Traumatic Brain Injury and Risk of Schizophrenia and Other Non-mood Psychotic Disorders: Findings From a Large Inpatient Database in the United States

Abstract

Background

Traumatic brain injury (TBI) is linked with an increased risk of schizophrenia and other non-mood psychotic disorders (psychotic disorders), but the prevalence and contributing factors of these psychiatric conditions post-TBI remain unclear. This study explores this link to identify key risk factors in TBI patients.

Methods

We used the 2017 National Inpatient Sample dataset. Patients with a history of TBI (n = 26 187) were identified and matched 1:1 by age and gender to controls without TBI (n = 26 187). We compared clinical and demographic characteristics between groups. The association between TBI and psychotic disorders was explored using the logistic regression analysis, and results were presented as Odds ratio (OR) and 95% confidence interval (CI).

Results

Psychotic disorders were significantly more prevalent in TBI patients (10.9%) vs controls (4.7%) (P < .001). Adjusted odds of psychotic disorders were 2.2 times higher for TBI patients (95% CI 2.05–2.43, P < .001). Male TBI patients had higher psychotic disorders prevalence than females (11.9% vs 8.4%). Younger age, bipolar disorder, anxiety disorders, substance abuse, personality disorders, and intellectual disability are associated with an increased risk of psychotic disorders in men.

Conclusion

Our study found that hospitalized TBI patients had 2.2 times higher odds of Schizophrenia non-mood psychotic disorder, indicating an association. This highlights the need for early screening of psychotic disorders and intervention in TBI patients, calling for more research.

Introduction

Traumatic brain injury (TBI), caused by external mechanical forces, stands as a significant source of mortality and disability. In 2017, the United States recorded approximately 224 000 TBI-related hospitalizations and 61 000 deaths.¹ Between 2006 and 2014, TBI-related emergency room visits increased by 54%, while hospitalization and death rates saw reductions of 8% and 6%, respectively.² These changing trends are likely due to advancements in the early detection and management of TBI.^3,4

As survival rates improve, it’s imperative to understand TBI’s long-term impacts. Survivors often face enduring cognitive, behavioral, and emotional symptoms.⁵ One such consequence is schizophrenia and other non-mood psychotic disorders, a rare but severe outcome of TBI, occurring in 3.4%–8.9% of cases which profoundly affects the quality of life and daily functioning.⁶ Previous research has identified a 65% increase in the likelihood of developing psychotic disorders in patients with a history of TBI.^7–9

Damage to specific brain regions, such as the frontotemporal areas, limbic structures, gray–white matter junctions, and white matter tracts, can precipitate psychiatric conditions.^10,11 However the precise biological process underlying non-mood psychotic disorders induced by TBI is not fully understood. Nonetheless, atrophy in the hippocampal head is hypothesized to disrupt dopaminergic networks after TBI. This disruption might explain clinical observations such as the extended delay before symptom emergence and the predominance of positive symptoms in TBI-related psychotic disorders.¹² Typically, non-mood psychotic disorders onset post-TBI happens after a 1–3 year latency, hinting at its association with post-acute changes rather than initial injuries.^8,13,14 The delay in psychiatric symptoms onset could be due to progressive neuron degeneration and cellular function impairment.¹⁵ Unlike schizophrenia without TBI, whose onset is in late adolescent and early adulthood, TBI-related non-mood psychotic disorders display a broader range in onset age, often in older individuals.^16–19

TBI-induced psychotic disorders not only impacts patients but also their families. Several factors, including genetics, family mental health history, and age, influence this condition. Past research presents mixed findings. Many studies report a significant TBI-psychotic disorders link, particularly in those with a psychiatric illness family history.^10,20–24 However, a meta-analysis by Molloy et al⁹ suggests these results are confounded by varying head injury details, injury severity, and psychotic disorders definitions. Contrarily, Silver et al¹¹ found no increased TBI-psychotic disorders correlation, and Nelson et al¹⁴ observed fewer TBIs in schizophrenia patients.

This study aims to further explore the TBI-psychotic disorders relationship and associated risk factors by (1) evaluating the prevalence of schizophrenia and other non-mood psychotic disorders in patients with TBI compared to those without TBI. (2) Investigating the relationship between TBI-related non-mood psychotic disorders, demographic variables, and coexisting psychiatric and intellectual conditions. (3) Determining key risk factors for the development of schizophrenia and other non-mood psychotic disorders in individuals post-TBI. In this paper, the authors have used the term “psychotic disorders” to encompass schizophrenia and other non-mood psychotic disorders, as detailed in Supplementary material 1.

