The effect of first and second premolar extractions on third molars: A retrospective longitudinal study

Author links open overlay panel

A.MiclotteaB.GrommenbM.Cadenas de Llano-PérulaaA.VerdonckaR.JacobsbcG.Willemsa

Show more

https://doi.org/10.1016/j.jdent.2017.03.007

Get rights and content

Under a Creative Commons license

open access

Abstract

Objectives

To analyse the effect of first and second premolar extractions on eruption space for upper and lower third molars and on third molar position and angulation during orthodontic treatment.

Methods

The sample consisted of 296 patients of which 218 patients were orthodontically treated without extraction and 78 patients with extraction of first or second premolars. The eruption space for third molars was measured on pre– and posttreatment lateral cephalograms, whereas the angulation, vertical position, the relation with the mandibular canal and the mineralization status of third molars were evaluated using pre– and posttreatment panoramic radiographs. All data were statistically analyzed.

Results

The increase in eruption space and the change in vertical position of upper and lower third molars significantly differed between patients treated with and without premolar extractions, whereas the change in angulation, relationship with the mandibular canal and mineralization status of the third molars did not significantly differ between patients treated with and without premolar extractions.

Conclusions

The retromolar space and the position of third molars significantly change during orthodontic treatment in growing patients. Premolar extractions have a positive influence on the eruption space and vertical position of third molars, whereas they do not influence the angular changes of third molars. Due to the retrospective character of the study, these conclusions should be carefully considered. Further prospective research is necessary for better insights into this complex topic.

Clinical significance

This study stresses the importance of considering the possible effects of orthodontic treatment on third molars during treatment planning.

Previous article
Next article

Keywords

OrthodonticsPremolar extractionRetromolar spaceThird molars

1. Introduction

Accounting for 98% of all impactions, third molars are the most frequently impacted teeth [1], [2]. A recent meta-analysis of 49 studies, involving 83 484 individuals reported a worldwide third molar impaction rate of 24.40% [3]. Different factors such as morphology, mesiodistal width, unfavorable uprighting and path of eruption, have been associated with third molar impaction [4]. However, the main reason for third molar impaction is assumed to be a lack of retromolar space[4], [5], [6], [7], [8], which was reported by Björk et al. as limited in 90% of third molar impaction cases, [8]. Retromolar space depends in the upper jaw on the growth of the maxillary tuberosity along with alveolar growth and the mesial drift of the upper first molars [9]. In the lower jaw, it depends on the resorption at the anterior border of the mandibular ramus and the direction in which the teeth erupt during the functional phase of eruption [10]. Furthermore, Björk identified several factors linked with the impaction of lower third molars [6]: a vertical direction of condylar growth, a reduced mandibular length, a backward-directed eruption of the mandibular dentition and a retarded maturation of the third molars [6]. The more anteriorly the posterior teeth erupt, the more the retromolar space will increase [6], [8]. Condylar growth in a predominantly vertical direction is associated with reduced resorption at the anterior aspect of the mandibular ramus and forward growth rotation of the mandible, whereas more backward-directed growth at the condyles is associated with increased resorption at the anterior border of the mandibular ramus and a posterior growth rotation [5], [6], [11], [12].

Besides natural growth, the retromolar space is also influenced by orthodontic treatment [7]. Distalization of the upper first molars has a negative influence on the space available for the eruption of upper third molars [13], [14], [15], whereas orthodontic treatment carried out with extraction therapy is often found to improve the chance of successful third molar eruption. Several authors reported that most lower and upper third molars erupted successfully after the extraction of second molars [16], [17], [18], [19]. Richardson and Richardson as well as De-la Rosa-Gay et al. also found that the less developed the third molar is at the time of second molar extraction, the higher the chances are for its eruption [17], [18]. Bayram et al., Livas et al. and Halicioglu et al. investigated the effect of first molar extractions on third molar eruption [20], [21], [22]. They concluded that the extraction of first permanent molars considerably reduces the frequency of third molar impaction. Consistent with these findings, several studies have shown that orthodontic treatment involving premolar extractions has a positive influence on the development and position of third molars [4], [5], [10], [23], [24], [25], [26], [27] by increasing the eruption space for the third molar due to mesial movement of the first and second molars during space closure.

However, as previously mentioned, third molar impaction has been associated with other factors, such as an unfavorable inclination of the third molar [4]. During development, third molars permanently change their inclination and undergo important pre-eruptive rotational movements [28], [29], [30], preventing impaction of the third molar. Several authors reported that orthodontic treatment involving premolar extraction significantly improved third molar angulation due to an increase in retromolar space [10], [31], [32], whereas other authors did not find significant differences [4], [33], [34].

Although several studies on the effect of premolar extractions on third molars are published, the sample sizes were often small, a lot of studies did not distinguish between first and second premolar extraction, only considered the effect of first premolar extractions or only investigated lower third molars. Therefore, this retrospective study aimed at investigating the effect of both first and second premolar extraction during orthodontic treatment on the space available for both upper and lower third molars. Furthermore, we investigated the possible change in angulation and vertical position of third molars in patients treated with and without premolar extractions. Additionally, we evaluated the relation between the lower third molars and the alveolar nerve before and after treatment.

