A traumatic brain injury (TBI) is characterised by a disruption of normal brain function or structure caused by an external force.1–3 It is theorised that TBI sets off a prolonged secondary injury process, driven by neuroinflammation, oxidative stress, and endoplasmic reticulum stress, ultimately leading to neurodegeneration which may lead to serious psychiatric problems.4 5

Despite significant research into the incidence of pTBI in Finland over the past decade, a notable gap remains in understanding the long-term neurological consequences following such injuries.6 7 A Finnish cohort study reported a 118% increase in pTBI incidence over 20 years (1998–2018), from 251 to 547 per 100 000.8

The current literature on pTBI-related attention-deficit/hyperactivity disorder (ADHD) reveals conflicting and limited findings. Prior studies, frequently with small cohorts and lacking nationwide representation, have produced inconsistent results. Some small population meta-analyses suggest a association between pTBI and ADHD, a trend supported by comparable findings in smaller cohort studies.9–11 However, a small cohort study involving 30 patients reported disparate results, indicating no apparent correlation between mild pTBI and ADHD.12

Exploring the potential causal relationship between pTBI and ADHD medication is an underexplored area and has not been studied within a nationwide population and in this large cohort before, nor with injury reference patients. This study is designed to address this gap by investigating if pTBI is linked to the occurrence ADHD medication and how the sex of patient with pTBI and operative treatment affect the ADHD medication occurrence.

This nationwide retrospective register-based cohort study used data from the Finnish Care Register for Healthcare and the Finnish Social Insurance Institution. Additionally, authorisation was obtained from Statistics Finland to access Population Information and the Register of Death Causes. The study spanned from January 1998 to December 2018 and adhered to the Strengthening the Reporting of Observational Studies in Epidemiology guidelines. The checklist is available in the online supplemental material 1 online supplemental materials.

The primary study cohort comprised individuals (aged <18 years) who experienced pTBI during the study period of 20 years. This group was hospitalised to specialised healthcare (secondary and/or tertiary level) for TBIs under International Classification of Diseases, Tenth Revision (ICD-10) codes starting with ‘S06*’. This included both inpatients and outpatients, indicating that patients may have been discharged directly from the emergency department or admitted to wards. Data were sourced from the Finnish Care Register for Healthcare, covering patient information from specialised healthcare visits, surgeries and hospitalisations nationwide.

The reference group comprised hospitalisations of patients under 18 years at the time of injury with ankle (ICD-10 S82*) or wrist fractures (ICD-10 S52*) between January 1998 and December 2018. This choice was based on their comparable behavioural profiles and leisure activities to children with TBI, ensuring a similar risk of physical injury in both groups.

Data collection involved recording hospitalisation dates for TBI or distal extremity fractures, along with details on age at injury and sex. Mortality and emigration data during the study period were also captured. TBI cases were categorised based on whether operative or non-operative treatment was received. The study assessed the incidence of neurosurgical procedures using Nordic Medico-Statistical Committee Classification of Surgical Procedures codes relevant to Finland. Online supplemental appendices 1 and 2 contain the full list of diagnosis and neurosurgical procedure codes used in the study.

The usage of ADHD medications were identified using specific Anatomical Therapeutic Chemical codes: N06BA02 (dextroamphetamine), N06BA04 (methylphenidate), N06BA09 (atomoxetine), N06BA12 (lisdexamfetamine) and C02AC02 (guanfacine), encompassing all ADHD medications in Finland. These codes were extracted from the Drug Purchase Register at the Finnish Social Insurance Institution. Acquiring these prescriptions in Finland mandates a thorough psychoneurological assessment by specialised healthcare providers, such as educational care doctors, paediatricians, paediatric neurologists and paediatric/youth psychiatrists. Given the prevalence of medication treatment for paediatric ADHD in Finland, our study includes a substantial portion of Finnish children with ADHD.13

Descriptive statistics were presented as counts (%), along with the median and IQR based on the distribution. Only the initial hospitalisation was considered for cases involving repeated injuries. Due to the lengthy ADHD diagnosis process in Finland, involving a comprehensive psychological cognitive examination by psychologists before the initial diagnosis, patients starting ADHD medications within 1 year post-trauma were excluded. A 1-year washout period was established. Follow-up began 1 year after pTBI or distal extremity fracture, effectively representing a 2-year interval from the event. This method enhances the likelihood of associating observed ADHD medication cases with the pTBI rather than pre-existing conditions.

