage control is a consequential problem in orthodontic treatment.
The most common mechanism for making retraction space available involves the extraction of one premolar in each quadrant. In order to retract the anterior teeth into the extraction space, most treatment strategies involve attaching the anterior teeth to some structure posterior to them. The only structures available for this purpose (prior to the recent development of TAD’s) have been the upper and lower molars. But forces applied to the molars to retract the anterior teeth tend to displace the molars forward into the extraction space. This forward displacement is called “anchorage loss” and its prevention is called “anchorage control”. In the lower dentition, anchorage loss is usually not a major problem because the lower molars are generally fairly resistant to mesial displacement. But in the maxillary dentition, consequential mesial displacement of the upper first molar occurs more readily and the problem can become severe. This is especially true in the treatment of Class II malocclusions.
Most techniques for retracting the anterior dentition involve preliminary bonding and leveling procedures soon after bicuspid extraction. After leveling has been completed, there are two general approaches to the problem of retracting anteriors with minimal mesial displacement of the upper first molar.
The most common approach is a sequential procedure in which the canines and incisors are retracted in two separate and distinct steps. In the first step, the canine in each quadrant is retracted to full contact with the tooth distal to the extraction space. In the second step, the canines are fastened to the teeth distal to them. The resulting grouping is then used as a single anchorage unit to retract the incisors. This procedure has been called the “two-step” technique [4-6].
In retracting the canine separately in the first step without adding the additional force that would be required to move the incisors at the same time, the advocates of the two-step approach assume that the load on the posterior teeth is reduced, thus reducing the tendency of the upper molars to displace forward. In the second step, the posterior segments, now buttressed by the incorporation of the canines, are pitted against the reduced resistance of the incisors alone. [7]
However, there are some conceivable disadvantages to the two-step approach. Closing space in two steps rather than one may make treatment take longer. Also, when canines are retracted individually, they may tend to tip and rotate more than when the anterior teeth are retracted as a single unit, thus requiring additional time and effort to re-level and re-align.
Therefore, an alternative approach called “en masse retraction” has come into use in which the incisors and canines are retracted as a single unit. One therapeutic technique that uses this approach is the MBT system developed by Bennett, McLaughlin and Trevisi [8-9]. This en masse technique has recently gained popularity because of its mechanical simplicity. But in theory, it might be expected to tax the posterior anchorage more than the two-step technique.
A randomized controlled clinical trial (RCT) was conducted to test the relative effectiveness of these two retraction techniques under actual clinical conditions. For this purpose, 64 maximum anchorage patents were randomized to treatment by eight experienced full time clinical instructors, each of whom was experienced in the use of both retraction techniques. Each clinician treated the same number of patients. Differences in mesial displacement of the upper first molar were measured between the two treatment techniques, between males and females, and between patients who started treatment at different stages of growth. Other parameters of interest were also investigated and will be reported at a later time. The trial was designed with the assistance of Dr. Edward L. Korn of the Biometric Research Branch, National Cancer Institute. It was conducted at the Orthodontic Clinic of the Department of Orthodontics, Peking University School of Stomatology in consultation with colleagues at the Craniofacial Research Instrumentation Laboratory of the Arthur A Dugoni School of Dentistry, University of the Pacific.
The primary purpose of the study was to test whether there were statistically significant between-treatment differences in mesial displacement of the upper first molar when maximum anchorage patients were randomized to both kinds of treatment. Our secondary purpose was to test the feasibility of performing an orthodontic RCT in this distributed setting.
Materials and Methods
Prior to beginning the study, a power analysis was performed by Dr. Korn to determine the required sample size. This analysis was based on the known variability of upper molar mesial displacement of the upper first molar as measured in previous extraction studies at UOP and the University of Medicine and Dentistry of New Jersey. It was determined that a sample size of 32 patients per group would be sufficient to detect a true mean difference of 1.75 mm between techniques at the p<0.05 level (two-sided) with 80% power. Therefore, the study was begun with the recruitment of a random sample consisting of 64 maximum anchorage patients.
Potential patients were identified during their initial visits to the Departmental clinic. Criteria for inclusion were that each selected patient:
1) Had a Class I or Class II malocclusion whose treatment required maximum anchorage control.
2) Had erupted permanent canines and no missing permanent teeth.