2. Materials and methods

2.1. Materials

The sample consisted of pre- and posttreatment panoramic radiographs and lateral cephalograms of growing patients, orthodontically treated with or without premolar extractions and with radiographic evidence of at least one third molar. Patients with craniofacial disorders, agenesis or missing teeth before the start of treatment were not included. All of the included patients were treated in the Department of Orthodontics of the University Hospitals Leuven, Leuven, Belgium and finished their treatment between January 2008 and December 2014. The cephalometric radiographs as well as the panoramic radiographs were generated by a Veraview, Morita (Kyoto, Japan) or a Cranex Tome, Soredex (Tuusula, Finland). All radiographs were stored as DICOM files. Patients with insufficient radiographic image quality were excluded (n = 6). Because of overlap of the left and right side on a cephalometric radiograph, patients with asymmetrical premolar extractions were excluded (n = 18). It has been reported that the space available for upper third molars might be influenced by orthodontic distalization appliances [15]. Therefore, we also excluded 188 patients in the non-extraction group and 29 patients in the extraction group who were treated with fixed appliances together with a distalization appliance, such as headgear. The final sample consisted of 296 patients, of which 116 patients were treated with functional and fixed appliances and 24 patients had expansion of the upper jaw before treatment with fixed appliances. Of the 296 patients, 218 patients (103 males, 115 females) were treated without extractions of premolars and 78 patients (37 males, 41 females) were treated with extractions of first or second premolars. Of the 78 patients treated with extractions, 23 patients only had extractions in the upper jaw, 7 patients only had extractions in the lower jaw and 48 patients had extractions in both jaws. In the upper jaw, the first premolar was extracted in 54 patients, whereas the second premolar was extracted in 17 patients. In the lower jaw, the first premolar was extracted in 25 patients and the second premolar in 30 patients.

2.2. Methods

On the lateral cephalograms, the mandibular plane angle and the space available for the upper and lower third molar was measured (Fig. 1). According to the mandibular plane angle, defined by Steiner as the SN-GoGn angle [35], the sample was divided into three groups: normal growth cases (27° < SN-GoGn < 37°), open growth cases (SN-GoGn > 37°) and closed growth cases (SN-GoGn < 27°). The eruption space in the upper jaw was defined as the distance from the pterygoid vertical (PTV) to the distal surface of the upper first molar (M1) along the occlusal plane (PTV-M1). The eruption space in the lower jaw was defined as the distance from Ricketts’ Xi-point to the distal surface of the lower second molar crown along the occlusal plane (Xi-M2). Both measurements rely on the cephalometric analysis of Ricketts [36]. Additionally, the eruption space in the lower jaw was also scored on panoramic radiographs using the classification suggested by Pell & Gregory [37] (horizontal classification; stages 1, 2 and 3) (Fig. 2). The angulation of the third molars was scored on the panoramic radiographs, the upper third molar using the Archer’s classification [38] (Fig. 3) and the lower based on Winter’s classification [39] (Fig. 4). Additionally, the angle between the long axis of the second and third molar (M2^M3) was measured. In case the third molar had a distoangular inclination, the angle was taken as a positive value in the upper jaw and as a negative value in the lower jaw, whereas a mesioangular inclination was classified as a negative angle in the upper jaw and a positive angle in the lower jaw. The vertical position of the third molars compared to the adjacent second molar was also scored on the panoramic radiographs, in the upper jaw using Archer’s classification (Fig. 5) and in the lower jaw using the classification suggested by Pell & Gregory [37] (vertical classification; stages 1, 2 and 3) (Fig. 2). Furthermore, the relation between lower third molars and the mandibular canal was evaluated by the classification suggested by Whaites [40]. A close relationship between the roots of the lower third molar and the mandibular canal was assumed when one of the following landmarks were seen on the panoramic radiograph: loss of tramlines, narrowing of the tramlines, alteration of direction of the inferior canal at root apex, and a radiolucent band across the roots (Fig. 6). Finally, Demirjian’s classifications was used to examine the mineralization status of the third molars [41] (Fig. 7).

1-s2.0-S030057121730074X-gr1.jpg

Download high-res image (259KB)Download full-size image

Fig. 1. Cephalometric measurements to analyse the eruption space for the third molars (PTV-M1, Xi-M2) and the mandibular growth pattern (GoGn-SN).

1-s2.0-S030057121730074X-gr2.jpg

Download high-res image (515KB)Download full-size image

Fig. 2. Pell & Gregory’s classification for lower third molars. Horizontal classification; PGH-1: normal apical area, PGH-2: moderate apical area, PGH-3: small apical area. Vertical classification: PGV-1: the occlusal plane of the third molar is at the same level as the occlusal plane of the second molar, PGV-2: the occlusal plane of the third molar is located between the occlusal plane and the cervical margin of the second molar, PGV-3: the occlusal plane of the third molar is below the cervical margin of the second molar.

1-s2.0-S030057121730074X-gr3.jpg

Download high-res image (584KB)Download full-size image

Fig. 3. Archer’s classification of upper third molars according to their inclination to the long axis of the upper second molar. (1) mesioangular, (2) distoangular, (3) vertical, (4) horizontal, (5) buccoangular, (6) linguoangular, (7) inverted.