After exclusions, the final study group comprised 128 006 patients (online supplemental appendix 3). Follow-up was continued until individuals relocated abroad, experienced mortality or reached the common closing date of the surveillance period (31 December 2018). Individuals experiencing both TBI and a fracture were included in the TBI group. TBI subgroup analysis distinguished non-operatively treated and operatively treated cases.

The cumulative incidece of ADHD-medication consumption was assessed using Kaplan-Meier (KM) survival analysis and a multivariable Cox regression model for both the pTBI and reference groups, including 95% CIs. This analysis spanned from 1 year after hospitalisation to the onset of ADHD medication, presenting results through cumulative incidence rates. Covariate selection for the regression analysis was guided by a directed acyclic graph (DAG)14 outlined in online supplemental appendix 4. HRs interpreted the Cox regression results. The dependent variable was ADHD-medication starting date, with only post-traumatic starting dates being included in the analysis. All analyses were adjusted for potential confounders, including age at hospitalisation and sex. Violations of proportional hazard (PH) assumptions were evaluated by examining the correlation of scaled Schoenfeld residuals with time.

Visual inspection of the correlation between scaled Schoenfeld residuals and log-log survival curves was conducted to assess PH assumptions. A time-stratified model15 was implemented to address PH assumption violations, although the large sample size limited resolution. Continuous evaluation of Schoenfeld residuals with time was performed for effective handling of non-proportionality. To address this more effectively, a time-dependent coefficients method was applied, dividing the study population into intervals of 0–1 year, 1–4 years and 4–20 years of follow-up.

Subgroup survival analyses compared ADHD medication-free survival between operatively and non-operatively treated pTBI cases. Cumulative ADHD medication incidence differences between males and females in both reference and pTBI groups were examined. Employing KM survival analysis and a multivariable Cox regression model, mirroring the main analysis, ADHD medication starting date was the primary dependent variable. Analyses were adjusted for potential confounders, including age at hospitalisation and sex.

All analyses and figures were performed using Windows R V.4.0.5 (R Foundation for Statistical Computing, Vienna, Austria) with the packages tidyverse, ggfortify, survival and survminer.

The primary cohort consisted of 66 594 patients with pTBI, alongside 61 412 references. Notably, 4012 patients experienced both TBI and distal extremity fractures.

During the follow-up, ADHD medication was initiated for 4924 patients. Among them, 3164 individuals began medication following either a TBI or distal extremity fracture and were included in the analysis. Of these, 2061 (3.09%) were from the pTBI group, while 1103 (1.80%) belonged to the reference group. Throughout the study, 590 patients passed away, including 20 who had commenced ADHD medication earlier. Additionally, 25 out of 2728 emigrants had a history of ADHD medication. The median age at hospitalisation for injury was 7 years in the pTBI group and 11 years in the reference group. In the pTBI group, 58% were males, compared with 62% in the reference group. Among patients with pTBI, 293 underwent operative treatment and 19 (6.48%) of them began ADHD medication post-injury (table 1).

Sex distribution, median year of hospitalisation and age demographics of ADHD associated with pTBI in Finland (1998–2018)

In the KM analyses, consistently higher cumulative incidence rates for ADHD medication were observed in the pTBI group throughout the study period, especially among those who underwent operative treatment for TBI. After 1 year of follow-up, the cumulative incidence rates were 0.35% in the pTBI group and 0.31% in the reference group. However, these rates increased to 3.89% in the pTBI group and 1.90% in the reference group after 10 years of follow-up, and further increased to 5.98% in the pTBI group and 3.05% in the reference group after 20 years of follow-up (figure 1A and table 2A).

