Nail biting in children with ADHD: exploring its clinical correlates, psychiatric comorbidities and association with other body-focused repetitive behaviors

Abstract

Objective

This study investigated whether children with ADHD who engage in nail-biting (NB) differ clinically and cognitively from those who do not, focusing on body-focused repetitive behaviors (BFRBs), psychiatric comorbidities, and cognitive profiles.

Methods

In this retrospective, cross-sectional chart-review study, 250 medication-naive children aged 8–12 years with ADHD were grouped by the presence (n = 100) or absence (n = 150) of NB. Co-occurring BFRBs, psychiatric comorbidities, cognitive performance (WISC-R subtest scaled scores) and symptom severity (RCADS-P, CPRS-R: S) were compared using Mann–Whitney U and chi-square tests, followed by multivariable binary logistic regression.

Results

Compared with children with ADHD without NB, those with NB showed higher rates of co-occurring BFRBs (thumb sucking, lip or cheek biting, bruxism), anxiety disorders, obsessive–compulsive disorder and speech disorders, and substantially higher parent-rated internalizing symptoms (RCADS-P Total Anxiety and Depression, r_rb = 0.55). They also obtained lower scores on selected WISC-R subtests, most consistently Vocabulary, Comprehension, Picture Completion and Picture Arrangement. In multivariable analysis, older age, higher internalizing symptoms and lower scores on these subtests were independently associated with NB.

Conclusion

In children with ADHD, NB and co-occurring BFRBs were associated with a greater burden of internalizing symptoms and with lower performance on selected verbal and perceptual WISC-R subtests, and may correspond to a distinct clinical profile. These associations are cross-sectional and correlational and do not imply causality.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1186/s40359-026-05343-2.

Introduction

Attention-Deficit/Hyperactivity Disorder (ADHD) is one of the most prevalent neurodevelopmental disorders of childhood, characterized by inattention, hyperactivity, and impulsivity [1]. Studies have found the prevalence of ADHD to be around 5% [2, 3]. Individuals with ADHD frequently exhibit high rates of psychiatric comorbidities, particularly anxiety disorders, as well as various behavioral problems [4, 5].

BFRBs are a group of involuntary, repetitive motor activities that may lead to self-inflicted damage and functional impairment. These behaviors often involve attempts to modify or remove parts of the body and are considered disruptive and distressing [6]. Behaviors such as hair pulling, skin-picking, NB, cheek-chewing, lip-biting, thumb-sucking, finger-cracking, and teeth grinding are commonly classified as BFRBs in the literature [7]. Their classification, however, is not entirely uniform: nail biting and lip or cheek biting are typically discussed within the BFRB spectrum, whereas thumb sucking and bruxism may also be conceptualized as repetitive oral or parafunctional habits [8]. These behaviors were examined together in the present study because of their shared repetitive oral characteristics, while acknowledging this heterogeneity. Trichotillomania and skin-picking disorder are categorized under obsessive-compulsive and related disorders in the DSM-5, whereas NB and other BFRBs do not have a distinct classification [1]. In recent years, research on this topic has increased, exploring its etiopathogenesis and associations with psychiatric disorders [9].

BFRBs are highly prevalent in the general population [10–12]. NB, a type of BFRB, is particularly common among children and adolescents. A study investigating these behaviours in children and adolescents found the prevalence of NB to be 37%, identifying it as the most commonly observed BFRB [13]. Beyond its high prevalence, NB has also been linked to emotional and behavioral difficulties. A study examining NB behavior and problem behaviors in a community sample found that children who engaged in NB exhibited higher levels of anxious/withdrawn traits and displayed more problem behaviors [14].

In the literature, studies examining BFRBs and psychiatric comorbidities have primarily focused on depression, anxiety, and obsessive-compulsive disorder [15]. However, there is a limited number of studies exploring the relationship between ADHD and BFRBs, particularly NB. There is one study investigating comorbidity in children referred to psychiatry for NB. The prevalence of ADHD was found to be 74.6%, making it the most frequently observed comorbid condition [16].