1-s2.0-S030057121730074X-gr4.jpg

Download high-res image (534KB)Download full-size image

Fig. 4. Winter’s classification: Third molars are classified according to their inclination to the long axis of the second molar. (1) vertical angulation, (2) horizontal angulation, (3) distoangular angulation, (4) mesioangular angulation, (5) transversal angulation, (6) inverse angulation.

1-s2.0-S030057121730074X-gr5.jpg

Download high-res image (356KB)Download full-size image

Fig. 5. Archer’s classification of upper third molars according to their vertical position compared to the adjacent second molar. (1) the occlusal surface of the third molar is at the same level as the occlusal surface of the second molar, (2) occlusal surface above the cementoenamel junction of the second molar, (3) occlusal surface at the same level of the cementoenamel junction, (4) occlusal surface underneath the cementoenamel junction, (5) occlusal surface above the apex of the second molar.

1-s2.0-S030057121730074X-gr6.jpg

Download high-res image (298KB)Download full-size image

Fig. 6. Whaites’ classification. Lower third molars are classified according to their position in relation to the mandibular canal. (1) normal relationship: tramlines across the root, (2) loss of tramlines, (3) narrowing of the tramlines, (4) alteration in direction of the mandibular canal at root apex, (5) radiolucent band across the roots.

1-s2.0-S030057121730074X-gr7.jpg

Download high-res image (429KB)Download full-size image

Fig. 7. Demirjian’s classification. Third molars are classified according to their developmental stage. (1) cusp tips are mineralized, (2) mineralized cusps are united, (3) crown is about half formed, (4) crown formation is complete, (5) formation of the inter-radicular bifurcation has begun and root length is less than the crown length, (6) root length is at least as great as crown length and roots have funnel-shaped endings, (7) root walls are parallel but apices remain open, (8) apical ends of the roots are completely closed.

2.3. Statistical analysis

All measurements were performed in a scoring program written in MATLAB™ which randomized the order of DICOM images and saved the results as comma separated value files [42]. This approach minimized bias, reduced the possibility of man made errors and meanwhile facilitated efficient data handling and statistical analyses.

A second observer randomly reassessed 20% of the radiographs for all mentioned classifications to determine the inter-observer variability, whereas the main observer also randomly reassessed 20% of the radiographs to determine intra-observer variability. Intra-class correlation (ICC) and the standard error of measurement (SEM) were calculated for the continuous measurements (PTV-M1, Xi-M2, M2^M3). Weighted kappa was used for the ordinal measurements and a simple kappa was calculated for the nominal measurements.

For the comparison of nominal, ordinal and continuous variables between patients treated with and without premolar extractions, Fisher’s exact tests and Mann-Whitney U tests were used. Associations between ordinal and/or continuous variables were evaluated with Spearman correlations. To evaluate the changes over time for the continuous measurement, a linear model for longitudinal measurements with an unstructured covariance matrix was used. Age and gender were added as confounders. The growth pattern was added as a time-varying factor in the analysis of Xi-M2. A similar approach is followed for the ordinal data and the nominal scores, but the linear model is replaced with a logistic regression model using generalized estimating equations (GEE) to handle the correlation between both time points and teeth. P-values smaller than 0.05 are considered statistically significant. All analyses have been performed using SAS software, version 9.4 of the SAS System for Windows. Copyright© 2016 SAS Institute Inc. SAS and all other SAS Institute Inc. product or service names are registered trademarks or trademarks of SAS Institute Inc., Cary, NC, USA.

This study was registered and approved by the medical ethics committee of the University Hospitals Leuven (registration number S56447).

3. Results

The descriptive data are summarized in Table 1. The orthodontic treatment took significantly more time for patients treated with extractions of premolars compared to patients treated without extractions (p = 0.011). The age at start and end of treatment was not significantly different between both groups (p = 0.791 and p = 0.143, respectively). General outcome information and pairwise comparisons for all mentioned classifications are summarized in Table 2.

Table 1. Sample distribution by gender, age, treatment duration and Angle classification. PM1: first premolar extraction; PM2: second premolar extraction.

		Non extraction (N = 218)	Extraction (N = 78; PM1 = 79, PM2 = 47)	p-value
Gender (n/N (%))	Male	103/218 (47.2)	37/78 (47.4)	0.977
Gender (n/N (%))	Female	115/218 (52.8)	41/78 (52.6)	0.977
Age pretreatment (years)	Mean	12.9	13.1	0.791
Age pretreatment (years)	Range	7.9–18.2	7.3–19.3	0.791
Age post-treatment (years)	Mean	15.5	16.0	0.143
Age post-treatment (years)	Range	12.2–20.3	12.5–21.3	0.143
Treatment (years)	Mean	2.6	2.9	0.011*
Angle classification (n/N (%))	Class I	92/218 (42.2)	29/78 (37.2)	0.657
	Class II	116/218 (53.2)	44/78 (56.4)
	Class III	10/218 (4.6)	5/78 (6.4)

Table 2. General outcome information and pairwise comparisons for all mentioned classifications.