3) Had not yet reached his or her 16th birthday.
4) Was in good health with no chronic disease or disability.
A preliminary decision that each patient met these criteria was made by a single full-time project screener, an orthodontist with 8 years of clinical experience. A stratified bloc randomization was then used to assure the separate samples for the two treatment techniques were well balanced for sex, Angle class, and starting age. Following randomization to treatment technique, each patient was randomly assigned to care by one member of a panel of eight clinicians. Each member of the panel was experienced in the use of both the en masse and the two-step technique. However before assignment was confirmed, it was further required that the clinician to whose care each individual patient was assigned:
5) Agreed with the project screener that the patient required maximum anchorage control.
6) Agreed that it was appropriate to treat the patient using the treatment technique to which the patient had been randomized.
The last two requirements were included in order to meet the ethical and therapeutically important condition that no clinician be asked to treat a patient using a technique that he or she considered inappropriate for the care of that particular patient. [10]
Stratified randomization assured that the sub-samples for the two treatment techniques were well balanced for sex, Angle class, starting age, and pre-treatment crowding. (See Table I.)
Except for the requirement to use the prescribed retraction technique and some form of headgear, all treatment planning decisions for each patient were made by the treating clinician. These decisions explicitly included the type of headgear to be employed, the choice of extraction pattern, and whether or not to use a trans-palatal appliance (TPA) for additional anchorage support. Latitude in making all treatment decisions for each individual patient was delegated to the treating clinician because these decisions were considered to be part of the clinician’s unique treatment plan for that individual patient.
The samples for both techniques were treated with appliances of the same type (3M Unitek MBTTM prescription, 0.022×0.028-inch bracket slot) [9]. The treatment protocols of the two samples differed solely as follows: In the en masse sample, the canines were retracted with lace-back until crowding was eliminated and class I canine relationship was attained, after which the remaining extraction space was closed by retracting the 6 anterior teeth as a single unit. In the two-step sample, the canines were retracted first by lace back until they contacted the second premolars. The 4 incisors were then retracted using sliding mechanics.
All measurements of tooth displacement reported in this paper reflect differences in tooth position between beginning-of-treatment and end-of-treatment lateral cephalograms superimposed on the anatomical structures of the hard palate and the anterior maxillary process. (See Figure 3.) Displacements of the maxillary incisors and molars were measured parallel and perpendicular to the pretreatment Downs occlusal plane. All measurements are the averages of replicate independently performed landmark location and tracing superimposition operations by blinded and calibrated investigators using previously reported computer-assisted techniques developed at CRIL [11,12].
Figure 1: Measurement method. The T1 and T2 cephalograms are best fit on maxillary structures with primary attention given to the alignment of palatal structures and the anterior surface of the maxilla. Horizontal, vertical, and tipping displacements of the first molar and the central incisor were measured parallel and perpendicular to the Downs occlusal plane of the pretreatment image. The pre-treatment and end-of-treatment positions of Point A and the upper incisor and first molar cusps and apices are averages of the actual values for the 63 patients but the lines are interpolated and not necessarily to scale.
Reflection will make it apparent that what is being measured here is not “anchorage loss” per se. Rather it is the total mesial displacement of the upper first molar between the start of treatment and the time the second cephalogram is generated at the end of fixed appliance therapy. Hence our measurement of molar displacement includes not only treatment-associated changes during the periods of active space closure and subsequent finishing procedures but also the effects of inter-current growth changes throughout the active treatment period. A more precise measurement of anchorage loss during active space closure itself could have been obtained by generating an additional comparison cephalogram at the precise time that space closure was deemed complete. But this would have been technically difficult and of only academic value. For the purposes of the present investigation, total mesial displacement of the upper first molar has been used as an excellent surrogate measurement for “anchorage loss”. But beyond its surrogate role, we believe that total mesial displacement of the molar during treatment is the variable of greater interest to the clinical practitioner.
Statistical analysis:
The primary hypothesis investigated in this study was the belief of most orthodontic clinicians that on average there would be significantly less mesial displacement of the upper first molar during two-step retraction than during en masse retraction. This hypothesis was tested in the null form, i.e., that there would be no statistically significant difference in mesial displacement of the upper first molar between a sample of patients treated using the “en masse” technique and an equivalent sample treated using the two-step technique. The frequency, correlation and t-test procedures of the SAS statistical package (version 9.2, SAS Institute, Cary, NC) were used to analyze the data.
