Mandibular bone changes in children with ADHD: evaluation using fractal analysis and mandibular indices

Abstract

Background

Previous studies have shown that abnormalities in the oxytocin system in children with attention deficit hyperactivity disorder (ADHD) may affect bone mineral density, and that medications used to treat ADHD may cause growth suppression. This study aimed to evaluate bone structure and microarchitecture in children diagnosed with ADHD using mandibular indices and fractal dimension analysis as well as to compare these results with those of healthy children.

Methods

The study included 61 children aged 4–15 years who had been diagnosed with ADHD, along with 122 healthy children matched for age and gender. Measurements of the mandibular cortical index (MCI), mental index (MI) and panoramic mandibular index (PMI) were obtained from panoramic radiographs of both the study and control groups, and fractal dimension (FD) analysis was performed in regions of interest (ROI) corresponding to the condyle, angulus and interdental areas. All measurements were performed on digital images using the ImageJ programme (ImageJ 1.38, NIH, Maryland, USA). Data were analysed using SPSS 27. A significance level of p < 0.05 was accepted.

Results

Significant differences were observed for interdental FD and for all mandibular indices (PMI, MI, and MCI). İnterdental FD measurements were higher in the study group (ADHD group) than in the control group (healthy group). Type 1 MCI was observed more frequently in the study group, while Type 2 MCI was observed more frequently in the control group. According to the PMI and MI analyses, the study group’s measurements were higher than those of the control group’s. There were no significant differences in condyle FD and angulus FD values between the study groups (ADHD and healthy control group).

Conclusion

The differences observed in mandibular indices and interdental FD are thought to reflect region-specific changes. The potential influence of parafunctional habits cannot be confirmed without clinical evaluation.

Background

Attention deficit hyperactivity disorder (ADHD) is a neurodevelopmental disorder characterised by persistent patterns of inattention, hyperactivity, and impulsivity. More common in males, this disorder is becoming increasingly prevalent and affects approximately 5–10% of children worldwide, making it one of the most common childhood psychiatric disorders [1, 2]. Stimulant (e.g., methylphenidate and amphetamine) and non-stimulant (e.g., guanfacine and atomoxetine) medications are commonly used to reduce symptoms by increasing synaptic dopamine and norepinephrine levels, with psychostimulants being the primary pharmacological treatment option [3].

Changes in bone mineral density (BMD) and bone mineral content (BMC) during childhood have long-term effects on skeletal health, particularly regarding bone strength and integrity [4]. The impact of psychostimulant drugs on bone health is an area of ongoing research, and the specific mechanisms are not yet fully understood [2]. The use of psychostimulant medications has been associated with lower serum levels of vitamin D and magnesium. It has also been reported that these medications may cause almost complete growth arrest by reducing growth hormone levels [5, 6]. Since bone mass accumulation continues until early adulthood, evaluating the possible effects of stimulant medications on bone density in growing children is of particular importance [7]. Dental imaging obtained during routine dental visits may enable early detection of mandibular bone alterations and support timely referral and preventive management.

Mandibular indices derived from panoramic radiographs are useful parameters for assessing bone loss and for predicting possible decreases in bone mineral density [8]. Fractal analysis (FA) is a mathematical technique that allows quantitative characterization of complex structures that cannot be described by integral dimensions; the resulting fractal dimension (FD) acts as a numerical indicator [9]. By applying FA, the irregular trabecular architecture of bone can be examined, providing insight into bone quantity. Previous studies have reported that FD values derived from radiographs are associated with variations in bone density and may reflect mineral loss in osseous tissue [10].

Our study aimed to measure bone quantity and microarchitecture in children diagnosed with ADHD, using mandibular indices and fractal dimension analysis and to compare these results with those obtained from healthy children. The null hypothesis was that there would be no significant difference in bone quantity and microarchitecture between children with ADHD and healthy controls.

Methods

Study design

This retrospective, single-center cross-sectional study was conducted in accordance with the Declaration of Helsinki and received approval from the Non-Interventional Clinical Research Ethics Committee of Bolu Abant Izzet Baysal University (2024/176).

The required sample size was calculated using G*Power 3.1.9.2 software [11]. Based on a previously published study [12], the standardized effect size (PMI-s) was determined as 0.7242 at a 95% confidence level (α = 0.05). A minimum of 31 participants per group was required to achieve 80% statistical power. To strengthen the reliability of the results, the ADHD group comprised 61 children, while 122 age- and gender-matched individuals formed the control group.

