OUP user menu

45,X/46,XX karyotype mitigates the aberrant craniofacial morphology in Turner syndrome

Sara Rizell, Marie-Louise Barrenäs, Anna Andlin-Sobocki, Christina Stecksén-Blicks, Heidrun Kjellberg
DOI: http://dx.doi.org/10.1093/ejo/cjs014 467-474 First published online: 24 April 2012

Abstract

The aim of this project was to study the impact on craniofacial morphology from Turner syndrome (TS) karyotype, number of intact X chromosomal p-arms, and age as well as to compare craniofacial morphology in TS with healthy females. Lateral radiographs from 108 females with TS, ranging from 5.4 to 61.6 years, were analysed. The TS females were divided into four karyotype groups: 1. monosomy (45,X), 2. mosaic (45,X/46,XX), 3. isochromosome, and 4. other, as well as according to the number of intact X chromosomal p-arms. The karyotype was found to have an impact on craniofacial growth, where the mosaic group, with presence of 46,XX cell lines, seems to exhibit less mandibular retrognathism as well as fewer statistically significant differences compared to the reference group than the 45,X karyotype. Isochromosomes had more significant differences versus the reference group than 45,X/46,XX but fewer than 45,X. To our knowledge, this is the first time the 45,X/46,XX and isochromosome karyotypes are divided into separate groups studying craniofacial morphology. Impact from p-arm was found on both maxillary and mandibular length. Compared to healthy females, TS expressed a shorter posterior and flattened cranial base, retrognathic, short and posteriorly rotated maxilla and mandible, increased height of ramus, and relatively shorter posterior facial height. The impact of age was found mainly on mandibular morphology since mandibular retrognathism and length were more discrepant in older TS females than younger.

Introduction

Turner syndrome (TS) has a prevalence of 1/2000–1/3000 live born females and is caused by a complete or partial absence of one of the X chromosomes, frequently accompanied by cell line mosaicism (Nielsen and Wohlert, 1990; Gravholt et al., 1996). The main features are short stature and ovarian failure but also the skeletal morphology and constitution are affected since various skeletal malformations and osteoporosis are over-represented (Karlberg et al., 1991; Lippe, 1991; Landin-Wilhelmsen et al., 1999; Kim et al., 2001; Bryman et al., 2011).

It is also evident that the craniofacial morphology in TS is influenced by the underlying chromosomal defect since literature is unanimous that females with TS have a retrognathic maxilla and mandible (Jensen, 1985; Peltomäki et al., 1989; Rongen-Westerlaken et al., 1992; Babic et al., 1993; Midtbø et al., 1996; Simmons, 1999; Hass et al., 2001; Perkiomäki et al., 2005). It is also evident from the above mentioned studies that the cranial base is more flattened and the posterior cranial base is shorter in TS females. The majority of previously published reports on craniofacial growth were performed either solely on 45,X karyotype or a mixture of different karyotypes. To our knowledge, only a few studies intended to investigate the influence of karyotype on craniofacial morphology (Table 1). The limited number of participants in these studies deteriorates an optimal subdivision of the karyotypes. Without exception, 45,X is compared versus a highly heterogeneous group of karyotypes; therefore, the results must be interpreted with caution. Jensen (1985) found that 45,X had a more retrognathic maxilla than the group of mosaics, while the other authors presented in Table 1 found no significant differences between 45,X and various kinds of mosaics and isochromosomes. Only Midtbø et al. (1996) compared the karyotypes one by one versus controls and found more cephalometric variables with significant differences in the 45,X group. There are studies indicating that TS females with a mosaic karyotype including 46,XX cell-lines have a less aberrant general phenotype than other TS (El-Mansoury et al., 2007, 2009; Bryman et al., 2011). On the other hand, the TS isochromosomes seem to have a more aberrant phenotype in other respects, i.e. birth size, growth hormone (GH) deficiency, bone age delay, hypothyreosis, hearing loss, and tooth width (Zinman et al., 1984; Schmitt et al., 1997; Hultcrantz, 2003; Hagman et al., 2010; Rizell et al., 2012). Merging these two disparate karyotypes might complicate the interpretation of results. Of course, subdividing the material into rare groups as ring- or marker-chromosomes would be highly interesting, but a drastic increase in the number of participants would have been required.

View this table:
Table 1

Published studies on the impact of Turner syndrome karyotype on craniofacial morphology.