(A) Kaplan-Meier curve illustrating the cumulative rate of ADHD medication in individuals with pTBI in Finland from 1998 to 2018, along with corresponding 95% CIs. (B) Kaplan-Meier curve of the cumulative incidence rate of ADHD medication between operated and non-operated patients, with 95% CIs, following pTBI in Finland (1998–2018).

(A) Kaplan-Meier curve illustrating the cumulative rate of ADHD medication in individuals with pTBI in Finland from 1998 to 2018, along with corresponding 95% CIs; (B) ADHD at 1–20 years after paediatric trauma in TBI and distal extremity fracture groups in a Finnish nationwide sample

When comparing patients with pTBI who underwent operative treatment with those who did not, the cumulative incidence rates for ADHD medication were 2.08% in the operative treatment group and 0.33% in the non-operative treatment group after 1 year of follow-up. These rates increased to 6.83% in the operative treatment group and 3.81% in the non-operative treatment group after 10 years of follow-up, and further increased to 8.28% in the operative treatment group and 6.14% in the non-operative treatment group after 20 years of follow-up (figure 1B and table 2B).

Cox regression revealed that patients with pTBI had a higher risk of ADHD medication usage compared with the reference group in the long term. In the first year of follow-up, the reference group exhibited a higher risk of ADHD medication usage. After the initial year of follow-up, the HR for the pTBI group was 0.90 (95% CI 0.74 to 1.10) and 1.42 (95% CI 1.24 to 1.62) between years 1 and 4. However, the risk of ADHD medication increased after 4 years of follow-up, with an HR of 1.89 (95% CI 1.70 to 2.10) between 4 and 20 years of follow-up. When comparing the cumulative incidence differences of ADHD medication-free survival between operatively and non-operatively treated patients with pTBI, the risk was higher in the operative treatment group. After the first year of follow-up, the HR for the operative group was 6.31 (95% CI 2.80 to 14.20) and 1.39 (95% CI 0.80 to 2.39) between 1 and 20 years of follow-up.

Comparing the cumulative incidence differences in ADHD medication-free survival between female patients with pTBI and the female reference group revealed a higher risk in the pTBI group.

The cumulative incidence of ADHD medication was 0.14% in the pTBI female group and 0.13% in the reference female group after 1 year of follow-up, while the corresponding figures for men were 0.49% in the pTBI group and 0.39% in the reference group. After 10 years of follow-up, the cumulative incidence of ADHD medication increased among female patients to 2.14% in the pTBI group and 1.07% in the reference group. After 20 years of follow-up, these numbers further increased to 4.58% in the pTBI group and 2.42% in the reference group (figure 2 and table 3).

Kaplan-Meier curve of the cumulative incidence rate of attention-deficit/hyperactivity disorder (ADHD) medication between paediatric traumatic brain injury (pTBI) and distal extremity fracture in female and male patients in Finland (1998–2018), along with 95% CIs.

ADHD medication incidence at 1–20 years after paediatric trauma in TBI and distal extremity fracture groups in female and male population

In contrast, men with pTBI experienced a more pronounced increase in ADHD medication incidence. After 10 years of follow-up, the pTBI group had a cumulative incidence of 5.02%, compared with 2.35% in the reference group. After 20 years of follow-up, these numbers further increased to 7.26% in the pTBI group and 3.48% in the reference group (figure 2 and table 3).

Following the first year of follow-up, the HR for the pTBI female group was 1.04 (95% CI 0.64 to 1.69), escalating to 2.01 (95% CI 1.72 to 2.35) between 1 and 20 years of follow-up. Similar results were observed in the comparison between male patients with pTBI and the reference group. After the initial year of follow-up, the HR for the pTBI group was 1.25 (95% CI 1.01 to 1.56), further increasing to 2.23 (95% CI 2.04 to 2.45) between 1 and 20 years of follow-up.