The study sample consisted of patients who presented to the Department of Paediatric Dentistry between 2019 and 2024 and underwent digital panoramic radiography (DPR) for diagnostic purposes. Children aged 4–15 years with ADHD confirmed in the hospital database and with DPRs of adequate quality were included. Controls were systemically healthy individuals not receiving medication, matched for age and gender. Exclusion criteria included maxillofacial cysts/tumours, systemic conditions affecting bone metabolism, orthodontic treatment, mandibular trauma or surgery, and poor image quality.

Image analysis

The DPRs used in this retrospective study were acquired using the same Soredex Cranex Novus panoramic unit (Tuusula, Finland) by the same operator under standardized exposure parameters (70 kVp, 10 mA, 8 s, magnification ×1.34). The images were initially downloaded in JPEG format via the MetaSoft System (Planmeca, Helsinki, Finland) and saved with a resolution of 2441 × 1149 pixels, 300 dpi, and 8-bit color depth. Image measurements were performed by a single examiner (S. Ç.) using ImageJ software (v1.38; NIH, Maryland, USA) on a 15.6-inch, 1080p display.

Fractal analysis

Fractal dimension (FD) calculations followed the box-counting method as described by White and Rudolph [13]. Regions of interest (ROIs) were placed bilaterally in the mandibular condyle, angulus, and interdental regions, with each ROI sized at 30 × 30 pixels. Specifically, in the mandibular condyle, ROIs were positioned in the central point of the trabecular bone; in the angulus region, ROIs were located over trabecular bone at the angle corner; and in the interdental region, ROIs were taken strictly from trabecular bone between the first molar and second premolar, avoiding other anatomical structures (Fig. 1). For fractal analysis, each ROI was extracted and duplicated. A Gaussian filter was applied to the duplicated image to suppress small and medium-sized structures, leaving only large density differences. The blurred image was then subtracted from the original image. The resulting image was thresholded at a gray value of 128 using the Threshold option, converting it into a binary black-and-white image. Noise was reduced using the Erode option, and the Dilate option was applied to enlarge the existing structures. The Invert option converted white areas to black and black areas to white. Skeletonization was then performed to erode the image until only the central pixel lines remained, outlining the trabecular bone. Finally, the FD value, representing the degree of structural complexity, was calculated as the negative slope of the linear regression line generated using the Fractal Box Count option. (Fig. 2).

Fig. 1.

Steps of fractal analysis. (a) ROI selection, (b) Gaussian filter–based blurring, (c) subtraction of the ROI from the original image, (d) adjustment of pixel values by adding 128 Gy levels, (e) binarization, (f) erosion, (g) dilation, (h) inversion, (i) skeletonization

Fig. 2.

Illustration of six chosen ROIs on panoramic radiographs

Mandibular indices

Measurements were performed using the mandibular cortical index (MCI), mental index (MI) and panoramic mandibular index (PMI) from among the mandibular indices. Analyses including MI and PMI were performed separately on digital panoramic images of the left and right sides of the mandible. Data were recorded by calculating the arithmetic mean of measurements from both sides. A panoramic radiograph of a patient with a measured mesiodistal width of the mandibular central incisors was used for calibration.

The mandibular cortical index, as described by Klemetti et al. [14] (Fig. 3), the mental index, as described by Ledgerton et al. [15] (Fig. 4), and the panoramic mandibular index, as described by Benson et al. [16] (Fig. 4), were assessed, with detailed procedures described in our previous publication [17].

Fig. 3.

Radiographic examples of mandibular cortical index categories. (a) C1, (b) C2, (c) C3

Fig. 4.

Demonstration of panoramic mandibular index and mental index measurements

Statistical analysis

Descriptive statistics, including counts, percentages, means, standard deviations, medians, and 25th and 75th percentiles, were calculated for all variables. The Shapiro–Wilk test was used to assess the normality of continuous data. For variables that did not meet the normality assumption, comparisons between two independent groups were performed using the Mann–Whitney U test. Spearman’s correlation analysis was conducted to explore associations between continuous variables with non-normal distributions. For categorical variables, the Pearson Chi-square test was used when expected cell counts were greater than five; otherwise, Fisher’s Exact test was applied. Additionally, multiple Chi-square tests were conducted for participants taking multiple medications. All analyses were carried out using IBM SPSS Statistics version 27.