ReferenceAge (years) N 45,X versus all other karyotypes45,X or other karyotypes versus controls
Jensen (1985) Adults41s–n–ss: 45,X < other
Rongen-Westerlaken et al. (1992) 4–1769No differences
Rongen-Westerlaken et al. (1993) 3.6–16.619No differences
Midtbø et al. (1996) 7–1733No differences45,X had more variables differing from controls than others
Dumancic et al. (2010) 10–3336No differences

Studies on individuals with different abnormal sex chromosome constitutions state that the number of sex chromosomes affects both the cranial base angle and the mandibular prognathism (Peltomäki et al., 1989; Babic et al., 1993; Brown et al., 1993; Grön et al., 1997; Krusinskiene et al., 2005). One possible explanation for the differences in craniofacial growth comparing sex chromosome aberrations could be that the Short Stature Homeobox (SHOX) gene is located on the short arm (p-arm) of the X- or Y-chromosomes and is, in haploinsufficiency, believed to be involved in skeletal malformations and growth retardation in several short stature syndromes (Leka et al., 2006; Marchini et al., 2007). Influence from haploinsufficiency of the SHOX gene on craniofacial growth is plausible as it seems to be involved in skeletal morphology in several other parts of the body. The localization of the SHOX gene raises the question whether craniofacial morphology differs between TS females exhibiting two X-chromosomes with unaffected p-arms (e.g. TS with deletions on the long arm of the X-chromosome) and TS with only one X-chromosome with unaffected p-arm (e.g. 45,X). No published studies have addressed this question before.

The aim of this project was to

  1. compare craniofacial morphology in different TS karyotypes as well as in TS versus controls

  2. compare craniofacial morphology in TS individuals with one or two X chromosomes with intact p-arms

  3. study the effect of age on craniofacial morphology in TS

Subjects

One hundred thirty-two Swedish females with a diagnosis of TS, living in the regions of Göteborg, Uppsala, and Umeå in Sweden, granted consent to participate in this study. Nineteen individuals were excluded since their records were missing or of poor quality, four individuals were excluded because they had extensive dental restorations or severe tooth loss, and one girl since she had a unilateral cleft. The remaining 108 subjects ranged from 5.4 to 61.6 years. All participants underwent genetic karyotyping by routine chromosomal analysis of peripheral lymphocytes. The distribution of TS karyotype subgroups and number of intact p-arms is shown in Table 2. The TS females were grouped into four karyotype categories: 1. monosomy (45,X), 2. mosaic (45,X/46,XX), 3. isochromosome, and 4. other but also according to the presence of one or two unaffected X chromosomal p-arms (Table 2). The mosaic karyotype 45,X/46,XX was counted as having two unaffected X chromosomal p-arms even though this was only true for part of the cells lines. The TS isochromosome karyotype has one normal X chromosome and one X chromosome displaying two identical arms due to duplication of the long arm (q-arm) and loss of the p-arm, i.e. long arm trisomy and short arm monosomy. In the present group of TS females, 77.7 per cent had been treated with growth hormone (GH). Information about previous orthodontic treatment was obtained from patient files at the orthodontic clinics in Göteborg, Uppsala, and Umeå and 34 of the 108 participants (31 per cent) had previously had such treatment.

View this table:
Table 2

The distribution of Turner syndrome (TS) karyotype, number of X chromosomes with an intact p-arm and, age in the four karyotype categories.

TS karyotype and categoriesTotal (n)One X chromosome with intact p-arm (n)Two X chromosomes with intact p-arms (n)Age
MeanMedianRange
1. 45,X4040019.913.65.7–61.6
2. 45,X/46,XX1201224.216.25.4–60.6
3. Isochromosomes2828015.210.65.4–50.3
4. Other
    Deleted X chromosomes72516.9132.86.2–52.2
    Translocated X chromosomes312
    Inverted X chromosomes431
    Marker chromosomes505
    Ring chromosomes440
    Y chromosomal material440
    45,X/47,XXX101*
    Total108822618.412.75.4–61.6
  • * Case with three p-arms in 27 of 52 cells.

