Mental health problems, particularly major mental disorders (MMDs), have become a significant global health concern. Mental health problems affect 1 in 4 people, with about 450 million individuals having MMDs worldwide, including approximately 300 million with depression, 21 million with schizophrenia, and 46 million with bipolar disorder (BD) [1]. Importantly, people with MMDs may have an increased risk of certain types of cancer. A meta-analysis including 125,760 women demonstrated that schizophrenia was associated with a 1.32-fold increased risk of breast cancer compared with the general population [2]. However, the association between MMDs and cancer may vary across different types of cancer. For example, compared to the general population, one study suggested a lower prevalence of prostate cancer in male patients with schizophrenia but a higher prevalence of lung cancer in female patients [3]. Another study identified a higher incidence of colorectal cancer (CRC) in patients with schizophrenia compared to people with BD or the general population [4], and a crosscancer genome-wide association study (GWAS) demonstrated a moderate positive correlation between lung cancer with depression [5]. In addition, a positive genetic correlation was found between breast cancer and schizophrenia, while the associations between cancer and other MMDs were insignificant [5].

Among different cancers, CRC is of particular concern due to its high incidence and mortality. According to the 2018 Global Cancer Statistics, CRC is the second leading cause of cancer-related mortality worldwide [6], and the age-standardized incidence rate (62 per 100,000) of CRC is highest among all cancers in Taiwan [7]. The association between CRC and MMDs has been investigated in epidemiological studies. A literature review reported that the prevalence rates of anxiety and depressive disorders among patients with CRC ranged up to 47.2% and 57%, respectively [8]. A population cohort study in the US reported that patients with a CRC diagnosis had a 3.7-fold increased risk of any mental health disorder at 0 to 2 years after the diagnosis [9]. Another cohort study in Denmark found that patients with CRC also had a 12-fold higher risk of depression than people without cancer [10]. In addition, a Canadian cohort study in 2022 demonstrated a 1.17-fold higher risk of depression among males with CRC compared to the cancer-free group [11]. Many factors contribute to the association between CRC and MMDs. For instance, screen-detected cancer diagnosis can lead to short-term adverse psychological outcomes [12]. Another study indicates that anxiety, depression, and posttraumatic stress disorder are the most frequent psychological disorders in patients with breast cancer [13]. From a genetic point of view, a GWAS confirmed the significant relationship between the risk of CRC and schizophrenia [14].

Although previous studies have explored associations between MMDs and CRC, several key aspects of this relationship remain unclear and warrant further investigation. First, it remains unclear whether the association between CRC and MMDs extends to kinship, such as parents, siblings, or offspring. It is worthwhile to estimate the heredity effect on the association between CRC and MMD in a massive epidemiological study, which may support the genetic evidence [14]. In addition, no previous research has comprehensively assessed the association between CRC with various kinds of MMDs, such as schizophrenia, BD, major depressive disorder, autism spectrum disorder (ASD), attention-deficit/hyperactivity disorder (ADHD), obsessive-compulsive disorder, or general anxiety disorder (GAD) regarding kinship. To address this gap in the knowledge, we aimed to conduct a family linkage study with a nationally representative cohort to explore the associations between the risk of MMDs in first-degree relatives (FDRs) of individuals with CRC. Moreover, we also aimed to investigate sex differences in these associations.

The Taiwan Health and Welfare Data Science Center audits and releases the National Health Insurance Research Database (NHIRD) for research purposes upon formal application. This comprehensive database includes the healthcare data of approximately 99.7% of the population in Taiwan, comprising clinical visit dates, demographic details, and diagnoses. To protect patient privacy, the NHIRD contains only anonymized individual medical records. In the current study, we integrated three NHIRD datasets: the catastrophic illness database, a specialized dataset for mental disorders, and the registry of beneficiaries. The catastrophic illness database includes diagnoses of catastrophic illnesses such as malignancies and the specific diagnosis dates from 2000 to 2011 [15]. The diagnoses of malignant cancers are validated through histological confirmation, and patients issued with a catastrophic illness certificate are exempt from medical co-payment fees. The specialized dataset for mental disorders encompasses all medical records of insured individuals diagnosed with mental disorders from 2000 to 2011. The beneficiary registry database contains comprehensive demographic information including birthdates and residence for the entire Taiwanese population. Family kinships documented in the NHIRD were used for genealogical reconstruction based on the methodologies established by Cheng et al. [16] and Chen et al. [17] The diagnostic codes used in this study refer to International Classification of Diseases, 9th Revision, Clinical Modification (ICD-9-CM) codes. The Institutional Review Board of Taipei Veterans General Hospital approved the study protocol (approval number: TPEVGH-IRB-2018-07-016AC) and granted a waiver for informed consent due to the use of deidentified data, which required no direct contact with human participants. The NHIRD has been extensively used in various epidemiological studies conducted in Taiwan [18-20].