Neurocognitive research on BFRBs has highlighted impairments in executive functions, particularly in cognitive flexibility and inhibitory motor control. Cognitive flexibility, often assessed through reversal learning tasks, refers to the ability to adapt behavior in response to changing environmental contingencies [17]. Studies have shown that individuals with BFRBs, especially trichotillomania, demonstrate significant deficits in cognitive flexibility compared to healthy controls, while findings related to other executive domains, such as working memory or attention shifting, are less consistent [18]. Additionally, abnormalities in response inhibition—defined as a top-down failure to suppress motor impulses—have been investigated, though results remain inconclusive in child and adolescent populations [19]. Furthermore, delay discounting, a measure of impulsive decision-making reflecting the devaluation of delayed rewards, has been associated with the severity of NB behaviors, suggesting its potential relevance in BFRB pathophysiology [20]. Overall, current findings indicate that neurocognitive dysfunctions—particularly in behavioral inhibition and decision-making processes—may underlie BFRBs. Given that these neurobiological mechanisms are also implicated in the etiopathogenesis of ADHD, understanding the relationship between ADHD and NB—one of the most common BFRBs in children and adolescents—becomes particularly important. A review of the literature has not identified any studies that simultaneously examine the prevalence and duration of NB in ADHD, along with its association with other co-occurring BFRBs and psychiatric comorbidities.

This study aimed to compare the demographic, clinical, and psychometric characteristics of children with ADHD with and without NB. Rather than testing a formal a priori hypothesis, we adopted an exploratory, hypothesis-generating approach. Given that individuals with BFRBs are thought to experience difficulties in emotion regulation and elevated stress levels [21], we predicted that children with ADHD and NB would show higher rates of anxiety disorders than those without NB. Considering the neurocognitive mechanisms associated with BFRBs, we further anticipated that the ADHD with NB group would show a more impaired cognitive profile.

Materials and methods

Study design and ethics

This retrospective cross-sectional chart-review study was conducted with the ethical approval of the Ethics Committee of Ankara Etlik City Hospital (Date: 07/05/2025, No: AEŞH-BADEK1-2025-092). It was carried out following the Declaration of Helsinki. Participants were identified retrospectively from outpatient clinical records and grouped according to the presence or absence of nail biting.

Participants, inclusion and exclusion criteria

A total of 5,857 pediatric patients presenting to two Child and Adolescent Psychiatry outpatient clinics between June 2023 and March 2025 were screened through medical record review. Eligible patients were aged 8–12 years, met DSM-5 criteria for ADHD, had no prior psychotropic exposure, had WISC-R scores > 70, and had completed RCADS-P and CPRS-R: S during diagnostic evaluation, with documented information on NB and other BFRBs. NB was identified according to the criteria described in the Data Collection section. Sequential exclusions included age (n = 2,816), ineligible diagnoses or confounding comorbidities (n = 1,256), missing WISC-R data (n = 621), incomplete psychometrics (n = 529), and absent BFRB information (n = 385). The final sample comprised 150 ADHD patients without NB and 100 with NB. Figure 1 depicts the study flow.

Fig. 1.

Participant screening and inclusion criteria. Note. ADHD = Attention Deficit Hyperactivity Disorder, NB = nail-biting, RCADS-P = The Revised Children’s Anxiety and Depression Scale—Parent Version, CPRS-R: S = Conners Parent Rating Scale-Revised Short Form, WISC-R = Wechsler Intelligence Scale for Children-Revised

Data collection

Hospital records were used to collect data on age, gender, parental age and education, family history of physical or mental illness, socioeconomic status, clinical diagnoses and comorbidities, the presence and duration (in months) of NB and other BFRBs, and psychometric measures (WISC-R, CPRS-R: S, RCADS-P). As noted, diagnoses were based on expert psychiatric evaluation rather than structured interviews and were supplemented by baseline psychometric assessments conducted before any pharmacological or psychotherapeutic intervention. ADHD presentations were not differentiated; all participants were analyzed within a single ADHD diagnostic category.

There is no universally accepted standard for defining nail biting as pathological. In line with previous studies, and to ensure clinical relevance, nail biting was considered present when the behavior occurred at least three days per week for more than three months and was observed in at least two settings [22, 23]. The Wechsler Intelligence Scale for Children–Revised (WISC-R) is routinely administered to children with ADHD to assess cognitive capacity and consists of Verbal and Performance subtests. Verbal subtests (Information, Similarities, Arithmetic, Vocabulary, Comprehension, and Digit Span) evaluate language-based reasoning, general knowledge, verbal comprehension, and auditory working memory, whereas Performance subtests (Picture Completion, Picture Arrangement, Block Design, Object Assembly, and Coding) assess visual–spatial processing, nonverbal reasoning, perceptual organization, and processing speed and may provide indirect information about executive functions. Every subtest was converted to an age-standardized scaled score (M = 10, SD = 3) using the Turkish normative tables, so that scores are directly comparable across participants of different ages; the Turkish standardization of the WISC-R was carried out by Savaşır and Şahin [24]. Digit Span was administered as a supplementary subtest and, in accordance with standard practice, was not included in the composites. The verbal and performance composites reported here are the sums of the five verbal (Information, Similarities, Arithmetic, Vocabulary, Comprehension) and the five performance (Picture Completion, Picture Arrangement, Block Design, Object Assembly, Coding) subtest scaled scores, respectively, and their total is the sum of all ten scaled scores. These composites are therefore scaled-score sums and not deviation IQ indices (M = 100, SD = 15); they are labeled accordingly in Table 2 and throughout the text. Because the conversion from scaled-score sum to deviation IQ in the manual is a strictly monotonic transformation, every rank-based statistic reported here (Mann–Whitney U, z, p and r) is numerically identical in either metric. Deviation Full-Scale IQ taken from the clinical WISC-R report was used only to apply the inclusion threshold (Full-Scale IQ > 70).