Methods

Lateral cephalometric radiographs were taken in intercuspidal position, with the head fixated in a cephalostat. Since the participants were examined in different centers, multiple cephalostats with different focus-film as well as film–midline distances were used; therefore, the enlargement factor was adjusted to zero for all linear variables. The measurements were made in a computerized cephalometric software program (FACAD®; Ilexis AB, Linköping, Sweden) by one investigator (SR), who was blinded for the karyotype of the participants. The landmarks and lines used are defined by Thilander et al. (2005), except for the line s–ba, which is defined in the Bolton standards and displayed in Figure 1 (Broadbent et al., 1975). All cephalometric variables were converted into age- and gender-specific standard deviation scores (SDS) using a reference group of healthy females (Thilander et al., 2005) except for the single variable s–ba, where Bolton standards were used (Broadbent et al., 1975). The reference values were used as a ‘golden standard’ and counted as zero. The difference between the actual value and the mean of the specific age population was divided by the standard deviation (SD) for the specific age population. SDS displays how much the actual value differs from the golden standard, i.e. an SDS value close to zero reflects similarity with the reference group, a positive SDS means a higher value than the reference group and a negative SDS means a lower value compared to the reference group.

Figure 1

Cephalometric landmarks and lines.

The study was approved by the University of Gothenburg ethics committee.

Error of method

Calibration to locate the landmarks was made together with Ulf Adolfsson (UA), one of the investigators in the previously mentioned reference study (Thilander et al., 2005). The inter-individual error between SR and UA, calculated from measurements on 16 randomly chosen radiographs, did not exceed 0.7 degrees for the angular measurements and 0.8 mm for the linear measurements except for the variables n–s–ba and sp’–pm (1.4 degrees and 1.5 mm, respectively) (Dahlberg, 1940). To calculate the intra-individual error for the cephalometric measurements, the recordings were repeated by SR on 16 randomly chosen radiographs in 1 month intervals. The intra-individual error of measurement did not exceed 0.8 degrees for the angular measurements and 0.3 mm for the linear measurements, except for the variable pns–ans (1.0 mm) (Dahlberg, 1940).

Statistics

One sample t-test was used to test for statistically significant differences between SDS for the cephalometric measurements of the entire TS group and the reference groups of healthy females as well as for each karyotype group one by one versus the reference groups (Broadbent et al., 1975; Thilander et al., 2005). Analysis of covariance (ANCOVA) was used to test for impact from age, karyotype, and number of X chromosomes with intact p-arms on the cephalometric variables converted into age-specific SDS. ANCOVA was supplemented with Student–Newman–Keuls post hoc test, indicating which karyotype groups were divergent. The regression coefficient indicates a positive or negative correlation with a cephalometric variable and age. Two-sample t-test was used to test for significant differences between individuals with or without previous orthodontic treatment. P-values less than 0.05 were considered as statistically significant.

Results

No statistically significant differences were found between individuals with or without previous orthodontic treatment in any of the variables; therefore, the groups were merged. Impact from karyotype was found only on the mandibular retrognathism (s–n–sm) (Table 3). Student–Newman–Keuls post hoc test showed no significant differences for any of the tested variables. However, the most divergent groups concerning mandibular retrognathism were 45,X/46,XX and 45,X, where 45,X/46,XX displayed less mandibular retrognathism than 45,X (Table 3). The group by group versus controls comparison displayed that the 45,X/46,XX group had fewer significant differences from controls compared to all other groups, while isochromosomes exhibited more significant differences from controls than 45,X/46,XX but fewer than 45,X (Table 4). The majority of the cephalometric variables in TS differed significantly from the reference groups of healthy females (Table 4). The cranial base (n–s–ba and n–s–ar) was flattened and the posterior cranial base (s–ar and s–ba) was shorter, while the anterior cranial base (s–n) was longer compared to the controls. Both the maxilla and the mandible were found to be more retrognathic (s–n–ss, s–n–sm, and s–n–pg) and more posteriorly inclined (NSL–NL and NSL–ML) in TS. Additionally, the maxilla and mandible were shorter (sp’–pm, tgo–gn, and ar–pg), while ramus was longer (ar–tgo). The vertical relations were affected for the ratios between the upper anterior facial height and the entire facial height as well as between the anterior and posterior facial heights, both ratios being increased in TS.

View this table:
Table 3

Variables that showed statistically significant impact of age, karyotype, or number of intact p-arms on cephalometric variables, converted into age-specific standard deviation scores (SDS) using reference groups of healthy females (Broadbent et al., 1975; Thilander et al., 2005). ANCOVA, Analysis of covariance.