Methods

Dataset

The study data was obtained from the 2017 National Inpatient Sample (NIS) dataset.²⁵ The NIS dataset is part of the Healthcare Cost and Utilization Project (HCUP) sponsored by Agency for Healthcare Research and Quality (AHRQ). It contains information on clinical and nonclinical data elements for each hospital stay, including International Classification of Diseases, Tenth Revision, Clinical Modification/Procedure Coding System (ICD-10-CM/PCS) all listed diagnosis, procedures, external causes of morbidity, patient demographic characteristics (age, gender, race, median household income for ZIP code), hospital characteristics, payment source, discharge status, and hospitalization cost.²⁵

Study Population

Patients with a history of TBI were identified using the ICD-10 code “Z87820” (Personal history of TBI) from the records. Patients in the control group were selected from those without a history of TBI. To ensure a large sample size for meaningful results with minor differences, an age-gender-matched population from the control group was selected. Institutional review board (IRB) approval was not required as the dataset contains no protected health information.

Outcome

The primary outcome was the prevalence of psychotic disorders in patients with and without a history of TBI. Psychotic disorders diagnoses were obtained during hospital admissions using the diagnosis code provided in Supplementary table 1.

Variables

Data on age, gender, race, bipolar disorder, personality disorder, anxiety disorders, and substance use disorders (SUD) were obtained from the datasets. The ICD-10 codes for bipolar disorder and intellectual disability are provided in Supplementary table 1. The ICD-10 codes for personality disorders, anxiety disorders, and SUD were obtained from the NIS dataset’s Clinical Classifications Software (CCS) software.

Statistical Analysis

Mean (SE) was used for continuous data, and the percentage (95% CI) was used for categorical data. We performed group comparisons using the complex survey analysis technique implemented in SPSS software, version 26.0.²⁶ We applied the stratum, cluster, and weight variables as per the dataset guidelines. Independent sample t-test and Rao-adjusted Chi-square test were performed for group comparison. In our initial analysis, we applied a univariate unadjusted logistic regression model, incorporating a broad range of variables (Supplementary table 2). Subsequent to this phase, only variables showing statistical significance were retained for the final logistic regression model. This model included key predictors such as intellectual disability, personality disorders, age, gender, race, SUD, bipolar disorder, anxiety disorders, and history of TBI, to robustly determine the factors associated with psychotic disorders. Odds Ratio (OR) and 95% confidence interval (CI) were obtained. A P-value of less than .05 was considered statistically significant. Data was not reported if individual discharge records in any given cell of tabulated data were less than or equal to 10.

Matching Technique

The control group was selected through 1:1 propensity score matching with patients with a history of TBI. Matching was performed using the MatchIt package of R software. Age and gender were entered as predictors, and the nearest neighbor matching technique was used with a caliper size of 0.001.

Results

Baseline Characteristics

After applying the history of TBI diagnosis (ICD-10 code: Z87820), 26 187 discharge records were obtained. Baseline characteristics are shown in table 1. The average age of the study population was 52 years, and one-third of the population belonged to the 50–64 age group. The male patient population ranged from 69.1% to 70.4%, and white race patients accounted for 73.6% (72.4%–74.8%).

Table 1.

Demographic Characteristics of the Study Population

	Patients with the History of Traumatic Brain Injury
Unweighted/weighted count	26 187/130 935
Mean age (standard error)	52.2 (0.19)
Age groups
Age 12–17, %	2.2 (1.9–2.6)
Age 18–34, %	16.9 (16.3–17.5)
Age 35–49, %	22.7 (22.1–23.3)
Age 50–64, %	33.3 (32.6–33.9)
Age ≥65, %	24.9 (24.2–25.7)
Gender
Male, %	69.7 (69.1–70.4)
Female, %	30.3 (29.6–30.9)
Race
White, %	73.6 (72.4–74.8)
Hispanic, %	8.4 (7.7–9.1)
Black, %	12.8 (12.1–13.7)
Others, %	3.6 (3.1–4.2)

Data is provided as mean (standard error) or weighted percentage (95% confidence interval).