Variable	Measurement	NE	PM1	PM2	Pairwise comparisons
					p-value	p-value	p-value
					NE vs PM1	NE vs PM2	PM1 vs PM2
PTV-M1	start of treatment
	mean (mm)	16.7	17	17.4
	median (mm)	16.3	17	17.6
	IQR (mm)	(14.6;19.1)	(15.0;19.2)	(16.8;18.8)
	>18 mm (%)	30.6	38.8	31.3	10.310	11.000	10.767
	end of treatment
	mean (mm)	18.8	22	22.5
	median (mm)	18.9	21.8	22.9
	IQR (mm)	(16.4;21.0)	(19.3;24.2)	(21.2;23.9)
	>18 mm (%)	58.9	83.7	100	1<0.001*	1<0.001*	10.184
	change				2<0.001*	2<0.001*	20.889
	mean (mm)	2.1	4.9	5.6
	median (mm)	1.7	4.5	5.1
	IQR (mm)	(−0.9;4.6)	(1.6;8.7)	(2.9;8.2)
	PTV-M1 at end >18 mm when at start <18 mm (%)	58.6	83.3	100	10.012*	10.007*	10.300
Xi-M2	start of treatment
	mean (mm)	17.8	16.4	17.2
	median (mm)	17.4	15.1	16.8
	IQR (mm)	(15.6;19.8)	(13.5;18.8)	(14.6;19.2)
	>25 mm (%)	2.6	0	0	11.000	11.000	11.000
	end of treatment
	mean (mm)	22.6	24.6	25.4
	median (mm)	22.7	24.1	25.1
	IQR (mm)	(20.0;24.6)	(21.8;26.6)	(22.5;27.9)
	>25 mm (%)	21.3	48	55.2	10.006*	1<0.001*	10.785
	change				2<0.0001*	2<0.0001*	20.869
	mean (mm)	5	7.9	6.9
	median (mm)	5.1	7.3	6.3
	IQR (mm)	(1.7;7.9)	(5.6;9.9)	(4.2;9.1)
	PTV-M1 at end >25 mm when at start <25 mm (%)	20.5	45.8	55.2	10.009*	10.001*	10.586
PGH	start of treatment
	PGH-1 (%)	2	6	3
	PGH-2 (%)	28	27	23
	PGH-3 (%)	70	67	74
	end of treatment
	PGH-1 (%)	23	47	77
	PGH-2 (%)	66	49	23
	PGH-3 (%)	11	4	0
	change				30.011*	3<0.0001*	30.001*
Archer (inclination)	start of treatment
	stage 1 (%)	8	9	16
	stage 2 (%)	16	8	7
	stage 3 (%)	75	83	74
	stage 4,5,6 and 7 (%)	1	0	3
	end of treatment
	stage 1 (%)	12	5	6
	stage 2 (%)	20	13	13
	stage 3 (%)	66	82	81
	stage 4,5,6 and 7 (%)	2	0	0
	change
	stage 1				30.094	30.056	30.671
	stage 2				30.562	30.547	30.835
	stage 3				30.404	30.262	30.651
Winter	start of treatment
	stage 1 (%)	20	25	20
	stage 2 (%)	8	20	7
	stage 4 (%)	71	55	73
	stage 3,5 and 6 (%)	1	0	0
	end of treatment
	stage 1 (%)	17	29	30
	stage 2 (%)	2	4	2
	stage 4 (%)	81	67	68
	stage 3,5 and 6 (%)	1	0	0
	change
	stage 1				30.260	30.102	30.570
	stage 2				30.889	30.840	30.790
	stage 4				30.900	30.085	30.228
M2^M3	UPPER JAW:
	start of treatment
	mean (°)	13.7	13.2	9.4
	median (°)	13.4	15.6	12.6
	IQR (°)	(6.2;22.9)	(2.9;22.2)	(5.5;17.7)
	end of treatment
	mean (°)	14.3	12.9	11.3
	median (°)	12.9	16.4	13.5
	IQR (°)	(2.7;23.4)	(4.4;25.3)	(2.4;19.1)
	change				20.708	20.715	20.574
	LOWER JAW:
	start of treatment
	mean (°)	25.7	26.6	26.5
	median (°)	24.9	24.1	25.7
	IQR (°)	(16.8;33.9)	(15.5;37.0)	(17.5;33.0)
	end of treatment
	mean (°)	26.4	26.6	23.2
	median (°)	26.3	27.2	25.1
	IQR (°)	(19.0;33.3)	(16.6;35.6)	(13.6;31.2)
	change				20.817	20.100	20.319
Archer (vertical)	start of treatment
	stage 1,2 and 3 (%)	5	8	6
	stage 4 and 5 (%)	95	92	94
	end of treatment
	stage 1,2 and 3 (%)	15	43	42
	stage 4 and 5 (%)	85	57	58
	change				30.030*	30.049*	30.542
PGV	start of treatment
	PGV-1 (%)	0	4	0
	PGV-2 (%)	1	0	0
	PGV-3 (%)	99	96	100
	end of treatment
	PGV-1 (%)	2	16	5
	PGV-2 (%)	10	10	33
	PGV-3 (%)	88	74	62
	Probability of PGV=3 at end (%)	89	74	62	30.024*	30.0001*	30.331
Whaites	positive relationship at start of treatment (%)	20	24	20
	positive relationship at end of treatment (%)	40	34	33
	change				30.236	30.444	30.726
Demirjian	start of treatment
	Stage 1,2,3 and 4 (%)	86	80	84
	Stage 4,5,6,7 and 8 (%)	14	20	16
	end of treatment
	Stage 1,2,3 and 4 (%)	30	29	28
	Stage 4,5,6,7 and 8 (%)	70	71	72
	change				30.319	30.179	30.579

NE: non extraction; PM1: first premolar extraction; PM2: second premolar extraction; PTV-M1: distance from pterygoid vertical to the distal surface of the upper first molar; Xi-M2: distance from Xi-point to the distal surface of the lower second molar; PGH: Pell & Gregory horizontal classification; M2^M3: angle between the long axis of second and third molar; PGV: Pell & Gregory vertical classification.