Results

Table 1 summarized the distribution of demographic and medication use characteristics according to study group, and Pearson Chi-square and multiple Chi-square tests were applied to examine the relationships between them. There were a similar gender distribution in both groups (p > 0.05): 80.3% of individuals in the study group and 79.5% of individuals in the control group were male, while 19.7% and 20.5% were female, respectively. Examining the distribution by drug type revealed that none of the individuals in the control group used drugs (100.0%). In the study group, 26.2% did not use drugs; 65.6% used stimulants; and 73.8% used non-stimulant drugs. Some participants were receiving more than one class of medications; therefore, the total percentage of stimulant and non-stimulant users exceeded 100%. Regarding medication duration, 46.7% of individuals in the study group had taken medication for less than one year, 31.1% for one to five years, and 22.2% for more than five years. There were no individuals using medication in the control group. The analyses revealed a statistically significant difference in medication type between the study groups, while no significant difference was observed in gender distribution (p > 0.05).

Table 1.

Distribution of demographic and medication use characteristics by study group and relationships between them

	Category	Study group		Control group
		n	%	n	%	Test statistic	p	Effect size
Gender	Male	49	80.3	97	79.5	0.02	0.896	-
	Female	12	19.7	25	20.5
Type of medicine	Not using	16	26.2	122	100.0	341.07**	< 0.001*	0.86
	Stimulant	40	65.6	0	0.0
	Non-stimulant	45	73.8	0	0.0
Duration of drug use	< 1 year	21	46.7	0	0.0	-	-	-
	1–5 year	14	31.1	0	0.0
	> 5 year	10	22.2	0	0.0

*p < 0.05

**Multiple Chi Square test and %: Column percentage

Table 2 summarized the age and FD measurements for each group, with Mann–Whitney U tests were used for group comparisons. The mean age of the study group was 10.19 ± 2.92 years, and that of the control group was 10.17 ± 2.86 years. There was no significant difference between the groups (p > 0.05). A statistically significant difference was observed in the interdental FD measurements between the groups (p < 0.05), with higher values in the study group. No significant differences were detected in the condyle and angulus FD measurements (p > 0.05).

Table 2.

Distribution and comparison of age and FD measurements according to study groups

	Study group		Control group
	Mean ± SD	Median (25th–75th percentile)	Mean ± SD	Median (25th–75th percentile)	Test statistic	p	Effect size
Age	10.19 ± 2.92	10.19(8.08–12.48)	10.17 ± 2.86	10.17(8.04–12.53)	−0.15	0.878	-
Condyle	1.36 ± 0.06	1.36(1.32–1.41)	1.34 ± 0.07	1.35(1.30–1.39)	−1.96	0.050	-
Angulus	1.31 ± 0.08	1.31(1.26–1.35)	1.30 ± 0.08	1.31(1.25–1.36)	−0.33	0.740	-
Interdental	1.31 ± 0.08	1.32(1.25–1.37)	1.27 ± 0.09	1.29(1.22–1.34)	−2.39	0.017*	0.41

*p < 0.05

Table 3 summarized the distribution of MCI types across the groups, which was analyzed using Fisher’s Exact test. A significant association was found between the groups and MCI classifications (p < 0.05). In the literature, Type 1 has been associated with high cortical bone quality, whereas Type 3 has been linked to an increased risk of osteoporosis [14]. In the study group, 27.9% of participants had Type 1, 45.9% had Type 2, and 26.2% had Type 3 MCIs, whereas in the control group, 13.1% had Type 1, 70.5% had Type 2, and 16.4% had Type 3 MCIs. The distributions of Types 1 and 2 differed significantly between groups: Type 1 was more common in the study group, while Type 2 predominated in the control group. No significant difference was observed for Type 3 MCIs.

Table 3.

Distribution of MCI groups according to study groups and the relationships between them

MCI	Study group		Control group
	n	%	n	%	Test statistic	p	Effect size
Type 1	17a	27.9	16b	13.1	10.76	0.004*
Type 2	28a	45.9	86b	70.5			0.24
Type 3	16a	26.2	20a	16.4

*p < 0.05, %: Column percentage and different letters indicate different column rates based on rows

Table 4 summarized PMI and MI measurements by group. Comparisons using the Mann–Whitney U test indicated statistically significant higher values for both PMI and MI in the study group compared to the control group (p < 0.05). Within the study group, no significant correlations were observed between the duration of drug therapy and the evaluated radiographic parameters (Table 5).

