This article has Open Peer Review reports available.
Root resorption, treatment time and extraction rate during orthodontic treatment with self-ligating and conventional brackets
© Jacobs et al.; licensee BioMed Central Ltd. 2014
Received: 25 November 2013
Accepted: 6 January 2014
Published: 23 January 2014
This study determined the amount and severity of EARR (external apical root resorption) after orthodontic treatment with self-ligating (SL) and conventional (Non-SL) brackets. Differences regarding rate of extraction cases, appointments and treatment time were evaluated.
Material and methods
213 patients with a mean age of 12.4 ± 2.2 years were evaluated retrospectively. The treatments were performed with SL brackets (n = 139, Smartclip, 3 M Unitek, USA) or Non-SL brackets (n = 74, Victory Series, 3 M Unitek, USA). Measurements of the crown and root length of the incisors were taken using panoramic radiographs. Three-factor analysis of variance (ANOVA) was performed for an appliance effect.
There was no difference between patients treated with Non-SL or SL brackets regarding the amount (in percentage) of EARR (Non-SL: 4.5 ± 6.6 vs. SL: 3.0 ± 5.6). Occurrence of severe EARR (sEARR) did also not differ between the two groups (Non-SL 0.5 vs. SL: 0.3). The percentage of patients with need of tooth extraction for treatment (Non SL: 8.1 vs. SL: 6.9) and the number of appointments (Non-SL: 12.4 ± 3.4 vs. SL: 13.9 ± 3.3) did not show any differences. The treatment time was shorter with Non-SL brackets (Non-SL: 18.1 ± 5.3 vs. SL: 20.7 ± 4.9 months).
This is the largest study showing that there is no difference in the amount of EARR, number of appointments and extraction rate between conventional and self-ligating brackets. For the first time we could demonstrate that occurrence of sEARR does not differ between the two types of brackets.
External apical root resorption (EARR) is defined as either a physiologic or pathologic process with the loss of cementum or dentine resulting in a shortening of the root apex. This process is often associated with orthodontic treatment .
Since EARR is a serious iatrogenic problem, there has been intensive research about EARR as an adverse effect during orthodontic treatment. As a result many studies have underlined that EARR often develops during treatment with the fixed multibracket appliance [1–5]. According to this and the fact that mechanical forces are a key factor in the occurrence of EARR, studies have shown, that the appliance or technique used for an orthodontic treatment can be related to the degree of EARR [4, 6–8].
Furthermore several other factors have been implicated in the initiation and progression of EARR’s during orthodontic treatment, such as the duration of treatment [9–11], the level of force applied [12, 13], idiopathic EARR before treatment [14, 15] and the type of movement, e.g. torque, intrusion or bodily movement [8, 13, 16–23]. In addition a genetic predisposition is assumed. Since orthodontic treatment with a fixed appliance can act as a trigger for severe EARR (sEARR) in genetically predisposed individuals , it is estimated, that the proportion of the hereditary component is 60%-80% for EARR .
The teeth most affected by EARR are the maxillary and mandibular incisors indicating that mechanical factors might play an important role in the development of EARR [18, 20, 25, 26]. SEARR are present when more than 1/3 of the root is resorpted. With conventional brackets (Non-SL) sEARR occurred in about 0.5% of patients under orthodontic treatment .
The search for improved efficiency in orthodontic treatment with less adverse effects has generated new types of brackets . Self-ligating (SL) brackets are an innovation, which have been pioneered in the 1930s. These brackets provide a mechanism of closure for the inserted archwire, so no additional rubber elastics or steel ligatures are needed. Still they have undergone a revival in the recent years with a variety of new features being developed and advanced. A host of advantages such as shorter treatment time, higher rate of tooth movement and fewer appointments have been claimed relating to reduced frictional resistance of these brackets .
By now there are a few studies, which observed no significant benefits of SL brackets compared to Non-SL brackets regarding the occurrence of EARR, treatment time or number of appointments. However most of these studies are with a small amount of patients or compare brackets of different companies or with different prescriptions between them [28, 30–34]. The occurrence of sEARR regarding SL- and Non-SL brackets has not yet been compared.
The objective of this study was to determine the occurrence and severity of EARR on maxillary and mandibular incisors during treatment with SL and Non-SL brackets with the same prescription and from the same company. We furthermore aimed to analyze the amount of extraction cases, number of appointments and treatment time.