The FDRs of all individuals diagnosed with CRC (ICD-9-CM codes: 153, 154) were identified as the study group, and included parents, offspring, and siblings (Supplementary Figure 1). To minimize potential confounding effects related to age and sex, a 1:4 matched case-control analysis was performed based on age, sex, and familial relationships (Supplementary Figure 1). For example, a 15-year-old son of a mother with CRC would be matched with four 15-year-old sons of mothers without CRC. If this individual had two distinct familial relationships, such as son-mother and brother-sister, they would be counted in each relationship and matched on both counts. Participants in both the study and control groups were assessed for major psychiatric disorders and neurodevelopmental disorders including ASD (ICD-9-CM code: 299), ADHD (ICD-9-CM code: 314), schizophrenia (ICD-9-CM code: 295), BD (ICD-9-CM codes: 296 excluding 296.2, 296.3, 296.9, and 296.82), major depressive disorder (ICD-9-CM codes: 296.2 and 296.3), as well as any malignancy (ICD-9-CM codes: 140- 208). These psychiatric and neurodevelopmental disorders were diagnosed at least twice by board-certified psychiatrists. The urbanization level of residence (levels 1-5, most to least urbanized) was assessed as a proxy for healthcare availability in Taiwan [21].

For between-group comparisons, independent t-tests were used for continuous variables, while Pearson’s χ² test was used for nominal variables. Prevalence rates across the two groups were evaluated. Relative risks (RRs) and 95% confidence intervals (CIs) were calculated, adjusting for sex, year of birth, and level of urbanization, to determine the risks of malignancy, major psychiatric disorders, and neurodevelopmental disorders between the groups. In addition, analyses were conducted with further adjustments for individual cancer comorbidities to assess the coaggregation of CRC with major psychiatric disorders and neurodevelopmental disorders. As previously noted, each family cluster could encompass multiple familial relationships. To account for the effects of clustering, we used a Poisson regression model with robust error variance to estimate the RRs for the clustered data [22]. Sub-analyses stratified by kinship (parents, offspring, and siblings) were conducted to investigate differences in the risks of major psychiatric disorders and neurodevelopmental disorders between the groups. All statistical analyses were performed using SAS 9.2 (SAS Institute), with PROC GENMOD to estimate the adjusted RRs (ARRs). All tests were two-tailed, and a p-value of <0.05 was considered statistically significant.

The NHIRD is released and audited by the Department of Health and Bureau of the NHI Program for the purpose of scientific research (https://www.apre.mohw.gov.tw/). The NHIRD can be accessed through a formal application that is regulated by the Health and Welfare Data Science Center of Ministry of Health and Welfare, Taiwan.

In this study, 68,965 FDRs of individuals with CRC (38,024 males and 30,941 females; FDR-CRC group) and 275,860 FDRs of individuals without CRC (152,096 males and 123,764 females; matched controls) were included. There were no significant differences in sex and age between the case and the control groups. The FDR-CRC group had higher levels of monthly income and urbanization than the controls. Other details stratified by sex are presented in Table 1.

After adjustments for demographic variables, Charlson Comorbidity Index, monthly income, area of residence, psychiatric disorders, and cancer comorbidities, the FDR-CRC group had a significantly higher risk of cancer (ARR: 1.30, 95% CI: 1.23-1.37) than the controls (Table 2), including in both females (ARR: 1.25, 95% CI: 1.16-1.35) and males (ARR: 1.35, 95% CI: 1.25-1.45). Regarding MMDs, the FDR-CRC group had a higher risk of ASD (ARR: 1.34, 95% CI: 1.01-1.77) than the controls after adjusting for the abovementioned confounders. In addition, the male individuals in the FDR-CRC group had a higher risk of ASD (ARR: 1.47, 95% CI: 1.08-2.01) than the controls after adjusting for potential confounders. The risks of MMDs stratified by kinship (parents, offspring, and siblings) in the FDR-CRC group are shown in Figure 1. In addition, details of the risks of cancer and different MMDs in the FDR-CRC group stratified by kinship are presented in Supplementary Tables 1-7. In summary, the risks of CRC were mostly significant among all of the FDRs of individuals with CRC, including parents, offspring, and siblings (Supplementary Table 1). Sons of the CRC individuals had a higher risk of BD (ARR: 2.39, 95% CI: 1.28-4.44) than the controls, and especially for mothers of CRC (ARR: 2.63, 95% CI: 1.24-5.56) (Supplementary Table 3). Fathers of individuals with CRC had a higher risk of ASD (ARR: 1.54, 95% CI: 1.07-2.21) than the controls, and especially sons with CRC (ARR: 1.73, 95% CI: 1.15-2.61) (Supplementary Table 5). Mothers of male individuals with CRC had higher risks of ADHD (ARR: 1.35, 95% CI: 1.07-1.71) (Supplementary Table 6) and GAD (ARR: 1.12, 95% CI: 1.02-1.24) than the matched controls (Supplementary Table 8).