Table 4.

Distribution and comparison of PMI and MI measurements according to study groups

	Study group		Control group
	Mean ± SD	Median (25th–75th percentile)	Mean ± SD	Median (25th–75th percentile)	Test statistic	p	Effect size
PMI	0.27±0.06	0.27(0.23–0.31.23.31)	0.22±0.04	0.21(0.18–0.25.18.25)	−5.89	<0.001*	1.08
MI	6.33±1.26	6.17(5.46–7.07.46.07)	5.63±1.01	5.52(4.95–6.34.95.34)	−3.59	<0.001*	0.63

*p<0.05

Table 5.

Relationships between medication duration and clinical measurements in the study group

		Study group
		Duration of drug use
Condyle FD	r	0.11
	p	0.46
Angulus FD	r	−0.06
	p	0.68
Interdental FD	r	−0.01
	p	0.94
MCI	r	0.23
	p	0.13
MI	r	0.07
	p	0.66
PMI	r	−0.08
	p	0.60

*p<0.05 and r: correlation coefficient

Discussion

Inadequate bone development during childhood is associated with reduced peak bone mass and a higher risk of osteoporosis in adulthood [18]. For pediatric practitioners, early recognition of altered bone development is important due to its potential lifelong skeletal consequences In dental practice, this has significant implications, as low bone density increases the risk of fractures following trauma and necessitates careful planning to avoid of potential bone fracture during surgical procedures. Additionally, evaluating bone structure in orthodontic procedures helps predict biomechanical responses [19]. Detecting bone changes on panoramic radiographs supports treatment planning and, when necessary, enables timely referral for further evaluation. In this study, FA and mandibular indices were used as non-invasive tools to obtain quantitative data and detect early changes in bone structure in children with ADHD. There are few studies [20] in the literature on mandibular bone quantity and microarchitecture in children with ADHD, and our study is pioneering in terms of the parameters used.

ADHD medications have been suggested to contribute to bone loss and increase fracture risk [20, 21]. However, no clear consensus exists in the literature regarding the effects of ADHD medications on bone health. Perry et al. [21] reported that treatment of ADHD with either stimulant or non-stimulant medications is associated with a reduced risk of fractures in individuals under 40 years of age. In a study conducted by Lahat et al. [22], 10 boys aged 3–10 years who were treated with methylphenidate for 1–2 years showed no statistically significant difference in bone turnover markers and BMD compared to healthy children. In a study conducted by Feuer et al. [7], it was reported that children and adolescents aged 8 to 20 years treated with psychostimulants had lower bone mass, mean lumbar spine BMD, bone mineral content, femoral neck bone mineral content, and total femur BMD compared to children and adolescents who were not treated with psychostimulants.

Recent data demonstrate the anabolic effect of oxytocin on bone. It increases bone formation and improves bone microarchitecture by supporting the differentiation and function of osteoblasts. It plays a role in bone formation in both sexes during development [23]. Oxytocin promotes osteoclast formation by activating NF-κB and MAP kinase pathways and increasing RANK-L expression, but it also suppresses bone resorption by mature osteoclasts through Ca²⁺ release and nitric oxide production [24]. In a study of children diagnosed with ADHD, serum oxytocin levels were found to be significantly lower in the study group than in the control group. When children with ADHD who were taking medication with those who were not, it was found that oxytocin levels were significantly higher in the group taking medication [25]. Therefore, our study included both children with ADHD who were receiving medication and those who were not.

Mandibular indices are used to assess mandibular morphological changes and bone volume on panoramic radiographs [26]. However, they do not provide a direct measurement of bone density [27]. Some studies have suggested that Type 3 MCI may serve as an indicator of low BMD [14, 28]. In our study, Type 3 MCI was similar between the two groups. The observed differences between Type 1 and Type 2 MCI may be attributed to the fact that this index is based on visual assessment rather than quantitative measurement, and its classification can vary depending on the examiner’s experience.