Results:
Findings for upper molar displacement during space closure are reported in Table II. It may be seen that relative to superimposition on palatal structures, no statistically significant difference in mesial displacement of the upper first molar was observed between the two retraction techniques. For the two treatment groups pooled, mean mesial displacement of the upper first molar at the molar cusp was slightly more than 4.3±2.1 mm with the apex moving forward 2.7±1.7mm. Indeed, in this particular study, mean mesial displacement of the upper first molar with respect to superimposition on palatal structures was slightly greater in the two-step sample than in the en masse sample, the best estimate of the average difference between the techniques being slightly less than 0.4 mm.
Extrusion of the upper molar, measured as the distance from the superimposed palatal structures to the molar cusp, was also similar for the two treatment groups. For the pooled sample of 63, its mean value increased an average of 2.1±1.5 mm, the average difference between the two treatment groups being less than 0.3 mm. At the molar apex, downward displacement was slightly less, the pooled average value being 1.8±1.4 mm with a mean between-treatment difference of 0.1 mm.
To form a more complete picture of maxillary dental changes during the process of space closure, corresponding statistics on incisor retraction and extrusion are included in the table. Retraction at the incisal edge was also extremely similar in the two samples, averaging 5.7±2.2 mm for the two samples pooled with a between-treatment mean difference of only 0.1 mm. The findings for retraction at the incisor apex were less anticipated and were indeed surprising. Essentially no retraction of the incisor apex was detected in either treatment group (mean = -0.1±17mm).
Extrusion at the incisal edge of the upper central incisor was greater in the en masse sample than in the two-step sample. The mean difference of 0.9 mm between the two treatments was statistically significant (p<0.03) but probably of little clinical importance. Part of the difference may be associated with the slightly greater lingual crown tipping observed in the en masse group.
As a further control, we checked for between-sample differences in incisor root resorption and found that the mean difference was less than 0.2 mm. We also checked the mean difference in treatment time between the two treatment samples and found it to be less than 1.3 months.
In evaluating these findings we investigated the possibility that there might have been consequential differences other than retraction technique between the en masse and two-step samples. Such differences could have resulted either from chance distortions in the randomization process or from systematic differences in the way treatment was delivered in the two samples. The second of these two possibilities is particularly important in experimental designs such as this one in which, following randomization, each clinician is asked to make all treatment decisions entirely based on his/her judgment of what treatment plan is best for each particular patient. In the present study it therefore became desirable to know whether the differences in the clinicians’ in-course treatment decisions for individual patients concerning headgear type, TPA use, and extraction pattern were balanced between the two subsamples. Table III provides answers to this question.
The tabulated distributions for headgear type, use or non-use of trans-palatal appliances and choice of extraction pattern were quite similar between techniques. The use of trans-palatal appliances seemed slightly more conservative in the en masse sample and the choice of extraction pattern seemed slightly more conservative in the two-step sample but neither finding was statistically significant.
Early in the analysis of the data from the study, an extensive cephalometric comparison of the pretreatment state of two treatment subsamples was made. This comparison examined 35 conventional cephalometric measurements. Results of the comparison are summarized in Table IV without correction for multiple comparisons.
Assuming independence among the 35 statistical tests, one would have expected by the nature of the definition of statistical significance that on average ~1.7 tests would be reported to be statistically significant at the p<0.05 level by chance alone. In this investigation, two of the 35 tests were found to be significant at that level. These variables were Condyle-Pogonion Distance, a measure of mandibular size, and Condyle-Point A distance, a measure of maxillary length.
These findings raised the question of whether differences in these two dimensions were actually associated with differences in upper first molar displacement. We were able to answer this question directly by testing the association between displacement of the upper first molar and each of these two variables. For the merged sample (n=63) and for the two-step sample considered separately, neither of these relationships was statistically significant, but within the en mass subsample considered separately there was a statistically significant increased tendency for the patients with longer original Condyle-pogonion distances to have less mesial displacement of the upper first molar. (See Table V.)