We found that the FDR-CRC group had a higher risk of cancer than the controls, regardless of sex. In addition, the FDR-CRC group had a higher risk of ASD than the controls, and this risk remained after multiple adjustments for potential confounders in males, especially in males who had a father with CRC. Moreover, the individuals, and especially females, who had a son with CRC had a higher risk of BD, and males who had a mother with CRC had higher risks of ADHD and GAD than the controls.

After adjusting for potential confounders, we found that the FDR-CRC group had a higher risk of ASD, especially in males who had a father with CRC. In addition to epidemiological evidence from a previous population-based study [23], multiple etiologies have been proposed to explain the association between ASD and cancer, including genetic dysregulation, nutritional habits, microbiota composition in the gut, stress, and inflammatory response [24]. With regards to genetic etiologies, recent evidence has shown a notable link between ASD and tumor progression, including cellular proliferation, differentiation, and migration [25,26]. On the other hand, inflammation may also be involved in the association between ASD and cancer, as inflammatory response plays an important role in neuronal development [27] and also in different stages of cancer progression [28]. Moreover, neurological dysregulation has also been implicated in the link between ASD and cancer, possibly through overlap in both carcinogenic and neural signaling pathways. For example, the PI3K/AKT/mTOR signaling axis has been associated with an inherited risk of both cancer and ASD [29,30]. Our finding of the association between FDRs of individuals with CRC and ASD may support the genetic evidence and inherited effect of a neurological etiology. Furthermore, we found that the risk of ASD was predominant in males who had a father with CRC. This finding may provide new insights for further genetic studies, such as Y-linked targets regarding cancer or ASD [31,32]. However, the limited effect sizes should also be considered in the interpretation of our findings.

In addition to ASD, we also found that males who had a mother with CRC had a significant risk of ADHD. A previous population-based study also demonstrated a significant association between CRC and ADHD [33]. We hypothesize that genetic factors may play a crucial role in this association due to the unique link between males and their mothers. Although research directly exploring the genetic association between ADHD and CRC is lacking, substantial genetic overlap between ADHD and lung cancer has been demonstrated [34]. Other potential etiologies for the association between CRC and ADHD include gut microbiota [35], inflammation [36,37], and dopamine dysregulation [38,39]. On the other hand, some of the individuals in the FDR-CRC group were also associated with the risk of BD or GAD. Several population-based studies have discussed the association between CRC and mood or anxiety disorder [40,41]. Many explanations have been proposed for the link between cancer and BD, including environmental and lifestyle factors [42]. However, genetic factors may be more relevant in our findings because only some of the FDRs were affected, especially mothers who had a son with CRC. Shared genetic predisposition has been proposed to enhance the association between BD and cancer [42-44]. Nevertheless, further genetic studies are warranted to confirm the genetic and hereditary factors regarding our findings.

There are several limitations to the current study. First, data on environmental and lifestyle factors are not available in the NHIRD, where these factors may also confound the interpretations of our results [42].

For instance, smoking and nicotine use disorder was highly associated with MMD [45,46], where smoking also increases risk of CRC [47]. Such history of problematic lifestyle and environmental factors may confound the association between the risk of an MMD and CRC. Second, although we adjusted for several confounders, residual confounding factors may still exist. However, a naturalistic study may better reflect real-world clinical practice. Third, as a case-control study, time effect of MMDs in FDRs on CRC cannot be identified, such as time ordering between them. For instance, whether the psychiatric diagnosis in the relative occurs before or after the CRC diagnosis date in the proband remains unclear. Fourth, detection or surveillance bias may exist in this registry-based study, as FDRs of individuals with CRC may have increased healthcare contact and thus a higher probability of receiving psychiatric diagnoses.

We identified an association between FDRs of individuals with CRC and the risk of MMDs, including ASD, ADHD, BD, and GAD. Moreover, some sex differences in the associations were identified. The clinical implication of this study is to extend the direct association between CRC and MMDs to FDRs of individuals with CRC. Previous evidence has discussed factors involved in the association between CRC and MMDs, such as an unhealthy lifestyle, substance use, adverse effects due to psychotropic drugs, chronic inflammation, insufficient access to preventative care, social isolation, and misdiagnosis from physical symptoms wrongly considered as psychiatric manifestations [42,48]. Our finding of the association in FDRs suggests the potential impact of genetic factors regarding the etiology between CRC and MMDs.

Moreover, we also identified some of sex differences in this association, further suggesting the potential role of genetic etiologies, such as targets on sex chromosomes. The novel association between the risk of MMDs and FDRs of individuals with CRC may provide further insights for future genetic studies to better understand the etiology underlying the association.