Preexisting Mental Disorders Increase the Risk of COVID-19 Infection and Associated Mortality

Coronavirus disease 2019 (COVID-19), a respiratory disease of unknown origin, has a high rate of morbidity and mortality. Individuals with mental disorders may have a higher risk of infection and worse clinical outcomes because of a variety of factors such as poorer general resilience and lower immune function. However, there have been no studies to date specifically investigating the risk of COVID-19 and associated mortality in these patients. This was addressed in the present study by analyzing the data of 473,958 subjects included in the UK Biobank, 14,877 of whom tested positive for COVID-19 infection. Logistic regression analysis was performed to evaluate the associations between mental disorders and risks of COVID-19 infection and associated mortality. The results showed that subjects who were diagnosed with a mental disorder had a significantly higher risk of developing COVID-19 and a worse outcome as evidenced by higher rates of COVID-19-related mortality, with the strongest effects observed for dementia. Among dementia subtypes, Alzheimer disease patients had the highest risks of COVID-19 infection (7.39-fold increase) and associated mortality (2.13-fold increase). Late-life anxiety only increased the risk of developing COVID-19 while late-life depression not only was associated with a higher risk of infection but also a worse outcome. These findings highlight the need to prioritize patients with mental disorders-especially those who experience these disorders later in life—when implementing preventive strategies such as vaccinations.


INTRODUCTION
Coronavirus disease 2019 (COVID-19), a respiratory disease of unknown origin caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), has a high rate of morbidity and mortality, affecting 163 million people and resulting in 3.3 million deaths in over 200 countries (https://www.worldometers.info/coronavirus/, accessed on 16th May 2021). Although neutralizing antibody rapidly develops after infection (1), a second episode of asymptomatic infection occurring 142 days after the first symptomatic episode in an apparently immunocompetent patient has been recently reported (2), suggesting that COVID-19 is a greater public health threat than initially assumed.
According to the World Health Organization, about 20-25% of the adult population is affected by mental disorders (https:// www.who.int/). These individuals may have a higher risk of various diseases and have worse physical health and treatment outcomes (3)(4)(5) because of a variety of factors such as shorter life expectancy, poorer general resilience, lower immunity (6,7), and higher susceptibility to infection (8). Recent studies have reported psychological risk factors for infections in general population and among frontline healthcare workers (9,10). Although several risk factors for COVID-19 infection and associated mortality have been documented [e.g., older age and the presence of complication; (11)(12)(13)], the risk for COVID-19 infection and associated mortality left largely unknown.
Recent studies have shown that individuals with a diagnosis of a mental disorder had a higher risk for COVID-19 infection and associated mortality. Li et al. found that patients with a psychiatric diagnosis had a higher risk of mortality compared with those without a psychiatric diagnosis, but the risk of subtypes of psychiatric diagnosis was not studied (14). Nemani et al. found that individuals with a recent psychiatric diagnosis of a schizophrenia spectrum was significantly associated with a higher risk of COVID-19 associated mortality in a small sample-sized cohort study, but those individuals with diagnoses of mood disorders and anxiety disorders were not associated with mortality (15). In a larger sample-sized cohort study by Wang et al. they found that individuals with a recent psychiatric diagnosis of an attention-deficit/hyperactivity disorder, a bipolar disorder, a depression and a schizophrenia had a significantly higher risk for COVID-19 infection (16). However, no such large sample-sized cohort studies concerning on systematically investigated the association of mental disorders (e.g., subtypes of dementia, anxiety, and depression) with COVID-19 infection and associated mortality, especially in a study.
Identifying high-risk populations is important as they can be prioritized in intervention strategies and receive timely and appropriate medical care. In the present larger sample-sized cohort study, we systematically evaluated the impact of middle aged and late-life adults diagnosed with a mental disorder-including psychotic disorders, substance use disorder, bipolar disorder, dementia, anxiety, depression, and depression comorbid with anxiety-on the risk for COVID-19 infection and associated mortality after adjusting for several confounding factors.