The findings of our nationwide cohort study revealed an association between pTBI and an increased risk of post-traumatic ADHD medication compared with the distal extremity group. This risk was particularly prominent after 4 years of follow-up. Notably, we observed an elevated risk, especially after 1 year of follow-up, among patients with pTBI who underwent operative treatment for pTBI. However, it is crucial to consider the wide CIs due to small number of operatively treated patients, which introduces some imprecision to the results. On individual examination, both male and female pTBI groups exhibit a higher risk of developing ADHD compared with their respective reference groups.

The potential correlation observed could be linked to post-TBI neuroinflammation and oxidative stress, which in turn lead to neurodegeneration. This process may affect brain development and neurotransmitter function, consequently increasing the risk of neurodevelopmental disorders. These insights align with previous literature.4 5 16 17 Additionally, TBI-induced diffuse axonal injury has the potential to disrupt neural networks crucial for attention and cognitive function.18 Consequently, individuals may encounter challenges in maintaining attention, controlling impulses and regulating behaviour, which are all characteristic features of ADHD.

Existing findings on pTBI-related ADHD are limited and to some extent inconsistent. Unlike specific prior investigations, our findings closely align with a meta-analysis of 24 cohort studies encompassing 12 374 patients with pTBI. This meta-analysis reported an association solely with severe pTBI and post-traumatic ADHD. The meta-analysis relied on the Glasgow Coma Scale (GCS) as an indicator of pTBI severity, potentially introducing diagnostic bias stemming from exclusive clinical judgement. The odds for ADHD following severe TBI were elevated at both time points post-trauma compared with other injuries (≤1 year: OR 4.81 (95% CI 1.66 to 11.03); >1 year: OR 6.70 (95% CI 2.02 to 16.82)) and non-injured references (≤1 year: OR 2.62 (95% CI 0.76 to 6.64); >1 year: OR 6.25 (95% CI 2.06 to 15.06)), as well as those with mild TBI (≤1 year OR 5.69 (95% CI 1.46 to 15.67); >1 year OR 6.65 (95% CI 2.14 to 16.44)).9 However, this meta-analysis had several limitations. First, the diagnoses of ADHD were often unreliable due to many studies relying on self-reported diagnoses. Additionally, the wide CIs observed in the studies may introduce some uncertainty to the results, affecting the overall robustness of the findings. Furthermore, the majority of the studies included in the analysis had under 1000 patients with pTBI, which could impact the generalisability of the findings. Moreover, most of the studies compared patients with pTBI with controls without any form of injury, rather than using injury controls, potentially influencing the interpretation of the results. Our study revealed a consistent association between pTBI and ADHD, which increased with severity of TBI. In a smaller American cohort study (187 patients with pTBI), a association was also observed only with severe pTBI and ADHD, although the results had wide CIs (HR 3.62 (95% CI 1.59 to 8.26)).11 Additionally, another smaller meta-analysis (5 studies, 3023 patients with mild TBI) reported a pooled relative risk of 2.2 (z=3.5, p<0.0005), aligning with our study but on a smaller scale.10 This included studies with populations under 1000. However, our study focused solely on post-traumatic ADHD medication usage rather than ADHD diagnosis, as seen in previous studies. Therefore, our findings may differ from those of previous research. These collective findings highlight the complexities of understanding the pTBI severity-ADHD relationship and underscore the need for more extensive research with larger cohorts and robust statistical power, considering the limitations of previous studies.

ADHD emerges as a notable complication following pTBI, posing a threat to the patient’s overall quality of life. Previous research suggests that patients with pTBI who develop ADHD may experience worse adolescent adjustment and executive functioning.19 20 Identifying individuals at risk for ADHD becomes crucial, underscoring the importance of long-term post-traumatic surveillance. Remarkably, the risk of ADHD persists for up to 20 years post-trauma, even in cases of mild TBI. Additionally, current studies signal a substantial rise in the incidence of pTBI, emphasising the urgency of comprehensive understanding and proactive measures in addressing this growing public health concern.8