P-values based on Fishers exact test.

P-values based on linear model without corrections of confounders.

logistic regression model without correction of confounders.

p-values smaller than 0.05 are considered significant.

3.1. Intra-observer reliability

For the continuous measurements, ICC ranged between 0.92 and 0.97. For PTV-M1, the SEM equals 1.14 mm; for Xi-M2, the SEM equals 1.09 mm; for M2^M3, the SEM equals 2.76°. Weighted kappa was higher than 0.91 for the ordinal measurements (i.e. PGH, PGV, Archer classification (vertical position upper third molar), Demirjian classification). For the nominal measurements (Archer classification (inclination of upper third molar), Winter classification, Whaites classification), simple kappa was higher than 0.92.

3.2. Inter-observer reliability

For the continuous measurements, ICC ranged between 0.69 and 0.92. For PTV-M1, the SEM equals 2.17 mm; for Xi-M2, the SEM equals 1.33 mm; for M2^M3, the SEM equals 5.64°. Weighted kappa was higher than 0.81 for the ordinal measurements. For the nominal measurements, simple kappa was higher than 0.87.

3.3. Eruption space for third molars

In the upper jaw, PTV-M1 values were higher for older patients (Spearman rho = 0.36, p < 0.0001) and for males (p = 0.028) at the start of treatment. As can be derived from Table 2, the mean values of PTV-M1 at the start of treatment were respectively 16.7 mm, 17.0 mm and 17.4 mm for patients treated without extractions, extraction of the first premolar and extraction of the second premolar; whereas at the end of treatment these values were respectively 18.8 mm, 22.0 mm and 22.5 mm. The linear model revealed a significant higher increase in PTV-M1 values for patients with premolar extractions compared to patients treated without extractions (p < 0.0001). The increases equal 2.10 mm (95%CI: 1.77 to 2.44, p < 0.0001), 5.02 mm (95% confidence interval (CI): 4.31 to 5.73, p < 0.0001) and 5.12 mm (95%CI: 3.88 to 6.36, p < 0.0001), respectively. The difference in change between extraction of first and second premolar was not statistically significant (p = 0.89). Also after correction for age and gender the change in PTV-M1 differs significantly between patients treated with and without premolar extractions (p < 0.0001). After correction of these confounders, the changes equal 0.64 mm (95%CI: 0.08 to 1.21, p = 0.03), 3.29 mm (95%CI: 2.40 to 4.18, p < 0.0001) and 3.77 mm (95%CI: 2.47 to 5.07, p < 0.0001) for patients treated without extractions, extractions of first and second premolars respectively. The results derived from the linear model for PTV-M1 are summarized in Table 3 and graphically presented in Fig. 8.

Table 3. Detailed results from the bivariate linear model for PTV-M1 and Xi-M2. Age and gender were added as confounders.

	measurement	Outcome (1 = in change start-end)	Without correction of confounders		With correction of confounders
			estimate (CI)	p-value	estimate (CI)	p-value
UPPER JAW	PTV-M1	INTERACTION EFFECT NE1 PM11 PM21 Δ NE-PM11 Δ NE-PM21 Δ PM1-PM21	2.10 (1.77;2.44) 5.02 (4.31;5.73) 5.12 (3.88;6.36) 2.92 (2.13;3.70) 3.02 (1.73;4.31) 0.10 (−1.33;1.53)	<0.0001* <0.0001* <0.0001* <0.0001* <0.0001* <0.0001* 0.889	0.64 (0.08;1.21) 3.29 (2.40;4.18) 3.77 (2.47;5.07) 2.65 (1.86;3.43) 3.13 (1.85;4.41) 0.48 (−0.94;1.91)	<0.0001* 0.0261* <0.0001* <0.0001* <0.0001* <0.0001* 0.507
LOWER JAW	Xi-M2	INTERACTION EFFECT NE1 PM11 PM21 Δ NE-PM11 Δ NE-PM21 Δ PM1-PM21	4.85 (4.47;5.23) 8.15 (7.00;9.29) 8.28 (7.21;9.34) 3.30 (2.09;4.50) 3.43 (2.30;4.56) 0.13 (−1.44;1.70)	<0.0001* <0.0001* <0.0001* <0.0001* <0.0001* <0.0001* 0.869	2.55 (1.96;3.14) 5.38 (4.13;6.64) 5.91 (4.76;7.06) 2.83 (1.65;4.02) 3.36 (2.25;4.47) 0.53 (−1.01,2.06)	<0.0001* <0.0001* <0.0001* <0.0001* <0.0001* <0.0001* 0.502

p-values smaller than 0.05 are considered significant.