VariableANCOVA (p-value)Impacting variableRegression coefficientSDS (mean)
s–n–sm0.006Age−0.033−2.37
0.041Karyotype 45,X/46XX−1.76
Others−2.12
Isochromosomes−2.16
45,X−2.87
s–n–pg0.001Age−0.042−2.45
ss–n–sm0.002Age0.037−0.02
s–n0.002Age−0.0310.82
sp′–pm0.028p-arm One intact p-arm−0.51
Two intact p-arms−1.32
tgo–gn0.001Age−0.040−2.05
ar–pg<0.000Age−0.049−0.59
0.048p-arm One intact p-arm−0.54
Two intact p-arms−0.74
  • For the remaining cephalometric variables not displayed in Table 3, no impact was found from the tested factors. A positive regression coefficient indicates a positive correlation and conversely, a negative regression coefficient indicates a negative correlation.

View this table:
Table 4

Mean standard deviation scores and results from comparison of the cephalometric variables from the entire group of Turner syndrome (TS) as well as each karyotype one by one versus the reference groups (Broadbent et al., 1975; Thilander et al., 2005) using one sample t-test.

All TS45,X45,X/46,XXIsochromosomesOther
s–n–ss−2.14***−2.30***−1.41**−2.33***−2.03***
s–n–sm−2.37***−2.87***−1.76**−2.16***−2.12***
s–n–pg−2.45***−2.91***−2.12**−2.12***−2.25***
ss–n–sm−0.020.520.10−0.58−0.29
n–s–ba1.13***1.14***0.89*0.99***1.34**
n–s–ar0.91***0.90***0.510.87***1.12**
s–n0.82***1.10***0.030.79*0.80**
s–ar−0.61***−0.45*−0.87−0.77*−0.57
s–ba−1.65***−1.54***−1.80*−1.45***−1.93***
NSL/NL1.57***1.66***1.55**1.59***1.43***
NSL/ML1.17***1.01***1.58**1.14**1.25***
NL/ML0.18−0.120.670.230.34
ML/RL0.08−0.280.510.270.24
n–sp’/n–gn0.25*0.36*0.150.370.03
sp’–gn/n–gn−0.02−0.130.04−0.140.20
sp’–pm−0.71***−0.74***−1.18*−0.48−0.70
n–gn/s–tgo2.17***1.66***2.93**2.05**2.68***
ar–tgo0.51***0.75***−0.040.66*0.28
tgo–gn−2.05***−2.28***−2.06*−1.91***−1.88***
ar–pg−0.59***−0.92***−0.71−0.30−0.37
  • Asterisks indicate level of significance (*P < 0.05, **P < 0.01, ***P < 0.001). The group consisting from all TS had a flattened cranial base with a shorter posterior portion and more retrognathic and posteriorly inclined maxilla and mandible versus controls. The 45,X/46,XX karyotype exhibited fewer variables differing from controls compared to 45,X, while isochromosomes were situated in between.

Impact was found from the number of X chromosomes with intact p-arms on both maxillary and mandibular length (sp’–pm and ar–pg) and from age on intermaxillary relation, anterior cranial base length, and mandibular prognathism and length (ss–n–sm, s–n, s–n–sm, s–n–pg, tgo–gn, and ar–pg) (Table 3).

Negative correlation was found for age and the variables s–n–sm, s–n–pg, s–n, tgo–gn, and ar–pg and a positive correlation with ss–n-sm (Table 3). In older individuals, the mandible was more retrognathic compared to controls than in younger individuals, which also gave a larger sagittal intermaxillary discrepancy (ss–n–sm). The anterior cranial base, which was longer in TS than controls, displayed less discrepancy in older individuals. The shorter mandibular length found in TS remained with age and this difference was larger for older patients.

Isochromosomes exhibited more significant differences from controls than 45,X/46,XX but fewer than 45,X.

