Comparison between two screws for maxillary expansion: a multicenter randomized controlled trial on patient’s reported outcome measures

Summary

Objective

The objective of this study was to compare an expansion screw of the maxilla that generates moderate and continuous forces versus a conventional screw for rapid maxillary expansion (RME) on patient-reported outcome measure during the first 12 weeks of treatment.

Trial design

This is a superiority, multicenter, two-arm parallel balanced randomization trial.

Methods

Patients in the mixed dentition were included with a transverse discrepancy between the two arches of at least 3 mm. An expansion screw using moderate continuous forces (Leaf group, treated with an expansion screw with Ni–Ti springs) was compared to a conventional RME screw that generates intermittent heavy forces (RME group). The primary response variable was the visual analogue scale (VAS) on pain calculated in the first 12 weeks of therapy. The VAS on difficulty on speaking and oral hygiene, patient satisfaction, and complications were also evaluated. A computer-generated block randomization was used with allocation concealed in sequentially numbered opaque-sealed envelopes. Blinding was not applicable. Linear models were used for statistical analysis.

Results

Twenty-eight patients in the Leaf group and 28 patients in the RME group were randomized and included in the study. There were no dropouts. The mean of the VAS for pain was 0.3 ± 0.4 in the Leaf group and 0.6 ± 0.5 in the RME group. The difference was −0.3 (95 per cent CI from −0.5 to −0.0; P = 0.017) in favour of the Leaf group. The difference in pain was marked in the first week (Leaf group 2.2 ± 2.3; RME group 3.7 ± 2.6; difference −1.5; 95 per cent CI from −2.7 to −0.3; P = 0.019).

Conclusions

Patients in the Leaf group experienced a lower degree of pain, especially during the first week following the application of the expander. For the other variables, no significant differences were reported between the two treatments.

Registration

The study was registered in the ISRCTN register on 8 November 2016 with the number ISRCTN18263886.

Introduction

Maxillary expansion is a very common clinical procedure to treat posterior crossbite and more generally maxillary transverse deficiency in growing patients (1–5). During rapid maxillary expansion (RME), there are sometimes side effects, such as pain, non-opening of the intermaxillary suture, and oral ulcerations (3, 6–9), especially when RME is performed in pubertal or post-pubertal periods. The perception of pain affects almost all patients; it is greater in the first 3 days of activation and it is normally described as mild or moderate (6, 10). Pain seems to be related to the inflammatory reaction generated during the opening of the midpalatal suture and in relation to the compression exerted on the periodontal ligament (11). Headaches and dizziness have rarely been described (12). A randomized trial showed a reduction in pain with the use of ketoprofen (10). Some studies have shown that pain could be decreased by changing the number of daily activations during palatal expansion (9, 12, 13).

Recently, an expansion device called Leaf expander has been introduced. The Leaf expander is similar to a conventional rapid palatal expander. However, instead of the conventional screw, it incorporates a nickel–titanium (Ni–Ti) leaf spring that resumes its shape after activation, resulting in a calibrated expansion of the maxilla (14). Generally, the Leaf expander is anchored to the deciduous teeth so that the first permanent molars expand spontaneously (14). Moreover, its activation is performed by the clinician and it does not require the cooperation of the patient’s parents. Ugolini et al. (15) performed a randomized clinical trial (RCT) in which patients treated with the Leaf expander reported significantly lower pain than patients treated with RME. In this RCT, however, only the first 7 days of treatment were evaluated.

The objective of this multicenter RCT was to compare the Leaf expander screw versus the conventional RME screw on patient-reported outcome measure (PROM) during the first 12 weeks of treatment. The present RCT follows the guidelines of Consort 2010 (16).

Materials and methods

Trial design

This is a superiority, two-centre, two-arm parallel balanced randomization trial. This study was conducted in the Orthodontic Clinics of two University Hospitals. The coordinating centre was the University of Florence, while the other centre was the University of Rome Tor Vergata. The study was registered in the ISRCTN register on 8 November 2016 with the ISRCTN18263886 number.

Participants

Patients had to be pre-pubertal (Cervical Stage [CS] 1 or 2 in cervical vertebral maturation) (17) and in the early or intermediate mixed dentition stage (18) with fully erupted upper and lower first permanent molars. To be included in the study, patients had to present with a posterior transverse interarch discrepancy (19) of at least 3 mm. Transverse interarch discrepancy was calculated on dental casts as the difference between the maxillary intermolar width (distance between the central fossae of right and left permanent first maxillary molars) and the mandibular intermolar width (distance between the tips of the distobuccal cusps of right and left permanent first mandibular molars).