Materials and methods
Patients and inclusion/exclusion criteria
For this retrospective study 213 patients, being treated in a private praxis between 2008 and 2012, were included by the following criteria:
completed treatment with a multibracket appliance
presence of panoramic radiograph before and after treatment
completed root growth of the maxillary and mandibular incisors before treatment
no evidence of EARR of the maxillary and mandibular incisors on the pretreatment panoramic radiograph
no severely dilacerated incisor roots
caries-free maxillary and mandibular incisors
trauma before and during active treatment
bends of first, second or third order in the archwire
need for orthognathic surgery
fixed class II appliance
139 patients (females n = 83; males n = 56) were treated with SL brackets (Smart Clip, 3 M Unitek, Monrovia, CA, USA). 74 patients (females n = 51; males n = 23) functioned as a control group treated with Non-SL brackets (Victory, 3 M Unitek, Monrovia, CA, USA). Both bracket types provided 0.022 slots sizes and MBT prescriptions. Treatments of all patients were performed by the same practitioner with a general arch-wire sequence of a 0.015 twistflex (stainless steel), 0.016 nickel-titanium, 0.016 × 0.022 nickel-titanium, 0.017 × 0.025 nickel-titanium and 0.019 × 0.025 stainless steel (Resilient Orthoform III OVOID, 3 M Unitek, Monrovia, CA, USA). Ligation in the control group was performed with steel ligatures for the conventional brackets.
X-ray measurement and patient informed consent
EARR was defined as any reduction in the radiographic lengths of the maxillary and mandibular incisor teeth from the tip of the incisal edge to the apex of the root. Quantitative measurements of the crown and root length of the maxillary and mandibular central and lateral incisors were taken. Any image distortion between the pre- and post-treatment radiographs was calculated using the crown length registrations. This measurement method was established by Linge and Linge and has been described by several studies [26, 27] (Figure 1).
Correction Factor (CF) = C1/C2
C1 = Crown length on pretreatment radiograph
The EARR in millimeter was calculated as following.
EARR = R1 – (R2 x CF)
R1 = Root length before treatment
R2 = Root length after treatment
Pre- and post-treatment radiographs were present due to the usual treatment procedure in Germany, where it is required for diagnostic and treatment procedure. All panoramic radiographs were taken by the same operator. No radiographs were taken out of study reasons and all data were analyzed anonymously to protect the rights of the patients. All patients were informed about using their data for research and gave their informed consent.
The association of SL brackets with EARR. This was investigated using a repeated measures mixed linear model with fixed and random effects. The model was further adjusted for extraction, treatment mechanism, type of tooth, and interaction between tooth location and self-ligating brackets.
The impact of SL brackets on the occurrence of sEARR. For this we used generalized estimating equations to fit a repeated measures binary logistic regression model.
The global significance level of 0.05 was adjusted using a Bonferroni correction which led to local significance levels of α = 0.025.
Secondary research questions included the association of SL brackets with (a) the number of visits within a treatment, (b) the duration of treatment, and (c) the number of extraction cases. This was analyzed using multivariable ANCOVA models additionally adjusted for extraction and treatment mechanic and a binary logistic regression model.
These analyses can be regarded as explorative and respective p-values as descriptive.
To determine systematic error, each of 10 panoramic radiographs taken before and after treatment were reanalyzed five months after the first assessment, a time period long enough for the original measurements to be forgotten in order to determine statistical error according to Dahlberg’s formula. The error in the panoramic radiographs assessment fell within a very good range of (0.14 mm).