Table 2.

Comparison of WISC-R subtest scaled scores and subtest scaled-score sums between groups

	ADHD Without Nail-Biting (N = 150)		ADHD With Nail-Biting (N = 100)		Statistics*
	Mean (SD)	Mdn (IQR)	Mean (SD)	Mdn (IQR)	U	Z	p	r rb
Total subtest scaled-score sum	97.76 (8.34)	98.00 (10.00)	93.43 (7.53)	94.00 (9.75)	5226.5	-4.063	< 0.001	0.303
Verbal subtest scaled-score sum	48.98 (5.36)	49.00 (7.25)	47.16 (4.97)	48.00 (6.00)	6145.5	-2.423	0.015	0.181
Information	9.53 (2.24)	10.00 (3.00)	9.74 (2.29)	10.00 (3.00)	7019.5	− 0.867	0.386	-
Similarities	9.65 (2.67)	9.50 (3.00)	9.70 (2.61)	10.00 (4.00)	7353.0	− 0.264	0.792	-
Arithmetic	9.63 (2.25)	9.00 (3.00)	9.39 (2.65)	10.00 (3.00)	7210.5	− 0.521	0.602	-
Vocabulary	10.04 (2.12)	10.00 (2.25)	8.93 (2.36)	9.00 (3.00)	5424.5	-3.742	< 0.001	0.277
Comprehension	10.13 (2.52)	10.00 (4.00)	9.40 (2.30)	9.00 (3.00)	6286.5	-2.184	0.029	0.162
Performance subtest scaled-score sum	48.78 (5.25)	48.00 (6.00)	46.27 (5.03)	46.00 (7.00)	5484.0	-3.607	< 0.001	0.269
Picture Completion	10.11 (2.46)	10.00 (4.00)	8.94 (2.51)	9.00 (3.75)	5579.5	-3.454	0.001	0.256
Picture Arrangement	9.87 (2.48)	10.00 (4.00)	8.55 (2.29)	8.00 (3.00)	5238.0	-4.070	< 0.001	0.302
Block Design	9.65 (2.67)	9.00 (3.00)	9.73 (2.59)	10.00 (4.00)	7371.0	− 0.231	0.823	-
Object Assembly	9.58 (2.55)	9.50 (3.00)	9.32 (2.39)	10.00 (3.75)	7229.5	− 0.486	0.627	-
Coding	9.58 (2.47)	9.00 (3.00)	9.73 (2.59)	10.00 (4.00)	7156.5	− 0.618	0.537	-

The verbal and performance subtest scaled-score sums are the sums of the five verbal and the five performance subtest scaled scores (each M = 10, SD = 3), and the total subtest scaled-score sum is the sum of all ten; these are sums of scaled scores, not deviation IQ indices

ADHD  Attention Deficit Hyperactivity Disorder, SD  Standard deviation, Mdn  Median, IQR  Interquartile range, WISC-R  Wechsler Intelligence Scale for Children-Revised

*The Mann–Whitney U test was used. Effect sizes are rank-biserial correlations (r_rb = 1 − 2U/(n₁n₂)) and are reported only for statistically significant comparisons; values are given as magnitudes, and the direction of each difference can be read from the group means and medians

Bold p values indicate p < 0.05

The Conners’ Parent Rating Scale–Revised: Short Form (CPRS-R: S), developed by Conners et al. [25], assesses ADHD symptom domains across defiance, cognitive difficulties/inattention, and hyperactivity, as well as an ADHD index. Its Turkish adaptation and psychometric validation were conducted by Kaner et al. [26].