Exclusion criteria were:

1. pubertal or post-pubertal stage of development (CS 3–6);

2. late deciduous or mixed dentition;

3. agenesis of upper second premolars (assessed on initial panoramic radiograph);

4. cleft lip and/or palate and craniofacial syndromes;

5. patients unable to be followed for at least 1 year.

All patients’ parents signed an informed consent before starting the trial. The study was approved by the Pediatric Ethics Committee of Tuscany (No. 57/2016 of 28/04/2016) and by the Independent Ethics Committee of the University of Rome Tor Vergata (No. 6/16 of 15/01/2016).

Interventions

The design of the appliance for maxillary expansion was the butterfly expander (20). The expanders were constructed with the two different expansion screws with lateral arms soldered to the bands cemented on the upper second deciduous molars. The device under investigation was an expansion screw using continuous moderate forces (Leaf Expander—Leone SpA, Sesto Fiorentino, Firenze, Italy; Figure 1). The comparison device was a conventional RME screw using intermittent heavy forces (expansion screw with telescopic guides A2620—Leone SpA, Sesto Fiorentino, Firenze, Italy). Ten-millimetre screws were used in all patients (Figure 2).

As far as the Leaf expander is concerned, Ni–Ti screws that developed 900 g of force have been used. An initial expansion of 4.5 mm was achieved in about 2–3 months. Patients were checked every 2 weeks to monitor the deactivation of the spring. Reactivation was performed in the office by 10 quarter turns of the screw per month. It should be stressed that 1 quarter turn corresponds to 0.1 mm of activation (10 activations of the screw generated 1 mm of activation). Three clinicians treated the patients in the centre of Florence (centre #1), while two clinicians treated the patients in the centre of Rome (centre #2). The clinical experience was similar for all clinicians (5–10 years).

As far as the clinical management of the conventional RME screw, which represents the control, patient’s parents were instructed to activate the screw a quarter of a turn per day (which corresponded to 0.2 mm of expansion). The force developed was about 2.4 kg (21). Patients were checked every 2 weeks (during which approximately 3 mm of expansion was obtained) until the desired expansion was achieved.

Once the expander was applied, patients were given questionnaires to fill in at home and to bring back at follow-up visits, every 2 weeks, throughout the active expansion phase. For both protocols, the active expansion was ended when the palatal cusps of the upper second deciduous molars approximated the buccal cusps of the lower second deciduous molars. The expanders were then left in place as passive retainers and they were removed 1 year after the start of therapy. From the end of the active expansion phase until the removal of the expander, patients were visited once a month until the end of the study (1 year after the start of therapy).

The most frequent complications could be pain, ulcerations on the palatal mucosa from excessive pressure of the screw body or arms, and decementation of the expander. Pain was generally limited to temporary discomfort on the anchor teeth or the midpalatine suture when the screw or expander spring was activated and, therefore, no pharmacological intervention was performed. The treatment of ulceration at an early stage included, in addition to home hygiene manoeuvres and rinsing with sodium hyaluronate-based mouthwashes, the use of topical application of gel with 0.2 per cent chlorhexidine for 15 days. Patients had to inform the clinician immediately if they felt the mobility of the expander. If the expander decemented totally or partially, it had to be recemented as soon as possible. The recording of deviations from the clinical protocol was carried out by the investigators involved in each centre, who reported the type and nature of the deviation on the data collection forms.

Outcomes

The primary outcome of the study was pain measured on a visual analogue scale (VAS) in the first 12 weeks. Secondary outcomes were the difficulty in speaking and in expander hygiene during the first 12 weeks, patients’ and parents’ satisfaction at the end of the study, and complications.

Immediately after cementing the expander, patients were provided with a questionnaire, including a VAS for pain, difficulty in speaking, and difficulty in keeping the expander clean. Both patients and their parents were instructed on how to complete the questionnaire.

The VAS were recorded once a week for 12 weeks. Patients were instructed to estimate (retrospectively) an average level of pain/difficulty during the week.