1-s2.0-S030057121730074X-gr8.jpg

Download high-res image (168KB)Download full-size image

Fig. 8. Results of a bivariate regression model for PTV-M1 without (left panel) and with (right panel) correction for confounders.

The eruption space for lower third molars was measured on lateral cephalograms (Xi-M2) and panoramic radiographs (PGH). Both measurements revealed a statistically significant increase of eruption space during treatment and this increase was statistically higher for patients treated with extractions of premolars (p < 0.0001). The results derived from the linear model for Xi-M2 are summarized in Table 3 and graphically presented in Fig. 9. Without correction for age and gender the increase of Xi-M2 equals 4.85 mm (95%CI: 4.47 to 5.23, p < 0.0001), 8.15 mm (95%CI: 7.00 to 9.29, p < 0.0001) and 8.28 mm (95%CI: 7.21 to 9.34, p < 0.0001) for patients treated without extractions, extractions of first and second premolars, respectively. After correction for age and gender, the changes for Xi-M2 equal respectively 2.55 mm (95%CI: 1.96 to 3.14, p < 0.0001), 5.38 mm (95%CI: 4.13 to 6.64, p < 0.0001) and 5.91 mm (95%CI: 4.76 to 7.06, p < 0.0001). In accordance with the results for PTV-M1, there was no difference for Xi-M2 between patients treated with extractions of first and second premolars (p = 0.502). When considering the growth pattern, the increase of Xi-M2 values was higher in patients with an open growth pattern compared to patients with a closed growth pattern (1.93 mm (95% CI: 0.88 to 2.99, p = 0.0004)). The findings for Xi-M2 are confirmed by the results of the Pell & Gregory Horizontal classification (PGH) except for the difference in change between first and second premolar extractions, where PGH shows a significant difference (p = 0.001). After correction for age and gender, the probability of having moderate (PGH-2) or small (PGH-3) apical areas at the end of treatment, equals 86.7%, 70.5% and 34.1% in patients treated without extractions, extractions of first and second premolars, respectively.

1-s2.0-S030057121730074X-gr9.jpg

Download high-res image (190KB)Download full-size image

Fig. 9. Results of a bivariate regression model for Xi-M2 without (left panel) and with (right panel) correction for confounders.

3.4. Third molar orientation

The orientation of the third molar was scored on panoramic radiographs. M2^M3 changed throughout treatment, however this change was statistically insignificant. M2^M3 showed a large spread both at the start of treatment as well as at the end of treatment (Table 2). The results of the linear model revealed that the change in M2^M3 did not significantly differ between patients treated with and without premolar extractions (p = 0.63 and p = 0.65, respectively with and without corrections for confounders). The same conclusions were drawn after statistically analyzing the outcome of the Winter’s classification and the classification suggested by Archer (p > 0.24 and p > 0.29).

3.5. Vertical position of third molars

In the upper jaw, the vertical position of the third molars was evaluated using the classification suggested by Archer [38]. For the statistical analysis, only subjects with fully erupted second molars at the start of treatment were considered. At the start of treatment, the third molar was situated under the level of the cementoenamel junction of the second molar in 95%, 92% and 94% of patients treated without extractions, and with first and second premolar extractions, respectively. Whereas at the end of treatment these percentages reduced to respectively 85%, 57% and 58%. After correction for age and gender, the change over time differed significantly between groups (p = 0.047).

In the lower jaw, the vertical position of the third molars was evaluated using the classification of Pell & Gregory (PGV). The probability of having a PGV-score of 3 at the end of treatment, equals 88.7%, 73.5% and 61.7% in patients treated without premolar extractions, and with first and second premolar extractions, respectively. This probability is significantly higher in the non-extraction group compared to the group of patients with an extraction of the first premolar (odds ratio (OR) = 2.841 (95%CI: 1.15 to 7.04, p = 0.024)) and compared to the group of patients with an extraction of the second premolar (OR = 4.89 (95%CI: 2.16 to 11.09, p = 0.0001)). There was no statistical significant difference between patients treated with first or second premolar extractions (p = 0.331).

3.6. Relation between lower third molar and alveolar nerve

At the start of treatment, in approximately 20% of the patients a relationship between the third molar and the alveolar nerve was noticed in all of the groups. At the end of treatment, a close relationship between the nerve and the third molar was seen in 40%, 34% and 33% of patients treated without extractions, extraction of the first premolar and extraction of the second premolar, respectively. The logistic regression model revealed no statistically significant difference between patients treated with and without extractions with regard to the relation between the lower third molar and the alveolar nerve (p = 0.38).

3.7. Third molar development

The results of the Demirjian classification revealed no significant difference between patients treated with or without premolar extractions (p = 0.39). There was a significant increase in development during treatment (p < 0.0001), but this increase did not differ between both groups (p = 0.342).