Psychotic Disorders

For patients without a history of TBI, the mean age was 57 years, and the majority were female (58%) (n = 6 169 996). To avoid large sample size and selection bias, age-gender matched populations were randomly selected from patients without a history of TBI diagnosis. The prevalence of psychotic disorders was 4.7% in the control group compared to 10.9% in patients with a TBI history (P < .001). Patients with a TBI history were 2.2 times more likely to have psychotic disorders than the control group after considering age, gender, race, and SUD. The prevalence of psychotic disorders was highest among the 18–34 and 35–49 age groups of patients with a history of TBI compared to the control group (figure 1). For patients aged ≥65, the prevalence of psychotic disorders was less among those with TBI but still higher than the control group. The 12–17 age group did not show any significant differences. In the stratified analysis by gender, 11.9% of male patients with TBI had psychotic disorders, compared to 8.4% of female patients with TBI. In the control group, 5.6% of male patients and 2.7% of female patients had psychotic disorders (figure 2).

Fig. 1.

Prevalence of psychotic disorders by history of traumatic brain injury (TBI) among different age groups.

Fig. 2.

Prevalence of psychotic disorders by history of traumatic brain injury (TBI) among male and female.

Our analysis included a sub-analysis for schizophrenia diagnoses, specifically for diagnostic codes F200, F201, F202, F203, F205, F2081, F2089, and F209. The prevalence rates of psychotic disorders were 4.5% in the study group and 2.2% in the control group. A separate analysis for Schizotypal (F21) was not feasible due to insufficient sample size in one group, adhering to HCUP NIS dataset guidelines of not reporting cell counts under 10.

In the univariate logistic regression analysis, age, gender, intellectual disability, personality disorders, SUD, race, anxiety disorders, and bipolar disorder were found to be significantly associated with psychotic disorders, as shown in Supplementary table 2. In the univariate analysis, the odds of developing psychosis in patients with a history of TBI were 2.48 (95% CI: 2.29–2.70). After adjusting for relevant covariates in the multivariable logistic regression analysis, a history of TBI still showed a significant association with psychosis, with an OR of 2.23 (95% CI: 2.05–2.43) (table 2).

Table 2.

Association Between History of Traumatic Brain Injury and Psychotic Disorders After Controlling for Age, Gender, Race, and Substance Use Disorders

	Psychotic Disorders (%)	No Psychotic Disorders (%)	Odds Ratio	P values
Age groups				<.001
Age 12–17	2.4	97.6	0.55 (0.37–0.82)
Age 18–64	9.4	90.6	1.81 (1.62–2.02)
Age ≥65	3.5	96.5	Reference
Gender				<.001
Male	8.8	91.2	Reference
Female	5.5	94.5	0.67 (0.62–0.73)
Past history of TBI				<.001
No	4.7	95.3	Reference
Yes	10.9	89.1	2.23 (2.05–2.43)
Substance use disorder				<.001
No	5.0	95.0	Reference
Yes	13.6	86.4	2.29 (2.12–2.46)
Race				<.001
White	6.6	93.4	Reference
Non-white	10.2	89.8	1.87 (1.72–2.04)
Bipolar disorder				<.001
No	7.1	92.9	Reference
Yes	17.1	82.9	1.71 (1.52–1.93)
Anxiety disorders				<.001
No	6.9	93.1	Reference
Yes	10.9	89.1	1.25 (1.15–1.36)
Personality disorders				<.001
No	7.5	92.5	Reference
Yes	24.1	75.9	2.22 (1.88–2.62)
Intellectual disability				<.001
No	7.5	92.5	Reference
Yes	26.0	74.0	4.21 (3.53–5.02)

TBI, traumatic brain injury.

In adolescents (age 12–17), the risk of psychosis was lower compared to those aged ≥65 (OR: 0.55, P < .001). Female gender was less likely to be associated with psychosis (OR: 0.67, P < .001). Substance use disorder was strongly associated with the risk of psychosis (OR: 2.29, 95% CI 2.12–2.46). Bipolar disorders, anxiety disorders, personality disorders, race, and intellectual disability also showed significant associations in the multivariable logistic regression analysis (P < .001).

Further, we analyzed codes F28, F29, F22, and F23, covering brief psychotic disorder, delusional disorder, and other unspecified psychotic disorders. Here, psychotic disorders prevalence was 2.9% in the study group and 1.0% in the control group.

Lastly, for schizoaffective disorders (codes F250, F251, F258, and F259), the prevalence was 4.1% in the study group and 1.7% in the control group.

In a focused sub-analysis on mood disorders with psychotic features, the TBI group showed higher prevalence: 1.0% for major depressive disorder with psychotic features and 1.2% for bipolar disorder with psychotic features, vs 0.4% in controls for both (P < .001).