The VAS ranged from 0 to 10. In the case of pain, 0 was ‘no pain’ and 10 was ‘worst imaginable pain’. The Wong–Baker Faces Pain Scale was also used to facilitate interpretation (22). In case of difficulty in speaking, 0 corresponded to ‘no difficulty’ and 10 to ‘the greatest imaginable difficulty’. In the case of the difficulty in expander hygiene, 0 corresponded to ‘no difficulty’ and 10 to the ‘the greatest imaginable difficulty’ in cleaning the expander. The satisfaction of the patients and the parents were also measured with a VAS in which 0 meant ‘maximum dissatisfaction’ and 10 ‘maximum satisfaction’ with the result achieved. Complications reported by patients and their parents, in particular, oral ulcerations and expander decementations, were also recorded.

Sample size

In order to highlight a difference between the two treatment modalities of 1.5 points on the VAS pain scale, considering a standard deviation (SD) of 1.87 (9), a two-tailed significance level of 5 per cent, and a power of 80 per cent, a sample size of 56 patients (28 per treatment) was required, taking into account a 10 per cent withdrawal rate from the study. Twenty-eight patients (14 per treatment) were enrolled in each centre.

Randomization and allocation concealment

A computer-generated random number list was used to allocate patients to treatments. Block randomization was used to assign the same number of patients to each treatment in the two centres.

The allocation sequence was concealed by the statistician (MN), who used opaque and sealed envelopes, sequentially numbered for each of the two centres. The envelopes were opened by the operators only when the expander was prepared for cementation.

Blinding

The clinicians could not be blinded about the treatment that they were providing. The patients and parents did not know whether the expander used was the Leaf or conventional RME expander. However, they could not be blinded because they could observe the expander that required different activation. Complications were recorded and treated by clinicians who were not blinded on the applied expander. The variables considered in this study could not be blinded since they referred to the questionnaire that was filled in by the patients and their parents who were not blinded on the treatment used.

Statistical methods

Descriptive statistics were performed using means and SDs for quantitative variables and frequency and percentage for qualitative variables. For the VAS, linear models were used considering the centre as a fixed effect. The explanatory variable was the treatment (Leaf versus RME expander) and the response variable was the average of the 12 measurements (one per week) of the VAS. The interaction between the treatment and the centre was considered in the models only if statistically significant.

Similar models were also prepared for the first-, second-, third- and fourth-week VAS and for the average of 5th to 12th week in order to show differences between treatments in a narrower time span. The presence of complications was analysed using logistic regressions, also considering the centre. The terms of interaction were added to the models only if significant.

For each statistical model, the estimate of the treatment effect (the estimate of the difference between the two treatments used), the P-value, and the 95 per cent confidence interval (CI) were provided. When the interaction was statistically significant, interaction graph was reported to analyse in more detail the different trend in the two centres.

Statistical analysis was carried out according to the intention-to-treat method. The unit of analysis was represented by the patient. A sensitivity non-parametric analysis (Mann–Whitney test) was performed for the pain VAS.

Results

Fifty-six patients were enrolled in the trial and randomly allocated to maxillary expansion with the two types of screw. Twenty-eight patients were allocated to the Leaf expander (Leaf group) and 28 patients were allocated to the conventional RME expander (RME group). Twenty-eight patients, 14 per group, were treated in the centre of Florence (centre #1), and 28 patients, 14 per group, were treated in the centre of Rome (centre #2).

The patients were recruited and treated in the two University Hospitals from October 2016 to November 2018. The last 12-month follow-up was carried out in November 2019.

The demographic characteristics of patients at the beginning of therapy are shown in Table 1. No statistically significant differences between the two groups and between the two centres for gender, cervical vertebral maturation (Fisher exact text), or age (independent sample t-test) were found. For the transverse interarch discrepancy, no statistically significant difference between the 2 groups was recorded while a statistically significant difference between the two centres was found (P = 0.001) with centre #1 showing a greater negative transverse interarch discrepancy (−1.3 mm).

All patients received the treatment assigned by randomization. There were no withdrawals from the trial and no deviations from protocol (Figure 3). Duration of active therapy was 4.5 ± 1.1 months in the Leaf group and 1.0 ± 0.4 months in the RME group (difference 3.5 months P <0.0001).

Pain

The results for the VAS for pain are presented in Table 2 and Figures 4 and 5. The mean of the VAS for pain was 0.3 ± 0.4 in the Leaf group and 0.6 ± 0.5 in the RME group. The difference was −0.3 (95 per cent CI = −0.5 to −0.0; P = 0.017) in favour of the Leaf group. The interaction between the centre and the treatment was not significant and there was no centre effect. Twenty-two patients in the Leaf group (79 per cent) and 24 patients in the RME group (86 per cent) reported pain.