SL (n = 139)
Non-SL (n = 74)
Total (N = 213)
Mean ± SD
Mean ± SD
Mean ± SD
Amount of teeth
Age at start of treatment (years)
12.6 ± 2.3
12.1 ± 2.2
12.1 ± 2.2
Amount of rRR between SL and Non-SL (tooth level)
SL (n = 1112)
Non-SL (n = 592)
Total (n = 1704)
Mean ± SD
Mean ± SD
Mean ± SD
3.0 ± 5.6
4.5 ± 6.6
3.5 ± 6.0
Percentage of teeth affected by severe EARR (tooth level)
SL (n = 1112)
Non-SL (n = 592)
Total (N = 1704)
Multivariable analysis evaluating the impact of SL brackets on rRR and sEARR
Parameter estimate (β)c
Number of teeth affected by severe EARR according to tooth group (grade-IV RR according to Malmgren) (tooth level)
Teeth affected by sEARR (n = 6)
Teeth non-affected by sEARR (n = 1698)
Total (N = 1704)
Treatment data (patient level)
(n = 139)
(n = 74)
(N = 213)
Mean ± SD
Mean ± SD
Mean ± SD
Treatment time (months)
20.7 ± 4.9
18.1 ± 5.3
19.8 ± 5.2
Number of visits
13.9 ± 3.3
12.4 ± 3.4
13.4 ± 3.4
Number of extraction cases (patients)
Multivariable analysis evaluating the impact of self-ligating (SL) brackets on number of visits within a treatment, treatment duration, and number of extraction cases
Number of visitsa
Duration of treatmenta
Parameter estimate (β)c
Parameter estimate (β)c
The present retrospective study was intended to determine the amount of EARR, as well as the occurrence of sEARR on maxillary and mandibular incisors in a large amount of patients treated with SL or Non-SL brackets.
Extraoral x-rays such as panoramic radiographs are considered to be less precise than periapical radiographs for the determination of EARR. The main disadvantage of panoramic radiographs is an overestimation due to the magnification of the radiographs up to 20%. In the anterior region, where EARR was measured, the difference between two radiographs was observed to be less than 0.2 mm . Therefore direct metric evaluations of panoramic radiographs are generally accepted as being unreliable due to the large magnification and poor reproducibility . Thus, many investigators have opted to use relative values and classify their findings into resorption stages. By measuring the crown/root-ratio seemingly root shortening caused by proclination of incisors does not distort the relative values. Amongst a variety of techniques to grade EARR we chose to use the root resorption grades according to Malmgren et. al , as their classification has been employed in previous studies [10, 13, 20, 27, 37].
The quantity of patients in our study was in comparison to previous studies very large, giving the possibility to display a higher value of significance. Furthermore all patients were treated by one practitioner with the same bracket prescription and the same sequence of archwires, which offered a very good comparability.
The amount of EARR in the present study showed no statistically significant difference after orthodontic treatment with the SL or Non-SL brackets. Pandis et al. also did not observe a significant difference between SL and Non-SL brackets in the amount of EARR . Although they compared two totally different brackets and their study displayed a small patient collective, our findings are consistent with their results. The working group of Leite et al. analyzed the amount of EARR between SL and Non-SL brackets by using cone beam computed tomography. Only 19 patients were involved in the study, but all teeth were analyzed exactly by cone beam computed tomography. However they could also not detect any significant difference between the two types of brackets . As far as we know our study is based on the largest evaluated collective by now confirming previous results that the type of bracket does not have any impact on the amount of EARR.
To evaluate the occurrence of sEARR with a statistical relevance, a large amount of patients is needed due to the rare appearance. Sehr et al. managed to achieve statistical significant results about the occurrence of sEARR (grade IV according to Malmgren) in patients treated with fixed orthodontic appliances. They could show, that 0.5% of orthodontically treated teeth were affected by sEARR . Thus, the 0.5% of affected teeth in the group of patients with Non-SL brackets of our study is in line with their study. As a result we could approve their findings and show for the first time, that there is no difference in the occurrence of sEARR between SL and Non-SL brackets.
In the present study the central incisors of the mandible were most often affected by sEARR. Most of the previous studies show, that the maxillary and mandibular incisors are more affected by EARR. This might be caused by higher mechanical load due to their smaller root surfaces. The observations of different studies vary equally between central or lateral incisors being more affected by EARR [8, 19, 20, 27].
Treatment time of the two groups differed, whereby the treatment time was 2.6 months shorter with Non-SL brackets compared to SL brackets. This is in concordance with the randomized clinical trial of Fleming et al. . They analyzed the same brackets as we did and also observed that duration treatment with the SL brackets was about 3 months longer. Their difference was not statistically significant due to the smaller amount of patients. Scott et al. analyzed the time for alignment between Non-SL and SL brackets. They found a slightly longer time for the SL brackets, which was also not statistically significant . Due to the larger amount of patients in our study we were able to confirm their results with a statistical relevance. There are also studies describing no differences in treatment time or a faster treatment time for SL brackets [33, 39, 40]. Differences between the studies could be due to the use of different ligatures for Non-SL brackets. In our study Non-SL brackets were combined with steel ligatures whereas other studies use rubber elastics, which might have an impact on the friction values of the inserted archwire.