Associations between Pretreatment Demographics, Clinicians’ Treatment Designs, and Mesial Displacement of the Upper First Molar:
The data collected in the course of the study made it also possible to conduct retrospective tests aimed at the detection of differences in mesial displacement of the upper first molar associated with differences in Sex, Angle Class, extraction pattern, head gear type, TPA use, starting age, and crowding at the beginning of treatment. The results of such tests are summarized in Table VI.
It may be seen in this table that when the patients in the two treatment groups were pooled, each of the major demographic variables identified prior to treatment was associated with a statistically significant difference in displacement of the upper first molar during treatment. Boys had significantly more mesial displacement of the first molar than girls (mean = 1.2 mm, p<0.02); patients starting treatment below 13 years of age had significantly more mesial displacement of the first molar than patients starting treatment after 13 years of age (mean = 1.2 mm, p<.04). When sex differences concerning age at puberty were taken into consideration by grouping girls over 12 with boys over 14 and testing them against a group consisting of girls less than 12 and boys less than 14, the younger group again showed greater mesial displacement than the older group (mean = 1.2 mm, p<0.02). Mean differences in displacement of the upper first molar associated with clinician choice to use cervical or high-pull head gear averaged no more than 0.5 mm and were not statistically significant. Patients who wore TPA appliances averaged 0.6 mm less mesial displacement of the upper first molar than those who did not but this difference also fell substantially short of statistical significance. On average, patients who had extractions of upper first and lower second bicuspids had 1.1 mm less mesial displacement of the upper first molar than those treated with the removal of four upper first bicuspids (p<0.06).
In evaluating these differences between category for the pooled sample, it should be remembered that the samples for the two treatments are well balanced within sample with respect to each of the variables in Table VI.
Discussion:
The findings summarized in Table II, demonstrate fairly conclusively that there was no statistically significant difference in mesial displacement of the upper first molar between the two samples under examination. At the outset of this study, none of the investigators considered the possibility that en masse retraction might result in less mesial molar displacement than two-step retraction since conventional wisdom held that en masse retraction would result in more mesial molar displacement. But in this study the en masse retraction sample actually experienced slightly less mesial molar displacement on average than did the two-step sample with essentially no mean difference in treatment time. While this difference in molar displacement was too small to be statistically significant given the observed intra-technique variability, the data are inconsistent with mean mesial displacement of the upper first molar in en masse treated patients being greater than that for two-step treated patients by an amount greater than 0.71 mm at the 95% level of confidence. Therefore we assert with high confidence that average mesial displacement of the upper first molar from en masse treatment is unlikely to be smaller than that for the two-step treatment in patient populations similar to the one from which the present sample was drawn.
Our study design had been stratified so as to produce parity between the en masse and two-step samples with respect to starting age, sex/gender and maturation level. This made it possible to investigate differences in mesial molar displacement with respect to these important demographic variables. No consequential differences between the en masse and two-step samples were found with respect to any of these variables. However, when the en masse and two-step treatment groups were pooled, mean mesial displacement of the upper molar for all boys in the study was statistically significantly greater than mean mesial displacement for all girls. Also for the pooled samples, patients of both genders who started treatment before age 13 had significantly more mesial displacement of the upper first molar than patients who started treatment at ages greater than 13, the mean difference being 1.2 mm (p<0.04). When an adjustment for the between-gender differences in maturation rate was made by comparing girls under 12 and boys under 14 in one group with girls over 12 and boys over 14 in another group, the findings were substantially the same (mean difference = 1.2 mm, p<0.02).
The orthodontic literature contains no comparable data from prospective RCTs on mesial displacement of the upper first molar but a retrospective Korean study of anchorage loss in Class I adult female patients yielded results not inconsistent with our own [13]. In another recent retrospective study, McKinney and Harris [14] reported on differences in anchorage loss between boys and girls treated with Begg, Edgewise and Straightwire appliances; between sex differences were smaller than those we report but not dramatically so when differences in sampling and measuring technique are considered.
In our study, small differences in starting age were surprisingly strongly associated with differences in mesial displacement of the upper first molar, greater mesial displacement being observed in younger children. This finding is consistent with those of other investigators; in recent years three different groups of investigators using retrospective samples of differing character have each reported higher mean values for mesial displacement of the upper first molar for younger subjects than for more mature subjects [14-16].