The difference in pain was significantly different in the first week (Leaf group 2.2 ± 2.3; RME group 3.7 ± 2.6; difference −1.5, 95 per cent CI from −2.7 to −0.3; P = 0.019) in favour of the Leaf group, whose patients felt less pain. In the first week, a centre effect was also observed, with patients of centre #2 presenting a significantly greater pain (difference of 1.7 between the two centres, 95 per cent CI from 0.5 to 3.0, P = 0.008). The interaction between the centre and the treatment was not statistically significant.

In the second week, the interaction between the centre and the treatment was statistically significant (P = 0.025). This indicates that pain between the two groups was different in the two centres. In fact, while a significant difference in pain persisted in centre #2, it practically disappeared in centre #1 (Figure 6). In the other weeks, there were no significant differences between the two groups, between the two centres, and in the interaction group/centre. Pain decreased progressively over the 12 weeks after the start of treatment (Figure 4).

The sensitivity analysis using Mann–Whitney test showed similar results. In particular, for the mean VAS, P-value was 0.01 favouring Leaf group in the first and, in the second week, P-value was 0.03 favouring Leaf group, while, from the 3rd through the 12th week, it was not significant.

Difficulty in speaking

The results on the VAS on difficulty in speaking are presented in Table 3 and Figure 7. The average of the VAS over the 12 weeks for difficulty in speaking was 0.6 ± 0.7 in the Leaf group and 0.9 ± 0.9 in the RME group. The difference was −0.3 (95 per cent CI from −0.8 to 0.1; P = 0.132) in favour of the Leaf group. The interaction between the centre and the treatment was not significant and there was no centre effect.

In the first week, the interaction between the centre and the treatment was statistically significant (P = 0.002). This indicated that difficulty in speaking between the two groups was different in the two centres. In fact, in centre #2, there was a significantly greater difficulty in speaking in the RME group, while, in centre #1, the difficulty was greater in the Leaf group, although not statistically significant (Figure 8). In the second, third, and fourth weeks, there were no significant differences between the two groups, between the two centres, and in the interaction group/centre.

During Weeks 5–12, a significant centre effect was observed (P = 0.009). Centre #2 showed lower values on the VAS for difficulty in speaking (−0.4, 95 per cent CI from −0.8 to −0.1). The difficulty in speaking decreased progressively in the 12 weeks after the start of the treatment (Figure 7).

Difficulty in expander hygiene

The results for the VAS on difficulty in expander hygiene are presented in Table 4 and Figure 9. The average of the VAS over the 12 weeks for the difficulty in hygiene was 1.2 ± 1.3 in the Leaf group and 1.1 ± 1.0 in the RME group. The difference was 0.1 (95 per cent CI from −0.5 to 0.7; P = 0.802) in favour of the RME group. The interaction between the centre and the treatment was not significant and there was no centre effect.

In the first week, the interaction between the centre and the treatment was statistically significant (P = 0.0006). This indicates that the difference in hygiene difficulty between the two groups was different in the two centres. In fact, in centre #2, a significantly greater difficulty in expander hygiene was observed in the RME group, while, in centre #1, the difficulty was greater in the Leaf group, although not statistically significant (Figure 10). In the second, third, and fourth weeks, there were no significant differences between the two groups, between the two centres, and in the interaction group/centre.

During weeks 5–12, a centre effect was observed (P = 0.005). Centre #2 showed lower values on the VAS for the difficulty in expander hygiene: −0.8, 95 per cent CI from −1.4 to −0.2. The difficulty in expander hygiene progressively decreased in the 12 weeks after the beginning of the treatment and it settled around 1 (Figure 9).

Complications

The results of the complications are presented in Supplementary Table 1. One patient in the RME group had ulceration of the palatal mucosa and decementation.

Four patients in the Leaf group and eight patients in the RME group showed complications (odds ratio = 0.42, 95 per cent CI from 0.11 to 1.59, P = 0.200). No significant differences were found between the two groups, between the two centres, and in the interaction group/centre.

Satisfaction with the treatment

The results regarding satisfaction with the treatment performed are presented in Supplementary Table 2. No significant differences were found between the two groups, between the two centres, and in the interaction group/centre.

Discussion

The objective of this RCT was to compare the Leaf expander versus the conventional RME on PROM during the first 12 weeks of treatment. In particular, pain, difficulty in speaking, difficulty in expander hygiene, patient and parent satisfaction, and complications were analysed in this study.