Number of appointments differed slightly between the two groups but without statistical relevance in the multivariable analysis. This finding is similar to the results of Fleming et al., who observed two more appointments for the group of patients treated with SL brackets compared to Non-SL brackets . In this point the results of different studies are also contradictory, showing that the type of bracket does not influence the necessary appointments of the patients with the practitioner .
Companies promote a better alignment with less need of extraction when fixed appliances with SL brackets are used. We compared the number of extraction cases of both groups of patients and could not detect any difference in the number of extraction cases. Ong et al. analyzed the initial alignment of patients treated with SL and Non-SL brackets and focused on the changes of intercanine and intermolar width. SL brackets did not show a higher efficiency in initial alignment and changes of the arch dimension during treatment were similar between SL and Non-SL brackets . Taken together there is no evidence that SL brackets can reduce the need for tooth extraction or improve alignment compared to Non-SL brackets.
As far as we know this is the largest study by now, showing that there is no significant difference in the amount of EARR between patients treated with SL or Non-SL brackets. Furthermore this is the first study showing that there is no difference in the occurrence of sEARR between the two types of brackets.
There was no statistically significant difference in the amount of EARR and the occurrence of sEARR between the two types of brackets.
Central mandibular incisors were mostly affected by sEARR.
Number of appointments did not display any difference between Non-SL and SL brackets, whereas treatment time with SL brackets was almost three months longer, which was statistically relevant.
There was no evidence for a difference in the amount of extraction cases in the two groups.
We thank Dr. Claudia Jacobs-Müller for providing the patient pool from which the study was performed.
- Weltman B, Vig KW, Fields HW, Shanker S, Kaizar EE: Root resorption associated with orthodontic tooth movement: a systematic review. Am J Orthod Dentofacial Orthop. 2010, 137: 462-476. 10.1016/j.ajodo.2009.06.021. discussion 412AView ArticlePubMedGoogle Scholar
- Brezniak N, Wasserstein A: Root resorption after orthodontic treatment: part 2. Literature review. Am J Orthod Dentofacial Orthop. 1993, 103: 138-146. 10.1016/S0889-5406(05)81763-9.View ArticlePubMedGoogle Scholar
- Brezniak N, Wasserstein A: Root resorption after orthodontic treatment: part 1. Literature review. Am J Orthod Dentofacial Orthop. 1993, 103: 62-66. 10.1016/0889-5406(93)70106-X.View ArticlePubMedGoogle Scholar
- Sameshima GT, Sinclair PM: Predicting and preventing root resorption: part II. Treatment factors. Am J Orthod Dentofacial Orthop. 2001, 119: 511-515. 10.1067/mod.2001.113410.View ArticlePubMedGoogle Scholar
- Sameshima GT, Sinclair PM: Predicting and preventing root resorption: Part I. Diagnostic factors. Am J Orthod Dentofacial Orthop. 2001, 119: 505-510. 10.1067/mod.2001.113409.View ArticlePubMedGoogle Scholar
- Fritz U, Diedrich P, Wiechmann D: Apical root resorption after lingual orthodontic therapy. J Orofac Orthop. 2003, 64: 434-442. 10.1007/s00056-003-0243-5.View ArticlePubMedGoogle Scholar
- Mavragani M, Vergari A, Selliseth NJ, Boe OE, Wisth PL: A radiographic comparison of apical root resorption after orthodontic treatment with a standard edgewise and a straight-wire edgewise technique. Eur J Orthod. 2000, 22: 665-674. 10.1093/ejo/22.6.665.View ArticlePubMedGoogle Scholar