The Revised Child Anxiety and Depression Scale–Parent Version (RCADS-P), developed according to DSM-IV criteria, measures anxiety disorders and depressive symptoms, including separation anxiety disorder (SAD), social phobia disorder (SPD), generalized anxiety disorder (GAD), panic disorder (PD), obsessive–compulsive disorder (OCD), and major depressive disorder (MDD), along with the composite scales Total Anxiety (TA) and Total Anxiety and Depression (TAD) [27]. The Turkish validity and reliability study was conducted by Görmez et al. [28]. These psychometric assessment tools are administered before any pharmacological or psychotherapeutic intervention during the initial diagnostic phase. As such, they reflect the patient’s baseline psychological profile.

Statistical analysis

All analyses were performed using the Statistical Package for the Social Sciences (SPSS) version 26.0. Descriptive statistics summarized demographic and clinical characteristics; categorical variables were reported as frequencies and percentages, and continuous variables as means with standard deviations and medians with interquartile ranges (IQR). Group comparisons for categorical variables used the Chi-square or Fisher’s exact test. Normality was assessed with the Kolmogorov–Smirnov test and indicated non-normal distributions; therefore, the Mann–Whitney U test was applied to RCADS-P and CPRS-R: S scores. The same test was applied to WISC-R subtest scaled scores and scaled-score sums. Adjustment for internalizing symptoms was carried out within the multivariable logistic regression described below rather than by analysis of covariance: parent-reported anxiety is plausibly a mediator as well as a confounder of the association between nail biting and verbal test performance, and covarying a mediator would bias the estimate of the total association towards the null. This decision is based on prior evidence linking anxiety and depression to both ADHD and BFRBs. Because the group comparisons were rank based, effect sizes are reported as the rank-biserial correlation, r_rb = 1 − 2U/(n₁n₂), which expresses the difference between the probability that a randomly selected child from one group scores higher than a randomly selected child from the other and the probability of the reverse, and is therefore the effect size that corresponds directly to the Mann–Whitney U test (0.10 small, 0.30 medium, 0.50 large). Cohen’s d was not used, as a mean-based standardized difference is not consistent with rank-based analyses. Odds ratios are reported for 2 × 2 comparisons. Every effect size in the tables can be reproduced from the reported U and the two group sizes. Variables showing significant group differences were entered into binary logistic regression models to predict NB, with highly collinear subscales excluded. The model included seven predictors with 100 nail-biting cases as the limiting outcome, yielding approximately 14 events per variable — above the conventional minimum of ≥ 10 events per variable — indicating adequate statistical power and a low risk of overfitting [29]. Statistical significance was set at p < 0.05 (two-tailed). Given the exploratory, hypothesis-generating nature of the study, formal correction for multiple comparisons was not applied, so as not to inflate the risk of Type II error at the stage at which associations are being identified rather than confirmed; the principal findings were nonetheless associated with large effect sizes (r_rb up to 0.63) and significance levels well below conventional thresholds (p < 0.001), which reduces the likelihood that they reflect Type I error.

Results

Baseline demographic and clinical characteristics of the ADHD without nail-biting (N = 150) and ADHD with nail-biting (N = 100) groups are presented in Supplementary Table 1. Children in the ADHD with nail-biting group were older than those without nail-biting (mean ± SD, 11.1 ± 1.40 vs. 10.0 ± 1.59 years, p < 0.001), whereas the proportion of girls was similar between groups (42.0% vs. 44.7%, p =0.677). The nail-biting group was also in higher school grades and had older parents, while no group differences were observed for parental education levels, socioeconomic status, or parental psychiatric illness (all p > 0.05).

Comparisons of other body-focused repetitive behaviors are presented in Supplementary Table 2. Thumb sucking, lip or cheek biting, and teeth grinding/bruxism were all significantly more prevalent in the ADHD with nail-biting group (all p < 0.05). Among children who exhibited these behaviors, the mean duration of lip/cheek biting was shorter in the nail-biting group (4.7 ± 2.8 vs. 6.6 ± 2.4 months, p = 0.022), whereas teeth grinding/bruxism persisted for a longer period (15.6 ± 9.9 vs. 8.0 ± 3.4 months, p = 0.030). No group difference was observed in the duration of thumb sucking (p = 0.892).

Examination of BFRB patterns indicated that most participants (N = 112) exhibited no BFRBs, constituting the largest subgroup. NB alone (N = 29) and the combination of NB with teeth grinding/bruxism (N = 21) were the next most frequent patterns, whereas other combinations involving thumb sucking, lip/cheek biting, and teeth grinding were comparatively rare (see Supplementary Fig. 1).