Subjects
After removing the death data (n = 28,547) before 31-November 2019, the remaining data of 473,958 subjects from 22 centers between March 2006 and December 2010 in the UK Biobank were analyzed (Figure 1). Confirmed COVID-19 infection was defined as at least 1 positive test result, 60,446 subjects of whom received COVID-19 test results up to February 24, 2021, with 14,877 subjects confirmed positive for COVID-19 infection (Figure 2). Subjects who did not undergo COVID-19 testing were deemed to have a negative COVID-19 test result. Subjects who received positive COVID-19 test results were classified into two categories: death due to and not due to COVID-19. These data were used to study the risk of COVID-19 associated mortality. Ethical approval was obtained from the North West Multi-Centre Research Ethics Committee (REC reference: 16/NW/0274).
Education level (Field ID 6138) was classified into two categories depending on the presence of college or university degree. The definition of obesity was a body mass index (BMI) >30 kg/m 2 . The definition of diabetes included selfreported type 1 or type 2 diabetes and a primary or secondary hospital diagnosis related to diabetes (ICD-10 codes E10-E14.9). Respiratory disease was defined as pulmonary embolism (code I26), other pulmonary heart diseases (code I27), influenza (codes J09-J11), pneumonia (codes J12-J18), acute bronchitis (code J20), acute bronchiolitis (code J21), unspecified acute lower respiratory infection (code J22), other diseases of the upper respiratory tract (codes J30-J39), chronic lower respiratory diseases (codes J40-J47), lung diseases due to external agents (codes J60-J70), other respiratory diseases principally affecting the interstitium (codes J85-86), and other diseases of the pleura and respiratory system (ICD-10 codes J90-99).

Statistical Analysis
Continuous variables are presented as mean ± SD, and categorical variables are presented as a number (percentage). We used unpaired t-test and χ 2 -test to compare differences between groups where appropriate.
Logistic regression analysis was used to evaluate associations between mental disorders and risk of COVID-19 infection and associated mortality (among subjects who received positive COVID-19 test results). A univariate model (Model 1) and multivariate models (Model 2-4, adjusted for several confounding factors) were applied to determine the Odd Ratios (ORs) and 95% confidence intervals (95% CIs). Two-tailed p < 0.05 were considered statistically significant.

Sample Characteristics
The demographic characteristics of the study population are presented in Table 1. Total of 8.9% subjects (n = 42,264) had a diagnosis of mental disorders, among whom 19.8% subjects (n = 8,378) had a chance to receive the COVID-19 test (21.9% subjects had a positive COVID-19 test result; n = 1,838). Among the 1,838 subjects, 8.9% (n = 164) deaths due to the COVID-19. However, only 12.1% subjects (n = 52,064) had a chance to receive the COVID-19 test among the remaining subjects who did not have a diagnosis of mental disorders (n = 4,31,694).
In general, subjects diagnosed with a mental disorder had higher COVID-19 infection and COVID-19-associated mortality rates than those without a mental disorder; this was not true of COVID-19 morbidity rates in subjects with secondary dementia, middle-aged depression, and middle-aged depression comorbid middle-aged anxiety.
Of the 14,874 subjects in the database who were positive for COVID-19, 664 (4.5%) subjects died due to the COVID-19. Among subjects diagnosed with a mental disorder, the highest rates of COVID-19 infection were among subjects with substance use disorder (5.4%), depression (5.3%), anxiety (3.2%), and dementia (2.0%), whereas subjects with psychotic disorder (0.3%), and bipolar affective disorder (0.4%) had the lowest rates. Among cases of COVID-19-related death, depression (9.8%), dementia (9.3%), substance use disorder (9.2%), and anxiety (5.1%) were the most common mental disorders; there were no deaths from COVID-19 among subjects with mixed dementia, middle-aged anxiety and middle-aged depression comorbid with anxiety. Subjects with a diagnosis of psychotic disorder did not have a higher risk of COVID-19 infection (p = 0.21) or COVID-19-related death (p = 0.23) than those without a mental disorder (Tables 2, 3). Substance use disorder and bipolar affective disorder were associated with a significantly higher risk of COVID-19 infection compared to absence of a mental disorder after adjusting for age, sex, education level, ethnicity, and BMI in model 3; however, after further adjusting for medical comorbidities (model 4), the higher risk was not significant. In the finally model, subjects with middle-age anxiety, or middleage depression did not have a higher risk of COVID-19 infection; those with anxiety did not have a higher risk of death from COVID-19; and those with substance use disorder did not have higher risks for COVID-19 infection and associated mortality than subjects without these mental disorders.