Patients in the Leaf group experienced a lower degree of pain, especially in the first week following the application of the expander. For the other variables, no significant differences between the two groups were found. For many variables, differences between the two centres were observed.

Interpretation

The primary outcome variable of this study was the average pain of the first 12 weeks. Pain was lower in the Leaf group with a difference that was statistically significant though very small and probably not clinically significant (0.3 points on the VAS). This was due to the fact that, for both treatments, pain decreased significantly along with time and, from the fifth week, it was around 0.1 on a scale from 0 to 10. The difference in pain between the two groups was particularly noticeable only in the first week after application of the expander. In the first week, the difference on the VAS was 1.5 points in favour of the patients in the Leaf group who felt less pain than the patients in the RME group. This difference has some clinical relevance because it is also the difference that was set as clinically relevant for the calculation of the sample size. As for the second week, the difference in pain between the two treatments was more uncertain. In the second week, there was an interaction between the centre and the treatment. This indicates that the difference remained high only for centre #2 but not for centre #1 and, from the third week onwards, the difference in pain between the two groups was not statistically significant. The difference between the two centres in pain levels for the RME group during the second week is difficult to explain. The number of clinicians and their level of experience was similar in the two centres. Most probably, this difference can be ascribed to differences in perception of pain among patients in the two centres. These data are in line with the literature, which showed that the highest proportion of pain for RME was observed in the first few days after the application of the expander (6, 8–10, 12, 13, 15, 23).

In two studies, 98 per cent of patients reported pain after RME and the highest pain levels were found during the first six activations of the expansion screw, that is, in the first 3–6 days (6, 12). Subsequently, pain decreased rapidly. In addition, pain was related to the number of daily activations of the screw (two rather than one) (6). The same result was confirmed in another study (9). In the present RCT, in the RME group, the screw was activated only once per day and this would explain the slightly lower frequency of pain found compared to the Needleman et al. study (6).

It is still unclear whether pain is influenced by the patient’s gender and age. In fact, some studies have shown no differences in gender or age (6, 8, 12), while other studies have shown greater sensitivity to pain for girls (9, 13) and others have shown greater sensitivity to pain as age increases (9, 23). Ugolini et al. (15) performed a multicenter RCT that is very similar to ours. The major differences were that Ugolini et al. (15) evaluated pain and difficulty in speaking only during the first week. Moreover, the Leaf expander used in their RCT generated smaller forces (450 g). In the RME group, the screw was activated twice per day and child’s pain was measured 5 minutes after each turn of the screw. In addition, they did not evaluate differences between the two centres. The authors reported significantly greater levels of pain in the RME group during the first 4 days.

As for difficulty in speaking, in general no significant differences were obtained between the two treatments; however, a centre effect emerged that is difficult to explain. It could be due to small differences in the positioning of the expanders in the two centres or to differences at patients’ level. In general, the difficulty in speaking was maximum in the first week and decreased along with time, although it did not disappear completely even at the 12th week, reaching values of 0.2–0.4 on the VAS. Similar results were shown in another study (8). Ugolini et al. (15) investigated the difficulty in speaking in Leaf and RME groups during the first week and they also found no significant differences.

A similar trend to difficulty in speaking was also observed for the difficulty in the hygiene of the expander. Also, for this variable, in general, no significant differences were obtained between the two groups. A centre effect emerged whose cause could have the same interpretation as for difficulty in speaking. The difficulty in expander hygiene was maximum in the first week and decreased along with time, even if it remained far from 0 even at the 12th week, settling at values close to 1 on the VAS.

As for the complications, no differences emerged between the groups and between the centres, although probably the sample size does not allow to perform high-power statistical tests since the variable in question is dichotomous and the total number of complications that occurred was low. Patient and parent satisfaction were high for these procedures and there was no difference between groups and centres.

Limitations

A limitation of this study was the presence of several operators, which could have determined the high heterogeneity between the centres. On the other hand, however, this fact increases the possibility of extrapolating the results to the normal clinical routine. Moreover, the sample size may not be adequate for some variables, such as complications. In fact, it was calculated considering pain as the primary variable.

Conclusions

The Leaf expander decreases painful perception in the first week after the application of the expander in growing patients.

Funding

No funding or conflict of interest to be declared.

Conflict of interest

The authors declare that they have no conflict of interest

Data availability

Data are available on request

References