As part of an effort to check for possible over-all dissimilarities in treatment outcome between the en masse and two-step groups, we also examined changes in the incisor region. Change in the axial inclination of the upper incisor was similar in the two treatment groups. Root resorption was also similar and in the range reported in previous retrospective studies [17, 18]. A statistically significant mean difference of 0.9 mm in incisor extrusion was observed between techniques but was considered too small to be of clinical importance. Perhaps the most interesting finding in the incisor region was the minimal apical displacement observed in both samples. Relative to best fit of maxillary structures, mean displacement of the incisor apex was considerably less than 0.2 mm in either the vertical or horizontal direction in either sample. We conclude that in the samples for both techniques retraction of the incisor was accomplished almost entirely by controlled tipping with an average center of rotation very close to the apex.
Some early participants in the design of this study believed that we should rigidly prescribe the conditions under which the two techniques under investigation were to be used. Thus for example, the same type of head gear would be used for the same length of time in conjunction with the same use of TPA’s and the same extraction pattern. (Examples of this kind of strategy may be seen in recent Class II correction studies at North Carolina and Florida [19, 20].) Pursuing this course of action might well have reduced the variability of outcome within each treatment sample in our own study, making it possible to identify more statistically significant differences. But by the same token it would have made the results applicable only to the very small set of clinical cases treated in the same rigid manner, thus very much reducing the applicability of the findings to most actual clinical treatment.
For this reason, once a patient was assigned to a particular retraction technique, all of the treating clinician’s decisions concerning treatment planning and execution (including head gear type, TPA use, and extraction pattern) were considered part of the way in which that clinician used that technique in the treatment of that individual patient. We did measure differences in each of these variables between techniques but we consciously chose to treat clinicians’ choices regarding them as part of each clinician’s application of each technique to the treatment of each patient. Had we controlled the precise conditions of treatment more rigidly, we believe that the results of this study would have been rendered far less applicable to the actual conditions of clinical treatment as it is customarily delivered. In this sense, the present study is aimed at providing information on the “effectiveness” of each treatment as it is actually used under actual clinical conditions rather than the “efficacy” of the technique under an overly-rigidified design that adds constraints foreign to actual clinical practice in order to simplify the process of data acquisition and analysis.
It also seems noteworthy that no differences associated with mesial displacement of the upper first molar were found for several variables where conventional wisdom might have caused us to expect them. There were no statistically significant differences in molar displacement between Angle Class I and Class II patients, or among patients whose crowding was more severe before treatment. No significant differences in mesial molar displacement were noted between patients who had cervical or high-pull headgear, and no significant differences were found between patients treated with or without trans-palatal appliances. Patients in this sample who were treated with the extraction of upper first and lower second bicuspids did have less mesial displacement of the upper first molar than patients treated with the removal of all four first bicuspids, but this difference fell just short of statistical significance (p<0.06).
Generalizability and Limitations:
We have presented data on mesial molar displacement from an RCT comparing two techniques of space closure with samples of moderate size. We believe that the conclusions of this study are likely to be generalizable to those other techniques of two-step retraction that 1) use headgear and 2) incorporate the incisors into the arch before starting retraction of the canines. Our findings are less likely to be generalizable to two-step retraction systems that employ segmental retraction schemes in which the incisors are not tied into the arch until the canines have been retracted.
Conclusions:
In this prospective randomized trial, a group of patients treated using an en masse retraction technique experienced slightly less mesial molar displacement on average than was observed in a comparable group treated using the more conventionally used two-step technique of retraction. The observed difference was not large enough to be statistically significant from zero. Hence it would be inappropriate to assert that the en masse technique consistently produces less mesial displacement of the upper first molar on average than does the two-step technique. But at the same time, the findings imply that the assertion that the en masse technique produces substantially more mesial displacement of the upper first molar than the two-step technique is very unlikely to be true.
In extrapolating on these findings, it is important to note that headgear was used by all patients in both treatment groups and that auxiliary anchorage (TPA) was used in the majority of cases in both treatment groups. Despite these precautions, average mesial molar displacement from anchorage loss and growth approximated or exceeded half the width of the crown of the extracted bicuspid teeth in both treatment groups.
When both treatment groups were pooled, small but statistically significant sex and starting age differences in the magnitude of molar displacement were detected. Boys had more mesial displacement than girls and children starting treatment earlier had more mesial displacement than children starting treatment slightly later.
|