The mental index (MI) is a measure of the thickness of the mandibular cortical bone below the mental foramen. Several studies have reported that individuals with osteoporosis have lower MI and PMI values compared to healthy controls [29–32]. Kostiner et al. [20] compared the MI values measured from the DPRs of 38 children and adolescents with ADHD who had been treated with methylphenidate for at least 12 months with those of 58 healthy children and adolescents, and reported that the mean MI was significantly lower in the ADHD group. In contrast, in our study, MI and PMI values were observed to be higher in the ADHD group compared to controls. These findings suggest that the observed differences in bone structure may be influenced by additional biological or environmental factors.

An increase in FD indicates greater complexity of the examined structure, whereas a decrease reflects a flatter and less complex pattern [13, 33]. Some studies suggest that FD may increase due to higher trabecular bone complexity in osteoporotic conditions [34, 35], while others report that osteoporotic changes can reduce trabecular complexity, leading to a decrease in FD [36, 37]. However, several investigations have found no statistically significant differences in FD values between osteoporotic patients and healthy controls [38, 39]. Yaşar and Akgünlü [38] suggested that the similar FD values observed in osteoporotic and non-osteoporotic individuals could be influenced by factors such as younger mean age, higher body mass index, the measurement method, and the area selected for FD calculation. In our study, the lack of differences in angulus and condyle FD values may similarly be related to the young mean age of participants.

Malki et al. [40] reported a higher incidence of bruxism in children with ADHD receiving pharmacological treatment, particularly highlighting the association with stimulant medications. In a previous study on children with bruxism, fractal dimension values were found to be significantly higher compared to healthy controls [41]. Increased jaw activity, such as chewing or bruxism, or behavioral and lifestyle changes associated with medication use, may influence mandibular trabecular or cortical patterns. Additionally, the effects of oxytocin or medication may vary across skeletal sites due to differences in local remodelling mechanisms. In our study, interdental FD, MI, and PMI values may have been influenced by these changes; however, our study did not have sufficient sample size for the analysis of medication type and duration and was not designed to assess behavioral characteristics such as bruxism. These possible mechanisms are hypothetical and cannot be confirmed by our data.

A limitation of this study is its retrospective and cross-sectional design; therefore, causal inferences could not be made, and biochemical data such as drug doses, vitamin D levels, and parathyroid hormone levels could not be included. Another limitation is that statistical comparisons between the study groups in terms of medication type and duration of use could not be performed due to the limited sample size. This limitation may have restricted the ability to identify potential differences between the groups. The use of JPEG images instead of TIFF files is a limitation that may have affected image quality and measurement accuracy. Furthermore, intra- and inter-observer reliability could not be assessed. In addition, calibration was performed using the mesiodistal width of the mandibular central incisors on panoramic radiographs. Because panoramic images are subject to non-uniform magnification and geometric distortion, the use of a single dental reference for calibration assumes consistent magnification across the mandibular region. This approach, while commonly used, does not fully compensate for regional magnification differences inherent to panoramic imaging. This inherent limitation may have introduced systematic measurement bias, which could have affected the accuracy of MI and PMI measurements and limited direct comparability with previously reported values.

Conclusion

Differences observed in mandibular indices and interdental FD may reflect localized anatomical or developmental variations. These findings may also be influenced by parafunctional habits such as bruxism; however, this cannot be confirmed without clinical evaluation. Future research with larger sample sizes and prospective study designs should integrate assessments of lifestyle factors, medication dosage, bruxism assessments, and relevant biochemical markers to better understand mandibular bone characteristics in children with ADHD.

Abbreviations

ADHD

Attention Deficit Hyperactivity Disorder

MCI

Mandibular Cortical Index

MI

Mental Index

PMI

Panoramic Mandibular Index

FD

Fractal Dimension

ROI

Regions of Interest

BMD

Bone Mineral Density

BMC

Bone Mineral Content

DPR

Digital Panoramic Radiography

Authors’ contributions

NAÜ designed the research study. NAÜ, SÇ and DSÖÇ: performed the research and collected the data. All authors contributed to editorial changes in the manuscript. All authors read and approved the final manuscript and agree to be accountable for all aspects of the work, ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.

Funding

Not applicable.

Data availability

All data analyzed during this study are included in this manuscript.

Declarations

Ethics approval and consent to participate

This retrospective study was conducted in accordance with the principles of the Declaration of Helsinki, and was approved by the Non-Interventional Clinical Research Ethics Committee of Bolu Abant Izzet Baysal University (2024/176). Informed consent to participate was obtained from all individual participants included in the study. For participants under the age of 16, written informed consent was obtained from their parents or legal guardians.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.