Discussion

The main finding in this study was that individuals with 45,X/46,XX TS karyotype exhibited a less aberrant craniofacial morphology versus controls than other karyotypes. Our results confirm the results from Midtbø et al. (1996), who found a pattern of less deviation from controls among the group including mosaics. Still, the Norwegian material included not only 45,X/46,XX but also various kinds of mosaics and isochromosomes, which might have biased the result. To our knowledge, this is the first time studying craniofacial morphology, a material was collected which enabled a separation of the mosaics exhibiting 46,XX cell lines and the isochromosomes instead of merging them together. The separation of these karyotypes resulted in a finding of that isochromosomes exhibited more significant differences from controls than 45,X/46,XX but fewer than 45,X. Converting the values into age-specific SDS was additionally a presupposition for the possibility of assessing all individuals independent of age in one group and thereby avoiding division into different age groups. Additionally, impact from karyotype was found on mandibular retrognathism. Even if the differences could not be proven with Student–Newman–Keuls post hoc test, the mean SDS reveal that females with 45,X/46,XX karyotype express less mandibular retrognathism compared to females with 45,X karyotype (Table 3). This result is not previously found since only an increased maxillary retrognathism in 45,X compared to other karyotypes or no differences at all between different TS karyotypes has been reported (Table 1). The reports differing from our results is presumably related to the grouping of various karyotypes together while we aimed to achieve a more specific division. The present study showed that the presence of unaffected cell lines (46,XX) in the mosaic group mitigated the mandibular retrognathism; still, the mandibular morphology was not fully normalized (Table 4). Similar positive effects on karyotypes exhibiting 46,XX cell lines have been reported on e.g. spontaneous pregnancies, fine motor function, body balance, hearing, presence of TS stigmata, and tooth width (El-Mansoury et al., 2007, 2009; Bryman et al., 2011; Rizell et al., 2012).

The cranial base was flattened and the posterior cranial base was shorter in the entire TS group versus the reference group (Table 4), which confirms previous results (Jensen, 1985; Peltomäki et al., 1989; Rongen-Westerlaken et al., 1992; Babic et al., 1993; Midtbø et al., 1996; Simmons, 1999; Hass et al., 2001; Perkiomäki et al., 2005). These changes are already seen prenatally and seem to be present both in young girls and in adult TS females (Jensen, 1985; Rongen-Westerlaken et al., 1992; Midtbø et al., 1996; Andersen et al., 2000). From these observations, the conclusion is drawn that the growth disturbances occur already before birth and the pattern of craniofacial morphology is maintained during the growth period. The anterior part of the cranial base was found to be longer in TS than in controls in the present study (Table 4). This might be surprising. However, with one exception, the previously mentioned studies report the anterior cranial base as being longer or of the same size as controls. Additionally, we found both the maxilla and the mandible to be retrognathic and posteriorly inclined in the TS group (Table 4). As suggested by Björk (1955), there is an association between the relation of the jaws and the cranial base angulation, i.e. with an increased cranial base angle comes a more retrognathic face. The short mandibular length and retrognathism are the underlying measures for the micrognathia, which is one of the most significant features in TS (El-Mansoury et al., 2007).

The number of unaberrant p-arms was found to have an impact on both maxillary and mandibular length (sp′–pm and ar–pg); however, not in the way we expected, since it seemed that two intact p-arms did not decrease the deviation from the healthy women (Table 3). Until now, the effect on craniofacial growth from the number of X chromosomal p-arms has been carried out by comparing 45,X karyotype with 46,XX females, which might appear as the ultimate p-arm comparison. However, we wanted to test a new approach of grouping, which permitted inclusion of all kinds of TS karyotypes, also the rare ones. This categorization may possibly be too unspecific or biased since the p-arm can display an aberration but still express the causing gene(s) or other unknown factors with higher impact than the number of intact p-arms are involved. Presence of marker cells or 47,XXX cell lines might also have influenced the result in a negative direction. In a previous study on tooth width, no impact from the number of intact X-chromosomal p-arms was found (Rizell et al., 2012).

Age seemed to have an impact on several of the studied variables. With increased age, the mandibular retrognathism was more pronounced compared to controls, which also is associated with an enlarged sagittal intermaxillary discrepancy (ss–n–sm) (Table 3). Additionally, TS individuals displayed a shorter mandibular length than controls and this difference was larger for older patients (Table 3). As described above, the anterior cranial base was longer in TS, but the difference compared to controls decreased with increased age. All variables that age had an impact on, except for s–n, illustrate that mandibular growth impairment worsens with age and that the older females in this study seem to have a more discrepant mandibular morphology and retrognathism than younger females. Since the aim was to evaluate impact from age on craniofacial morphology, we included TS within a wide age range and therefore comprising females in different stages of GH treatment or with no history of such treatment. Among the older females, there was a larger portion without previous GH treatment, which might be part of the explanation for the more severe mandibular changes. However, to evaluate the effect on craniofacial growth, additional data of GH as dose, start age, and treatment length are needed.