Association Between Mental Disorders and COVID-19 Incidence
Subjects with a diagnosis of a dementia (

DISCUSSION
This study of a large sized UK Biobank cohort yielded three novel findings. First, we document that subject diagnosed with a mental disorder had a significantly higher risk of COVID-19 infection and worse outcome (COVID-19-related death) than those without a mental disorder. Second, among dementia subtypes, Alzheimer disease exhibited the highest risk of COVID-19 infection (7.39-fold risk) and COVID-19-related death (2.13fold risk). Mixed dementia and vascular dementia did not The most striking and strongest risk factors for COVID-19 infection and associated death was dementia. In particular, subjects with Alzheimer disease had the highest risks of COVID-19 infection and associated mortality among all dementia subtypes. This finding was confirmed by a previous study (17).
There are a few possible explanations for this finding. Firstly, SARS-CoV-2 has been detected in the central nervous system of patients with dementia (18) and may cause brain damage and dysfunction (19). Patients with severe neurological dysfunction and injury are more likely to develop pulmonary diseases that further worsen clinical outcomes, in what is known as the brainlung-brain axis theory (20). Secondly, the risk genotype for dementia was shown to predict the severity of COVID-19 (21) and risk of COVID-19-related death (22). Thirdly, patients with  dementia often require the support of others to communicate their needs; however, the COVID-19 pandemic has reduced the availability of social support services for these patients (23).
Our results also suggest that anxiety and depression were only risk factors in subjects diagnosed with these disorders in latelife but not in middle age. The negative effects of anxiety and depression on immunity are not fully understood. Age is a known risk factor for COVID-19 infection (24). The immune system undergoes numerous age-related changes in a process known as immune senescence (25); we speculated that this is the reason for the higher risks of COVID-19 infection and associated mortality in late-life anxiety and late-life depression relative to middleaged anxiety and middle-aged depression. Late-life depression has been linked to an elevated risk of respiratory conditions (26) and is a risk factor for or early symptom of dementia (27), but not at younger ages (28). These findings may also explain why anxiety and depression were found to be risk factors for COVID-19 infection and associated mortality only in subjects diagnosed at older ages.
The strengths of this study are concerning the COVID-19 risk and associated mortality for dementia subtypes and age stratification for depression and anxiety with a large sized cohort. However, there are several limitations that should be addressed. First, subjects in the UK Biobank have a restricted age range, and therefore our data could not represent the whole population. Second, individuals identified with positive COVID-19 tests from the UK Biobank cohort may represent the hospitalized and severe cases, which might limit the generalization. Third, subjects who did not undergo COVID-19 testing were deemed to have a negative COVID-19 test result, and we did not know if some subjects were tested but not captured in our dataset. Fourth, some medical co-morbidities may not be listed in the hospital diagnosis, and therefore the number of reported medical comorbidities may be lower than the actual number.
In conclusion, our findings indicate that public health measures for COVID-19 such as vaccinations should prioritize individuals with mental disorders, especially those with both COVID-19 infection and dementia or depression. Future studies should explore the risk factors for COVID-19 in individuals with late-life-related diseases (e.g., dementia), given the increasing prevalence of these disorders in the global population.

DATA AVAILABILITY STATEMENT
Publicly available datasets were analyzed in this study. This data can be found here: UKBIOBANK.

ETHICS STATEMENT
The studies involving human participants were reviewed and approved by ethical approval was obtained from the North West Multi-Centre Research Ethics Committee (REC reference: 16/NW/0274). The patients/participants provided their written informed consent to participate in this study.

AUTHOR CONTRIBUTIONS
X-jD and YW had the idea for and designed this study, had full access to all the data in this study, take responsibility for the integrity of the data and the accuracy of the data analysis, critically revised the manuscript for important intellectual content and gave final approval for the version to be published, drafted the paper, and did the analysis. YW, YY, LR, YS, and WT takes responsibility for double check of the data analysis. All authors agree to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately resolved.