- Parker RJ, Harris EF: Directions of orthodontic tooth movements associated with external apical root resorption of the maxillary central incisor. Am J Orthod Dentofacial Orthop. 1998, 114: 677-683. 10.1016/S0889-5406(98)70200-8.View ArticlePubMedGoogle Scholar
- Mohandesan H, Ravanmehr H, Valaei N: A radiographic analysis of external apical root resorption of maxillary incisors during active orthodontic treatment. Eur J Orthod. 2007, 29: 134-139. 10.1093/ejo/cjl090.View ArticlePubMedGoogle Scholar
- Levander E, Malmgren O: Long-term follow-up of maxillary incisors with severe apical root resorption. Eur J Orthod. 2000, 22: 85-92. 10.1093/ejo/22.1.85.View ArticlePubMedGoogle Scholar
- Artun J, Van ’t Hullenaar R R, Doppel D, Kuijpers-Jagtman AM: Identification of orthodontic patients at risk of severe apical root resorption. Am J Orthod Dentofacial Orthop. 2009, 135: 448-455. 10.1016/j.ajodo.2007.06.012.View ArticlePubMedGoogle Scholar
- Gonzales C, Hotokezaka H, Yoshimatsu M, Yozgatian JH, Darendeliler MA, Yoshida N: Force magnitude and duration effects on amount of tooth movement and root resorption in the rat molar. Angle Orthod. 2008, 78: 502-509. 10.2319/052007-240.1.View ArticlePubMedGoogle Scholar
- Beck BW, Harris EF: Apical root resorption in orthodontically treated subjects: analysis of edgewise and light wire mechanics. Am J Orthod Dentofacial Orthop. 1994, 105: 350-361. 10.1016/S0889-5406(94)70129-6.View ArticlePubMedGoogle Scholar
- Newman WG: Possible etiologic factors in external root resorption. Am J Orthod. 1975, 67: 522-539. 10.1016/0002-9416(75)90298-5.View ArticlePubMedGoogle Scholar
- Malmgren O, Goldson L, Hill C, Orwin A, Petrini L, Lundberg M: Root resorption after orthodontic treatment of traumatized teeth. Am J Orthod. 1982, 82: 487-491. 10.1016/0002-9416(82)90317-7.View ArticlePubMedGoogle Scholar
- Wehrbein H, Fuhrmann RA, Diedrich PR: Periodontal conditions after facial root tipping and palatal root torque of incisors. Am J Orthod Dentofacial Orthop. 1994, 106: 455-462. 10.1016/S0889-5406(94)70067-2.View ArticlePubMedGoogle Scholar
- van Loenen M, Dermaut LR, Degrieck J, De Pauw GA: Apical root resorption of upper incisors during the torquing stage of the tip-edge technique. Eur J Orthod. 2007, 29: 583-588. 10.1093/ejo/cjm060.View ArticlePubMedGoogle Scholar
- Mirabella AD, Artun J: Risk factors for apical root resorption of maxillary anterior teeth in adult orthodontic patients. Am J Orthod Dentofacial Orthop. 1995, 108: 48-55. 10.1016/S0889-5406(95)70065-X.View ArticlePubMedGoogle Scholar
- Linge L, Linge BO: Patient characteristics and treatment variables associated with apical root resorption during orthodontic treatment. Am J Orthod Dentofacial Orthop. 1991, 99: 35-43. 10.1016/S0889-5406(05)81678-6.View ArticlePubMedGoogle Scholar
- Levander E, Malmgren O: Evaluation of the risk of root resorption during orthodontic treatment: a study of upper incisors. Eur J Orthod. 1988, 10: 30-38. 10.1093/ejo/10.1.30.View ArticlePubMedGoogle Scholar
- Kjaer I: Morphological characteristics of dentitions developing excessive root resorption during orthodontic treatment. Eur J Orthod. 1995, 17: 25-34. 10.1093/ejo/17.1.25.View ArticlePubMedGoogle Scholar
- Kaley J, Phillips C: Factors related to root resorption in edgewise practice. Angle Orthod. 1991, 61: 125-132.PubMedGoogle Scholar
- Casa MA, Faltin RM, Faltin K, Sander FG, Arana-Chavez VE: Root resorptions in upper first premolars after application of continuous torque moment. Intra-individual study. J Orofac Orthop. 2001, 62: 285-295. 10.1007/PL00001936.View ArticlePubMedGoogle Scholar
- Harris EF, Kineret SE, Tolley EA: A heritable component for external apical root resorption in patients treated orthodontically. Am J Orthod Dentofacial Orthop. 1997, 111: 301-309. 10.1016/S0889-5406(97)70189-6.View ArticlePubMedGoogle Scholar