Our study has several notable strengths, with a primary highlight being the exceptional quality of the registers incorporated.21–23 The consistent use of ICD-10 classification in Finland since 1998, along with standardised coding measures, enhances the reliability of our data throughout the study period. Previous successful research on neurological conditions using the Finnish Care Register for Healthcare and data from the Finnish Social Insurance Institution adds to the credibility of our chosen data sources.24 25 Another strength stems from the accessibility of free specialised healthcare visits for children in Finland, supported by a social insurance system funded through universal fees based on individual income.26 This ensures equitable healthcare provision, minimising potential socioeconomic biases. Both the high-quality registers and the reimbursement system are integral components of the Finnish national healthcare system, facilitating rigorous register studies. The robustness of our study is further bolstered by the substantial population size of almost 70 000 patients with pTBI, making it the largest study population to date. The incorporation of a 1-year washout period is a notable strength. The diagnostic process for ADHD is time-consuming, involving a prolonged surveillance period of symptoms by parents, caretakers and teachers. This approach enhances the probability that identified ADHD cases are indeed associated with pTBI rather than being influenced by pre-existing conditions or immediate injury consequences.27 28 Additionally, the limited variation in sex and age across study groups contributes to minimising potential sources of bias, thereby enhancing the validity of our results. Another strength is the use of distal extremity fracture patients as a reference group, which limits the confounding bias. Moreover, the size of the reference group, comprising over 60 000 individuals, is the largest ever studied in this subject area.

Our study has some limitations. First, the Finnish Care Register for Healthcare lacks information on patients’ familial medical history, preventing the assessment of genetic factors’ impact on ADHD. Given the dual influence of genetic and environmental elements on ADHD, this represents a significant constraint.29 30 The Finnish Care Register for Healthcare also lacked information on the race, ethnicity and region of patients. Another limitation is the usage of only sex and age at the time of hospitalisation as covariates in the analysis. This limitation is due to our limited dataset and could potentially inflate the association between pTBI and ADHD medication utilisation by causing residual confounding. Additionally, registry data may contain diagnostic inaccuracies stemming from pTBI coding by healthcare practitioners. Third, the study data exclude information from primary care, potentially leading to an underestimation of less severe injuries typically managed by primary care providers. Fourth, the registers used lack GCS ratings and other severity indicators, limiting the assessment of TBI severity. Fifth, our data only capture the initial TBI, thus preventing the evaluation of cumulative risk associated with multiple TBIs of varying severities and their potential link with ADHD. Another limitation of our study is that we focused solely on the use of ADHD medication rather than directly diagnosing ADHD. This limitation arose primarily due to the constraints of our available data. Consequently, our study does not include individuals with ADHD who may not have sought medical assistance or who may have been managed conservatively. Finally, while our implementation of a 1-year washout period ensures a stronger likelihood that ADHD medication is linked to the traumatic event, the exclusion of this period may introduce some immortal time bias. A total of 421 patients were excluded due to 1-year washout period.

As the incidence of both pTBI and ADHD is on the rise, it is crucial to conduct more extensive research on TBI-associated ADHD. Future research should involve larger cohorts with injury reference groups and consider additional covariates such as genetic factors, pretraumatic cognitive performance and information on premature birth. Furthermore, the severity of pTBI and the occurrence of ADHD should be studied more extensively to explore the impact of different treatment strategies for pTBI. More extended surveillance periods should be considered to capture long-term outcomes. This approach would significantly advance our understanding of the complex dynamics involved in this interaction. Moreover, it would enable us to provide better psychoeducation to parents and caregivers, leading to improved management and support for affected individuals. Furthermore, our findings underscore the importance of monitoring ADHD symptoms post-trauma for many years, emphasising the necessity of long-term follow-up and care.

The results of this extensive nationwide cohort study reveal a link between pTBI and an increased likelihood of post-traumatic ADHD medication. This association was evident also when analysing males and females independently. These findings underscore the importance of preventive measures for TBIs and shed light on the potential impact of long-term post-traumatic monitoring or treatment in alleviating the incidence of ADHD.