4. Discussion

This study aimed at investigating the effect of first and second premolar extractions on the position and space available for upper and lower third molars. To investigate these possible effects, we combined both panoramic and lateral cephalograms taken before and at the end of orthodontic treatment. The eruption space for upper and lower third molars was analyzed on lateral cephalograms by measuring PTV-M1 and Xi-M2. Both measurements were applied by other researchers in the past [5], [10], [26], [27]. Behbehani et al. reported that measuring Xi-M2 has lower method errors than measuring eruption space from the distal border of the second molar to the anterior border of the mandibular ramus [5]. Furthermore, by only including patients with symmetrical extractions, we minimized the risk on errors due to overlap of the right and left teeth on lateral cephalograms. The angulation, vertical position, relation between the lower third molar and the mandibular canal and the mineralization status of the third molars were evaluated on panoramic radiographs. Although the use of panoramic radiographs has been criticized because of distortions and magnifications [1], several authors reported that angular measurements as well as the relationship between the third molar and the mandibular canal can be accurately analyzed in panoramic radiographs [43], [44].

Firstly, the effect of premolar extractions on the retromolar space was investigated. The results revealed that in both the upper and lower jaw, the increase in retromolar space is significantly higher in patients treated with premolar extractions compared to patients treated without premolar extractions. This increase in space can be explained by the mesial movement of the first and second molars during space closure. An earlier study reported, based on the Pell & Gregory classification, that in only 35.6% of patients treated without premolar extractions the eruption space for the lower third molar was sufficient (PGH-1) at the end of treatment, whereas in patients treated with premolar extractions the eruption space was sufficient in 55.6% of the cases [45]. In the current study, PGH-1 was noticed in 23%, 47% and 77% at the end of treatment in patients treated without premolar extractions, and with first and second premolar extractions, respectively. Kim et al. reported that PTV-M1 was 3 mm larger and Xi-M2 was 2.6 mm larger in patients treated with extractions of premolars compared to patients treated without extractions [10], matching our results (Table 2). In a group of non-growing subjects, Patel et al. [26] found a change of 4.47 mm for Xi-M2 in patients treated with first premolar extractions, compared to 1.93 mm in patients treated without extractions. The larger change in Xi-M2 seen in the current study can be explained by the fact that only growing patients were included, whereas Patel et al. only included non-growing patients [26]. It is important to mention that, although our results suggest a positive influence of premolar extraction on third molar impaction, no conclusions can be drawn with regard to the minimum retromolar space needed for predictable eruption since only radiographs of growing patients were used. However, several authors have tried to make a model to predict the eruption of third molars. Schulhof stated that third molar impaction was more likely to occur when Xi-M2 decreased below 25 mm and PTV-M1 decreased below 18 mm [46]. In the present study, in respectively 78.7%, 52.0% and 44.8% of patients Xi-M2 was smaller than 25 mm and in respectively 41.1%, 16.3% and 0.0% PTV-M1 was smaller than 18 mm at the end of treatment. The difference between patients treated with and without premolar extractions was significant, whereas the difference between first and second premolar extraction was not statistically significant. Note that, in accordance with these results, third molars in the upper jaw seem to have more chance to erupt after premolar extractions compared to third molars in the lower jaw although the rate of eruption was not investigated in this study as it requires a longer follow up. However, the data of Kim et al. indicated that the difference in impaction rate between both extraction and non-extraction groups was similar for upper and lower jaw [10]. Furthermore, the same authors questioned the predictive value of 18 mm and 25 mm suggested by Schulhof. 20% of their sample experienced impaction despite a distance of 18 mm or more. Besides, more than 60% of the patients in their sample with less than 23 mm for Xi-M2 experienced eruption of the lower third molars. Artun et al. [27] and Behbehani et al. [5] tried to identify risk factors for upper and lower third molar impaction by studying radiographs of orthodontic patients made before, after and at a minimum of 10 years post-retention. They reported that the decision to extract premolars in the upper jaw reduced the risk of impaction for the upper third molars by 76%, whereas extractions in the lower jaw reduced the risk by 63%. Furthermore, the study of Behbehani et al. [5] revealed that, in accordance with other authors [6], [12], [47], [48], forward mandibular growth rotation increased the risk of impaction. Likewise, our results revealed that the retromolar space was significantly larger in patients with an open growth pattern compared to patients with a closed growth pattern (p = 0.0004). Although Celikoglu [23] and Ong et al. [49] reported that extraction of second rather than first premolars appeared to be more favorable for third molar eruption, our results did not indicate a difference in eruption space for upper third molars between patients treated with first or second premolar extraction. For lower third molars our results were inconclusive, since the results of Xi-M2 did not show a difference between first and second premolar extractions whereas PGH did. The lack of a clear difference between first and second premolar extraction can be explained by the fact that the choice for a certain extraction pattern depends on multiple factors such as the amount of crowding, the need for incisor retraction, the underlying malocclusion and individual tooth conditions such as tooth decay, abnormal morphology or impaction. It is reasonable that these underlying factors could play a more crucial role in the final gain in retromolar space than the extraction pattern itself. Unfortunately, due to the retrospective character of this study, these factors could not be taken into account.