Discussion

We aimed to investigate the association between TBI and schizophrenia and other non-mood psychotic disorders in a large US population. After adjusting for covariates, we found that patients with TBI had 2.2 times higher odds of experiencing psychotic disorders than control subjects. This aligns with the findings of a previous meta-analysis conducted by Molloy et al⁹ that showed that those with TBI had higher odds of experiencing psychotic disorders. The meta-analysis indicated a significant association between TBI and schizophrenia, with an OR of 1.65.⁹ However, the studies included had notable heterogenicity in study methodologies and findings. A large population-based study from Taiwan, which showed a twofold increased schizophrenia risk within 5 years post-TBI admission, supports our finding.²⁴ Furthermore, the finding that female gender is less likely to be associated with psychosis aligns with a previous study.²⁴ Our research further identifies risk factors for psychotic disorders post-TBI, including substance use, bipolar disorder, anxiety, personality disorders, and intellectual disability. These results underscore the multifaceted nature of the TBI-schizophrenia link, highlighting the importance of early psychiatric assessment and intervention in TBI patients to address potential psychotic disorders.

Previous research has proposed two possibilities: TBI could increase the risk of psychotic disorders, or patients with psychotic disorders may have a higher risk of TBI.^10,24 TBI increases the risk of psychotic disorders, especially for those with a family history. Moreover, TBI patients with substance use or bipolar disorder face a higher risk of psychotic disorders. Interestingly, previous studies have not consistently found a connection between TBI severity and the development of psychotic disorders. Psychotic disorders can occur after both mild and severe TBIs, irrespective of the type or location of the injury.¹⁸ This suggests a complicated link between TBI and psychotic disorders, with the potential for psychotic disorders to develop years after the initial TBI.²⁷ The relationship between TBI and psychotic disorders may also be bidirectional. People related to schizophrenia patients without psychotic disorder symptoms are more likely to have TBI, indicating that family factors may increase the risk of both TBI and psychotic disorders, as well as potential exacerbation by genetic factors.⁹ On the other hand, schizophrenia and other psychoses could make patients more vulnerable to TBI.

The precise mechanism by which TBI is responsible for psychotic disorders is not yet fully understood.

This delayed onset could be attributed to ongoing degenerative brain changes, particularly in the hippocampi, positioning TBI as a neurodegenerative process.^28,29 Evidence indicates deteriorative brain changes in chronic stages of moderate-severe TBI.^28,30

Prodromal symptoms often precede psychotic disorders after TBI, with positive symptoms like hallucinations and delusions being most common.^10,13 These symptoms respond to D2 receptor antagonism, suggesting dopaminergic system dysregulation.^31–33

However, unlike schizophrenia and other non-mood psychotic disorders typically show fewer negative symptoms indicating a pathophysiology more focused on dopaminergic networks and less on other neurochemical systems involved in negative symptoms.^13,34–38

However, the study by Abdelmalik et al²² discussed that TBI can cause diffuse axonal injuries and disrupt connectivity, which has been implicated in schizophrenia. Additionally, a TBI during development may create a compromised brain vulnerable to later expressing schizophrenia when challenged. The threshold theory proposed in the same study suggests that TBI lowers reserve capacity, making those predisposed to schizophrenia more likely to exceed a neuropathological threshold.²² Furthermore, head injury may also interact with genetic risk in a pathway contributing to schizophrenia pathogenesis.

Our study used an extensive database to show how TBI and psychotic disorders are related. Our study is vital because it is one of the few to have used a large sample, relied on medical records, and carefully controlled for variables, including SUD. However, as a cross-sectional study, we are limited in our ability to establish causality between TBI and psychotic disorders. This study also has other limitations as the NIS data used in our study represents inpatient hospitalizations and not unique patients, potentially leading to the inclusion of the same patient multiple times. As an observational study, our findings are also susceptible to confounding and selection bias. Additionally, the dataset has limitations regarding the number of variables available, such as lacking information on current or past medications and the duration of illness. Due to these study limitations, we couldn’t establish a cause-and-effect relationship between TBI and psychotic disorders. Other factors may also influence the relationship we found that we didn’t consider. More research is necessary to comprehend the connection between TBI and psychotic disorders.

Conclusion

This study utilizing a large inpatient database, offers valuable insights into the association between TBI and the development of schizophrenia and other non-mood-related psychotic disorders. Clinicians should screen for psychotic symptoms when treating TBI patients, especially young men with additional psychiatric disorders. Early intervention may prevent adverse outcomes in high-risk TBI groups. While mechanisms relating TBI to psychotic disorders require further study, multi-disciplinary management is essential given the considerable psychiatric impact possible with head trauma.

Supplementary Material

Supplementary material is available at https://academic.oup.com/schizophreniabulletin/.

Conflict of Interest

None of the above authors declared any conflict of interest.