Symptom comparisons are shown in Table 1. Scores on every RCADS-P subscale and on both composite scales were significantly higher in the ADHD with NB group (all p < 0.001), with medium-to-large effects that were largest for Major Depression (r_rb = 0.63) and Total Anxiety and Depression (r_rb = 0.55). By contrast, CPRS-R: S Oppositional Behavior, Hyperactivity and ADHD Index scores did not differ between groups, and the difference on Cognitive Problems/Inattention was small and only marginally significant (p = 0.038, r_rb = 0.15).

Table 1.

Comparison of RCADS-P and CPRS-R: S subscales scores between groups

	ADHD Without Nail-Biting (N = 150)		ADHD With Nail-Biting (N = 100)		Statistics*			r rb
	Mean (SD)	Mdn (IQR)	Mean (SD)	Mdn (IQR)	U	Z	p
RCADS-P
SAD	51.85 (11.18)	49.00 (16.25)	61.63 (13.77)	59.00 (25.00)	4345.0	-5.640	< 0.001	0.421
SPD	47.61 (11.01)	49.00 (16.00)	56.40 (13.18)	57.50 (16.75)	4593.0	-5.193	< 0.001	0.388
GAD	54.11 (12.47)	52.50 (17.00)	66.71 (14.46)	71.00 (25.75)	3939.5	-6.385	< 0.001	0.475
PD	43.93 (12.88)	42.00 (18.25)	54.22 (14.17)	56.00 (18.75)	4451.0	-5.447	< 0.001	0.407
OCD	51.24 (11.89)	49.00 (18.00)	60.06 (13.12)	61.00 (20.00)	4629.5	-5.129	< 0.001	0.383
MDD	49.47 (12.71)	46.00 (20.50)	67.44 (14.60)	73.00 (22.00)	2781.5	-8.455	< 0.001	0.629
Total Anxiety	49.51 (13.79)	48.00 (22.00)	62.76 (14.98)	63.00 (24.75)	3922.0	-6.394	< 0.001	0.477
Total Anxiety & Depression	49.39 (13.82)	49.00 (23.00)	64.94 (15.09)	67.00 (25.75)	3377.0	-7.376	< 0.001	0.550
CPRS-R: S
Oppositional Behavior	10.36 (5.89)	10.00 (11.00)	9.52 (5.11)	10.00 (8.00)	7011.5	-0.876	0.381	-
Cognitive Problems /Inattention	12.31 (5.43)	14.00 (7.00)	10.42 (6.33)	12.50 (12.00)	6342.5	-2.078	0.038	0.154
Hyperactivity	8.77 (5.82)	9.00 (11.00)	8.37 (5.34)	8.00 (9.75)	7241.5	-0.462	0.644	-
ADHD Index	23.89 (9.54)	25.50 (12.00)	21.25 (11.04)	22.00 (22.00)	6559.5	-1.681	0.093	-

ADHD  Attention Deficit Hyperactivity Disorder, SD  Standard deviation, Mdn  Median, IQR  Interquartile range, RCADS-P  The Revised Children’s Anxiety and Depression Scale—Parent Version, SAD  Separation Anxiety Disorder, SPD  Social Phobia Disorder, GAD  Generalized Anxiety Disorder, PD  Panic Disorder, OCD  Obsessive Compulsive Disorder, MDD  Major Depressive Disorder, CPRS-R: S  Conners Parent Rating Scale-Revised Short Form

*The Mann–Whitney U test was used. Effect sizes are rank-biserial correlations (r_rb = 1 − 2U/(n₁n₂)) and are reported only for statistically significant comparisons; values are given as magnitudes, and the direction of each difference can be read from the group means and medians

Bold p values indicate p < 0.05

Table 2 presents the comparison of WISC-R scaled scores between children with ADHD with and without NB. The ADHD with NB group obtained a lower total subtest scaled-score sum (r_rb = 0.30) and a lower verbal subtest scaled-score sum (r_rb = 0.18); among the verbal subtests the difference was carried by Vocabulary (r_rb = 0.28) and, more weakly, Comprehension (r_rb = 0.16), whereas Information, Similarities and Arithmetic did not differ. The performance subtest scaled-score sum was likewise lower in the ADHD with NB group (r_rb = 0.27), with significant differences for Picture Completion (r_rb = 0.26) and Picture Arrangement (r_rb = 0.30) but not for Block Design, Object Assembly or Coding. Because the group difference in internalizing symptoms was itself substantial, the independence of these cognitive differences from anxiety and depression is addressed in the multivariable model below rather than by covariance adjustment.