- Mirabella AD, Artun J: Prevalence and severity of apical root resorption of maxillary anterior teeth in adult orthodontic patients. Eur J Orthod. 1995, 17: 93-99. 10.1093/ejo/17.2.93.View ArticlePubMedGoogle Scholar
- Linge BO, Linge L: Apical root resorption in upper anterior teeth. Eur J Orthod. 1983, 5: 173-183. 10.1093/ejo/5.3.173.View ArticlePubMedGoogle Scholar
- Sehr K, Bock NC, Serbesis C, Honemann M, Ruf S: Severe external apical root resorption–local cause or genetic predisposition?. J Orofac Orthop. 2011, 72: 321-331. 10.1007/s00056-011-0036-1.View ArticlePubMedGoogle Scholar
- Chen SS, Greenlee GM, Kim JE, Smith CL, Huang GJ: Systematic review of self-ligating brackets. Am J Orthod Dentofacial Orthop. 2010, 137: 726 e721-726 e718. discussion 726–727Google Scholar
- Miles PG: Self-ligating brackets in orthodontics: do they deliver what they claim?. Aust Dent J. 2009, 54: 9-11. 10.1111/j.1834-7819.2008.01081.x.View ArticlePubMedGoogle Scholar
- Celar A, Schedlberger M, Dorfler P, Bertl M: Systematic review on self-ligating vs. conventional brackets: initial pain, number of visits, treatment time. J Orofac Orthop. 2013, 74: 40-51. 10.1007/s00056-012-0116-x.View ArticlePubMedGoogle Scholar
- Fleming PS, DiBiase AT, Lee RT: Randomized clinical trial of orthodontic treatment efficiency with self-ligating and conventional fixed orthodontic appliances. Am J Orthod Dentofacial Orthop. 2010, 137: 738-742. 10.1016/j.ajodo.2009.06.023.View ArticlePubMedGoogle Scholar
- Leite V, Conti AC, Navarro R, Almeida M, Oltramari-Navarro P, Almeida R: Comparison of root resorption between self-ligating and conventional preadjusted brackets using cone beam computed tomography. Angle Orthod. 2012, 82: 1078-1082. 10.2319/080911-501.1.View ArticlePubMedGoogle Scholar
- Machibya FM, Bao X, Zhao L, Hu M: Treatment time, outcome, and anchorage loss comparisons of self-ligating and conventional brackets. Angle Orthod. 2013, 83: 280-285. 10.2319/041912-326.1.View ArticlePubMedGoogle Scholar
- Pandis N, Nasika M, Polychronopoulou A, Eliades T: External apical root resorption in patients treated with conventional and self-ligating brackets. Am J Orthod Dentofacial Orthop. 2008, 134: 646-651. 10.1016/j.ajodo.2007.01.032.View ArticlePubMedGoogle Scholar
- Sameshima GT, Asgarifar KO: Assessment of root resorption and root shape: periapical vs panoramic films. Angle Orthod. 2001, 71: 185-189.PubMedGoogle Scholar
- Goz G, Rakosi T: [Apical root resorption during orthodontic treatment]. Fortschr Kieferorthop. 1989, 50: 196-206. 10.1007/BF02168741.View ArticlePubMedGoogle Scholar
- Harris EF, Boggan BW, Wheeler DA: Apical root resorption in patients treated with comprehensive orthodontics. J Tenn Dent Assoc. 2001, 81: 30-33.PubMedGoogle Scholar
- Scott P, DiBiase AT, Sherriff M, Cobourne MT: Alignment efficiency of Damon3 self-ligating and conventional orthodontic bracket systems: a randomized clinical trial. Am J Orthod Dentofacial Orthop. 2008, 134: 470 e471-478.Google Scholar
- Harradine NW: Self-ligating brackets and treatment efficiency. Clin Orthod Res. 2001, 4: 220-227.View ArticlePubMedGoogle Scholar
- Eberting JJ, Straja SR, Tuncay OC: Treatment time, outcome, and patient satisfaction comparisons of Damon and conventional brackets. Clin Orthod Res. 2001, 4: 228-234. 10.1034/j.1600-0544.2001.40407.x.View ArticlePubMedGoogle Scholar
- Ong E, McCallum H, Griffin MP, Ho C: Efficiency of self-ligating vs conventionally ligated brackets during initial alignment. Am J Orthod Dentofacial Orthop. 2010, 138: 138 e131-137. discussion 138–139Google Scholar
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.