Secondly, the possible change in angulation of upper and lower third molars was investigated in all of the groups. The angulation changes over time, but there was no significant difference between patients treated with or without extractions. The findings of several studies are in accordance with our results [4], [33], [34], [50], [51]. Tarazona et al. investigated the influence of first and second premolar extraction on the angulation of lower third molars and concluded that the angulation of the third molars improves over time, regardless of treatment with or without premolar extractions [51]. In a retrospective study of 44 non-growing subjects with 22 patients treated with extractions of the first premolars, Türkoz found no significant change in angulation of the lower third molar during orthodontic treatment. However, Pearson’s correlation between this angle and eruption revealed a significant correlation between eruption of the lower third molar and pre- and posttreatment angulation of the lower third molar, indicating that the inclination of the third molar is strongly correlated with impaction [4]. Similar to our study, Russell et al. distinguished between first and second premolar extraction. They concluded that the changes in third molar angulation during orthodontic treatment did not significantly differ between patients without extractions, first premolar extraction and second premolar extraction because of a wide range of angulations and changes within each group [34]. However, in contrast to our results, several authors reported that extractions of premolars improved the angulation of the third molars [10], [31], [32], [52]. For example, Saysel et al. investigated the effects of first premolar extractions on the angulation of third molars by measuring M2^M3. The angulation of the lower third molar improved significantly more in patients treated with premolar extractions, whereas for the upper third molar no significant difference was found [52].

Thirdly, our study aimed at investigating the effect of premolar extraction on the vertical development of third molars. The results revealed that at the end of treatment significantly less third molars of patients treated with premolar extractions were situated under the cementoenamel junction of the second molar. Although there are very little studies to compare our results with, the findings of Elsey et al. [25] are in accordance with our results. They reported that after extractions of first premolars, the lower third molar significantly became closer to the occlusal plane compared to a group of patients who had not received any orthodontic treatment [25].

An additional aim was to investigate the relationship between the lower third molar and the inferior alveolar nerve. The results indicated that after treatment more lower third molars are in close contact with the alveolar canal, which can be explained by the growth and development of lower third molar roots in growing patients, as is the case for our sample. Although a close relationship was seen in 40% of patients treated without extractions compared to only 34% and 33% of patients treated with first or second premolar extractions, the logistic regression model revealed no significant difference between groups (p = 0.342). In contrast to our findings, a previous study showed that in patients treated without extractions, significantly more third molars were in close relation with the mandibular canal in comparison with patients treated with extractions of premolars [45]. They reported that at the end of treatment only 20.4% of the lower third molars were in close contact with the alveolar canal in the extraction group and 32.1% in the non-extraction group.

The fact that we only used pre- and posttreatment radiographs of growing patients is an important limitation of our study. Because of this, the time of follow-up was restrained to approximately 2.9 years. Furthermore, most of third molars are still in a developing stage at the end of treatment. A second limitation is that, due to the retrospective character of this study, we did not have detailed information on the orthodontic biomechanics applied to close the extraction spaces. Additionally, it must be emphasized that retrospective studies are typically more sensitive to overestimating treatment effects due to confounding and selection bias. We tried to minimize the confounding in this study by excluding known confounders such as age, gender and the effect of distalization appliances. However, as mentioned earlier in this paper, it is most likely that factors such as molar anchorage, the amount of initial crowding and the final position of the incisors could play an important role in the final gain in retromolar space. Therefore, future research should take into account these factors. At the same time a longer follow-up is necessary to draw conclusions on possible effects on impaction as well as to account for eruption rates and final angulation of the third molars. This study can be seen as an overview of the effects of premolar extractions on both upper and lower third molars, and is therefore a good start for future prospective research.

5. Conclusions

During orthodontic treatment in growing patients the retromolar space and the third molars experience important changes, not only influenced by natural growth but also by the type of orthodontic treatment. Based on the results of this retrospective study, the following conclusions can be made:

–

Extractions of both first and second premolars have a positive influence on the eruption space and vertical position of the upper and lower third molars.

–

The fact that this study did not reveal clear differences in effects between first and second premolar extractions suggests that the applied orthodontic biomechanics to close the extraction spaces could play an important role in the final gain in retromolar space.

–

The eruption space for the lower third molars increases significantly in patients with an open growth pattern.

–

The change in angulation of upper and lower third molars during orthodontic treatment did not significantly differ between patients treated with and without premolar extractions.

–

Although a relationship between the lower third molars and the mandibular canal is observed more frequently in patients treated without premolar extractions, the difference relative to the non-extraction group was not statistically significant.

In conclusion the results of the present study highlight the importance of taking third molars into account during orthodontic treatment planning. Due to the retrospective character of the study, the conclusions should be carefully reflected. Further prospective research is necessary to give further insight into this important but complicated topic.

Conflict of interest

None.

Acknowledgements

The authors would like to thank Dr. Steffen Fieuws for performing the statistical analyses, Steven Lauwereins for developing the measurement software and Prof. Dr. Ali Alqerban for comments that greatly improved the manuscript.

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

References

[1]

J.B. Fayad, J.C. Levy, C. Yazbeck, R. Cavezian, E.A. Cabanis

Eruption of third molars: relationship to inclination of adjacent molars

Am. J. Orthod. Dentofac. Orthop., 125 (2004), pp. 200-202, 10.1016/j.ajodo.2003.10.010

Article Download PDF View Record in Scopus

[2]

F.N. Hattab, M.A. Rawashdeh, M.S. Fahmy

Impaction status of third molars in Jordanian students

Oral Surg. Oral Med. Oral Pathol. Oral Radiol., 79 (1995), pp. 24-29, 10.1016/S1079-2104(05)80068-X

Article Download PDF View Record in Scopus