The anxiety-disorder profiles of the two groups are illustrated in Supplementary Fig. 2. Speech disorders, and particularly speech sound disorder, any anxiety disorder (including separation anxiety disorder, social phobia and performance anxiety) and obsessive–compulsive disorder were more prevalent in the ADHD with NB group. By contrast, no between-group differences were observed for oppositional defiant disorder, conduct disorder, specific learning disorders, tic disorders, nocturnal enuresis, depressive disorders, generalized anxiety disorder, panic disorder, specific phobia or sleep disorders (all p > 0.05). The pattern is therefore one of selective divergence in internalizing rather than externalizing comorbidity.

A binary logistic regression analysis was conducted to examine whether age, cognitive symptoms (CPRS-R: S), internalizing psychopathology (RCADS-P), and specific cognitive performance variables (WISC-R subtests) significantly predicted the presence of NB behavior in children with ADHD. Only the variables that showed significant group differences in prior comparative analyses were included in the logistic regression model. Due to high intercorrelations among the RCADS-P subscales, only the total internalizing score (RCADS-P Total Anxiety and Depression) was entered into the model to avoid multicollinearity. Similarly, because the WISC-R total, verbal IQ, and performance IQ scores were highly correlated with their respective subtests, only specific subscale scores were included to provide a more granular assessment of cognitive functioning (Table 3).

Table 3.

Logistic regression analysis of factors associated with the presence of nail biting in children diagnosed with ADHD

		B	SE	Wald	df	p	Exp(B)	95% Confidence Interval for Exp (B)
								Lower	Upper
Model 1	(Constant)	1.581	1.928	0.672	1	0.412	4.859
	Age (years)	0.379	0.111	11.570	1	< 0.001	1.461	1.174	1.818
	CPRS-R: S-Cognitive Problems /Inattention	− 0.066	0.030	4.883	1	0.027	0.936	0.883	0.993
	RCADS-P-Total Anxiety & Depression	0.071	0.012	34.891	1	< 0.001	1.073	1.048	1.099
	WISC-R-Vocabulary	− 0.257	0.081	10.067	1	0.002	0.774	0.660	0.907
	WISC-R- Comprehension	− 0.168	0.073	5.326	1	0.021	0.846	0.733	0.975
	WISC-R-Picture Completion	− 0.251	0.073	11.951	1	< 0.001	0.778	0.675	0.897
	WISC-R-Picture Arrangement	− 0.301	0.078	15.039	1	< 0.001	0.740	0.636	0.862

Independent variables associated with the presence of nail biting (dependent variable) were assessed using logistic regression analysis

ADHD  Attention Deficit Hyperactivity Disorder, SE  Standard error, CPRS-R: S  Conners Parent Rating Scale-Revised Short Form, RCADS-P  The Revised Children’s Anxiety and Depression Scale—Parent Version, WISC-R  Wechsler Intelligence Scale for Children-Revised

Bold p values indicate p < 0.05

The logistic regression model was significant (χ²(7) = 128.30, p < 0.001) and showed adequate fit (Hosmer–Lemeshow χ²(8) = 6.39, p = 0.603), explaining 40.1% (Cox & Snell R²) and 54.3% (Nagelkerke R²) of the variance, with an overall classification accuracy of 78.4%. Older age (OR = 1.46) and higher RCADS-P Total Anxiety and Depression scores (OR = 1.07) were independently associated with NB, whereas higher CPRS-R: S Cognitive Problems/Inattention scores were inversely related (OR = 0.94). Lower performance on WISC-R Vocabulary (OR = 0.77), Comprehension (OR = 0.85), Picture Completion (OR = 0.78), and Picture Arrangement (OR = 0.74) was also associated with increased odds of NB (all p < 0.05).

Discussion

In this cohort, children with ADHD with and without NB differed across several sociodemographic, behavioral, cognitive, and psychiatric domains; we interpret these differences below in light of the broader BFRB literature.

The ADHD with NB group was significantly older than the ADHD without NB group. This accords with reviews reporting that NB prevalence rises from early childhood and peaks around puberty [30], and likely reflects the age-related increase in NB as children approach adolescence.