Our findings elucidate that TS females exhibit a deviating craniofacial growth pattern compared to healthy females and the abnormal craniofacial morphology is presumably caused by several mechanisms. From literature, we know that the genes SHOX and SHOX2 are located in the pseudoautosomal region on the X chromosome where the haploinsufficiency causes growth retardation and several skeletal TS characteristics (Clement-Jones et al., 2000; Ross et al., 2001). Clement-Jones et al. (2000) found SHOX expressed in the mesodermal core in the pharyngeal arches in human embryos. It is claimed that this gene is coding for limb development but also for structures originating from the first and second pharyngeal arches, which is the origin of maxilla and mandible (except for the condylar cartilage) as well as structures in the middle ear (Clement-Jones et al., 2000). It is assumable that haploinsufficiency of SHOX gene might affect craniofacial morphology as well.

One of the striking cephalometric characteristics in TS is the morphology of the cranial base, which is flattened and has a shorter posterior part (Table 4). The cranial base origins from the chondrochranium, which acts as a template for the later bony cranium and is transformed into bone by endochondral ossification. Bands of cartilage, synchondroses, remain in the cranial base as growth sites between the ossifying centers. There are reports on both abnormal structure and histochemistry of the cartilage in TS (Stanescu et al., 1965). Geerkens et al. (1995) state that biglycan, which is a component of cartilage, is dependent on the number of X chromosomes and that the amount of biglycan is decreased in TS. The gene for biglycan (BGN) is mapped on the long arm of the X chromosome and is suggested to escape X inactivation (Latham and Scott, 1970; Geerkens et al., 1995; Wadhwa et al., 2004, 2005). Our result with a mitigation of the craniofacial deviations in the 45,X/46,XX group, is in accordance with these theories, since the 45,X/46,XX karyotype exhibit cell lines without haploinsufficiency of the SHOX or BGN genes. Accordingly, both 45,X and isochromosomes, which seem to have a more deviating craniofacial morphology, exhibit haploinsufficiency of the above mention genes.

It has been suggested that the short stature and embryonic growth retardation in 45,X can be explained by the hypothesis about the retardation of cell division, which results in a prolonged cell cycle. It is reported that the cell generation time in 45,X fibroblasts, as well as in trisomic cells, is significantly longer compared to unaffected cells (Paton et al., 1974; Simpson and Lebeau, 1981). The inability to upregulate the rate of the cell cycle during the short developmental time window available for the craniofacial structures causes a reduced cell number, which is assumed to be a reason for the growth deficiency (Barrenäs et al., 2000). This theory also suggests that growth disturbances could be caused by a delayed timing from signalling of growth factors important for craniofacial growth. A disturbance during a critical period during growth of the primary cartilages might cause irreversible effects on the morphology of the craniofacial skeleton (Diewert, 1985). This theory supports the results of an aberrant craniofacial morphology generally in TS but is also explaining why we found more deviations in 45,X compared to 45,X/46,XX.

Conclusions

This study has shown that karyotype has an impact on craniofacial growth, where the mosaic group with presence of 46,XX cell lines seems to exhibit less mandibular retrognathism as well as fewer significant differences from the reference group compared to the 45,X karyotype. For the first time studying craniofacial growth, it was possible to subdivide isochromosomes and 45,X/46,XX into two different groups and isochromosomes was found to exhibit a higher number of significant differences from controls than 45,X/46,XX but fewer than 45,X. Compared to healthy females, TS expressed a short posterior and flattened cranial base, retrognathic, short, and posteriorly rotated maxilla and mandible, increased ramus height, and relatively shorter posterior facial height. Impact from the number of p-arms was found on the length of maxilla and mandible while impact from age was found mainly on mandibular growth since mandibular retrognathism and length were more discrepant in elder TS females than younger females.

Funding

The Anders Otto Svärd Foundation; The Swedish Dental Society; The Gothenburg Dental Society.

Acknowledgments

The authors wish to thank Ulf Adolfsson for help with calibration against the Swedish reference material.

References

View Abstract