Children with NB also showed a higher prevalence of other BFRBs — thumb sucking, lip or cheek biting, and bruxism — consistent with the well-documented co-occurrence of multiple BFRBs within the same individual [31, 32]. This clustering is often attributed to shared neurobiological mechanisms and phenomenological features [33] and, for the oral behaviors in particular, to their overlapping classification as oral parafunctional habits [8] and to proposed abnormalities in oromandibular sensory processing [34, 35]. The shorter reported duration of lip or cheek biting may stem from its lower visibility and reliance on self-report, leading to later recognition than more overt behaviors such as thumb sucking and bruxism.

On the WISC-R, the ADHD with NB group scored lower on selected verbal and perceptual subtests, most consistently Vocabulary, Comprehension, Picture Completion and Picture Arrangement. Cognitive characterization of children with BFRBs is scarce, and to our knowledge no prior study has used IQ-based measures in this population. In adults with BFRBs, poorer cognitive flexibility has been reported relative to controls [18]. Although Vocabulary and Comprehension primarily index verbal knowledge, reasoning, and social judgment rather than cognitive flexibility directly, lower performance may still be associated with reduced flexibility [36], and deficits in motor inhibition — a core executive function — have likewise been described in BFRBs [37]. The lower Picture Completion and Picture Arrangement scores are similarly compatible with indirect demands on perceptual reasoning, planning, set-shifting, and inhibitory control [36]. These findings are consistent with a more impaired cognitive profile in the ADHD with NB group, particularly in executive-related domains. This interpretation warrants caution, however: speech and anxiety disorders were more prevalent in this group, and anxiety in particular can impair verbal working memory and verbal task performance through worry-related interference [38, 39]. The lower verbal performance may therefore partly reflect a greater anxiety burden rather than a primary cognitive deficit. For this reason we did not covary internalizing symptoms in the group comparisons: where a variable may lie on the causal path between the grouping factor and the outcome, adjusting for it removes part of the association of interest. Instead, the multivariable logistic model estimated the contribution of the WISC-R subtests with RCADS-P Total Anxiety and Depression in the model, and the subtest associations persisted, indicating that they are not wholly attributable to differences in internalizing symptoms. It should also be noted that the group differences are small in absolute terms — approximately one scaled-score point, well within the average range — and are not of a magnitude that would be clinically apparent in an individual child.

The ADHD with NB group also reported higher anxiety and depressive symptoms on the RCADS-P across multiple subscales (SAD, SPD, GAD, PD, OCD, and MDD). This fits the view of BFRBs as maladaptive strategies for regulating emotional distress [40] and the broadly reported emotion-regulation difficulties in BFRBs [41]. Boredom, tension, anxiety, and frustration are the affective states most likely to trigger and be modulated by BFRB episodes [42]; accordingly, NB has been observed to occur during boredom or demanding tasks [43], to provide tension relief in severe cases [44], and to be maintained by automatic reinforcement under low-stimulation or task-demand conditions [45]. Together these observations suggest that the elevated anxiety burden in the ADHD with NB group may both trigger and reinforce the behavior, making the higher anxiety scores an expected finding — although reliance on parent report remains a limitation. The selective elevation of anxiety disorders, alongside comparable rates of most other comorbidities and a smaller rise in speech disorders, is apparent in Supplementary Table 3, and the distribution across anxiety-disorder subtypes is illustrated in Supplementary Fig. 2.

By contrast, CPRS-R: S oppositional, hyperactivity, and ADHD-index scores did not differ between groups, suggesting that NB is not simply a marker of ADHD severity. Notably, the NB group’s lower objective cognitive performance contrasted with lower parent-rated cognitive problems/inattention — a discrepancy that may reflect the difference between objective testing and parental perception; teacher report, capturing difficulties more evident at school, could add precision to future assessments.

Consistent with the group comparisons, internalizing symptoms and cognitive performance also independently predicted NB in the regression analysis, reinforcing their role as key correlates of NB in children with ADHD.

Clinical implications

Three practical points follow from these findings. First, nail biting is worth asking about directly during the ADHD assessment, together with thumb sucking, lip or cheek biting and bruxism, since these behaviors clustered strongly with one another and are rarely volunteered by families who regard them as habits rather than symptoms. Second, where nail biting is present, an explicit screen for anxiety is warranted: any anxiety disorder was almost twice as common in this group, separation anxiety and social phobia roughly three to four times as common, and parent-rated internalizing symptoms were higher across every RCADS-P subscale. A brief parent-report instrument such as the RCADS-P is sufficient for this purpose and is already part of many ADHD assessment protocols. Third, and equally important, nail biting should not be read as a marker of more severe ADHD: CPRS-R: S Oppositional, Hyperactivity and ADHD Index scores were indistinguishable between the groups, so the presence of the behavior is not by itself a reason to intensify treatment directed at core ADHD symptoms. Management is instead likely to be better directed at the co-occurring internalizing symptoms and, where the behavior itself causes damage or distress, at behavioral approaches such as habit reversal training, with dental referral where bruxism is prominent. Whether such intervention alters outcome was not assessed here [46]. The absence of group differences in externalizing disorders further suggests that internalizing symptoms play a more prominent role in this subgroup, supporting tailored, integrative strategies that address both core ADHD symptoms and co-occurring emotional difficulties.

Strengths and limitations

This study adds to the limited literature on BFRBs in ADHD and offers a more detailed picture of their cognitive and psychiatric correlates. Its principal strengths are a relatively large clinical sample, the use of standardized and validated instruments (WISC-R, RCADS-P, and CPRS-R: S), and the simultaneous examination of multiple BFRBs alongside ADHD-related clinical profiles. These strengths, however, should be weighed against several limitations. Information on the frequency, severity, and functional impact of NB and other BFRBs was not available, and the absence of both a validated BFRB-specific scale and a universally accepted threshold for defining pathological nail biting may have affected measurement precision. As in any chart-based study, behaviors that were present but not recorded will have been misclassified as absent; because such under-recording is most likely for the less salient presentations, and therefore for the comparison group, it would tend to attenuate rather than create the group differences reported here. The assessment was limited to NB, thumb sucking, lip/cheek biting, and bruxism, excluding other BFRBs such as trichotillomania and skin-picking disorder. Internalizing symptoms were assessed only through parent reports, and structured diagnostic interviews were not used owing to the retrospective, naturalistic design. In addition, ADHD presentation subtypes were not distinguished; because the inattentive presentation may be associated with higher rates of anxiety, an unmeasured difference in the distribution of presentations between groups could have contributed to the internalizing-symptom findings, although the comparable proportion of girls across groups argues against a substantial imbalance. The proportion of girls was nonetheless high for a clinically referred ADHD sample in both groups (44.7% and 42.0%), which suggests that the cohort as a whole may over-represent the predominantly inattentive presentation; the findings may therefore generalize less well to samples in which the hyperactive-impulsive and combined presentations predominate. Children with a WISC-R Full-Scale IQ of 70 or below were excluded, so the findings may not generalize to children with ADHD and co-occurring borderline intellectual functioning or intellectual disability. The verbal and performance composites are sums of subtest scaled scores rather than deviation IQ indices and should not be read as Verbal or Performance IQ. Finally, the cognitive differences observed, although statistically reliable, were small in magnitude. The lack of a healthy control group further limits specificity to ADHD. Finally, some analyses were exploratory, and the observed associations do not imply causality.

Conclusion

This study provides insight into the associations between BFRBs, cognitive performance, psychiatric comorbidities, and emotional symptoms in children with ADHD. NB and other BFRBs appear to mark a clinical profile characterized by a greater burden of internalizing symptoms and anxiety disorders, higher rates of speech disorders, and modestly lower performance on selected cognitive subtests.

The co-occurrence of BFRBs and their links to anxiety and emotional regulation difficulties support a multidimensional approach to ADHD assessment; however, these associations should not be interpreted causally. Clinically, the findings highlight the importance of screening for BFRBs and related cognitive and emotional symptoms in children with ADHD, and of implementing early, targeted interventions addressing both core ADHD symptoms and emotional regulation difficulties. Longitudinal studies with healthy control groups are needed to clarify developmental trajectories and clinical implications.

Authors’ contributions

H.B.D. and M.E.Ö. contributed to the conceptualization and methodology of the study. H.B.D. and M.E.Ö were responsible for the investigation, data curation, and project administration. M.E.Ö. performed the formal statistical analysis and contributed to validation, visualization, and resources. H.B.D. drafted the initial manuscript, and M.E.Ö. critically reviewed and revised it for important intellectual content. Both authors reviewed the manuscript, approved the final version, and agree to be accountable for all aspects of the work.

Funding

No funding was received for this research.

Data availability

The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.

Declarations

Ethics approval and consent to participate

The study protocol was approved by the Clinical Research Ethics Committee of Ankara Etlik City Hospital (Date: 07/05/2025, No: AEŞH-BADEK1-2025-092) and was conducted in accordance with the Declaration of Helsinki. Owing to the retrospective design based on the review of existing medical records, the Ethics Committee waived the requirement for informed consent.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.