Suicide Attempt and Suicidal Drug Overdose in Chronic Obstructive Pulmonary Disease Patients With or Without Depression.

Background To determine differences in the incidence and risks of suicide attempt (SA) and suicidal drug overdose (SDO) between chronic obstructive pulmonary disease (COPD) patients with and without comorbid depression by using data from Taiwan’s National Health Insurance Research Database. Methods We analyzed the data of patients aged ≥20 years who had received a COPD diagnosis between 2000 and 2012. These COPD patients were divided into those with and without depression, and they were compared against a cohort from the general population. We calculated adjusted hazard ratios and the corresponding 95% confidence intervals for SA and SDO in the three cohorts after adjustment for age, sex, and comorbidities. Results Until the end of 2012, 5.81% of patients with COPD developed depression. The incidence of SA and SDO in COPD patients with and without depression was 29.7 and 4.69 per 10,000 person-years and 71.2 and 20.9 per 10,000 person-years, respectively. COPD patients with depression had 13.6- and 10.0-fold higher risks of SA and SDO, respectively, than did controls. Particularly, an increased risk of SA caused by the enhancement effects of depression on COPD was noted in patients aged less than 50 years. Conclusion SA and SDO risks are extremely high in Taiwanese COPD patients with depression. Our findings suggest that clinicians should be aware that for COPD patients with comorbid depression, prescribing a large amount of medications may be associated with SA risk through SDO.


INTRODUCTION
Globally, more than 800,000 people commit suicide every year, and the prevalence of completed suicide has progressively increased over the past decades (1). The annual prevalence of suicidal ideation and suicide attempt (SA) has been estimated to be 2.3% and 0.4% in Austria and 3.7% and 0.5% in the United States, respectively (2,3). In European countries, the estimated rate of suicidal lethality is much higher in men than in women (13.9% vs. 4.1%) (4). The main possible reason for this sex difference is the different methods of SA adopted by men and women; violent methods are more common among men, and self-intoxication or suicidal drug overdose (SDO) is the most frequently chosen method among women (5). Therefore, we believe that adequately identifying individuals at a high risk of SA and interrupting their SA are urgently required for preventing suicide (6).
Depression is a well-established predictor of suicidal behaviors (7,8). Researchers have investigated interactions between suicidality and medical conditions for years; however, only few researchers have conducted population-based studies on this topic (9,10). In a recent large population-based crossnational study, Scott et al. reported several physical conditions that are potential independent risk factors for suicidality (10). Whether the probability of suicide is high in patients with chronic medical conditions or whether it is noted only in patients with specific disorders remains controversial (11). Sanna et al. reported an association between suicidality and several medical disorders; however, no association was noted between suicidality and pulmonary disease (12).
The Taiwanese government established the National Health Insurance (NHI) program in 1995, which covers approximately 99% of the residents of Taiwan for the past 2 decades (13). Both chronic obstructive pulmonary disease (COPD) and depression are common chronic disorders in the National Health Insurance Research Database (NHIRD). Furthermore, pneumonia and COPD were the third and seventh causes of death in 2018 in Taiwan. As COPD is one of the major chronic physical disorders among elderly people, we believe that SA and SDO risks should be different between the general population and COPD patients with and without comorbid depression. Investigating these differences might be beneficial for the medical care systems in both Taiwan and other East Asian countries due to similarities in their ethnicity and cultural heritage (14).

Data Source
The NHIRD of Taiwan is maintained by the National Health Research Institutes of Taiwan. It contains comprehensive NHI records of almost 99% of the population of Taiwan (13). The database includes the de-identified medical records of inpatients and outpatients and their prescription drug and other medical service records. Only patients with hospitalization records were included in this study. Diagnoses were defined based on International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) codes. The Research Ethics Committee of China Medical University and Hospital in Taiwan approved the study (CMUH104-REC2-115-CR4).

Study Population
To assess the differences in SA and SDO risks between COPD patients with and without depression, 4 cohorts were defined in this study: the total COPD cohort comprising patients aged ≥20 years diagnosed with COPD (ICD-9-CM codes 491, 492, 494, and 496), the COPD with depression cohort comprising COPD patients with depression (ICD-9-CM codes 296.2, 296.3, 296.82, 300.4, and 311), the COPD without depression cohort comprising COPD patients without depression, and the comparison cohort comprising individuals without COPD and depression during 2000-2012. The index date was the date of COPD diagnosis. Controls (individuals without COPD and depression) were frequency matched based on sex, age, and index year at a ratio of 1:2. The primary outcome was the occurrence of SA (ICD-9-CM codes E950-E959) and SDO in NHIRD records. SDO was defined based on an emergency department visit, a clinic visit, or hospitalization with the ICD-9-CM codes 960-979 but without E codes. The study excluded patients aged <20 years and patients with SA or SDO before the index date. The same exclusion criteria were applied for the comparison cohort.

Statistical Analyses
Categorical and continuous variables in this study are presented by number (%) and mean ± standard deviation, respectively. The chi-square test for categorical variables and analysis of variance for continuous variables were performed to examine the differences in these variables among the cohorts. Incidence rates of SA and SDO were calculated by dividing SA and SDO events by person-years (every 10,000 person-years). Cumulative incidence curves of SA and SDO in the cohorts were plotted using the Kaplan-Meier method, and log-rank tests were applied to test the differences in the curves. Hazard ratios (HRs), adjusted HRs (aHRs), and 95% confidence intervals (95% CIs) of SA and SDO were calculated using Cox proportional hazard models to evaluate the risks of SA and  SDO among the cohorts. The level of significance was set at.05. All data were analyzed using SAS 9.4 software (SAS Institute Inc., Cary, NC, USA).
Moreover, to evaluate whether patients with COPD and depression might have higher SA and SDO risks than patients with depression without COPD, we further compared COPD patients with depression (n = 19,725) with non-COPD patients with depression who were frequency matched based on sex, age, and index year (Supplementary Tables). COPD patients with depression had a moderately high risk of SA (aHR, 3.59; 95% CI, 2.85-4.52) and SDO (aHR, 4.97; 95% CI, 4.20-5.88) than non-COPD patients with depression.
To assess the incidence and risk of different SDO types, patients with SDO were classified based on the different methods of suicide ( Specifically, COPD patients with depression had a higher incidence (10.5 per 10,000 person-years) and risk of poisoning through benzodiazepine-based tranquilizers (aHR, 4.97; 95% CI, 3.66-6.73) than those without depression.

DISCUSSION
Despite the urgency and necessity of reducing suicidality worldwide, several developed countries have not invested sufficient resources in SA research and prevention (5). An obvious imbalance exists between the magnitude of the suicide problem and the knowledge required to address it. The present study involved a 13-year follow-up of patients aged ≥20 years, and the results revealed that 5.81% of patients with COPD developed depression. One of our previous studies with a similar design showed that the incidence of SA and SDO in patients with depression was 3.76 and 6.97, respectively, per 10,000 personyears (15). The current study revealed that the incidence of SA and SDO in COPD patients with depression was 29.7 and 71. 10,000 person-years, respectively. COPD patients with depression had a 13.6-and 3.09-fold higher SA risk than did controls and COPD patients without depression, respectively. Furthermore, COPD patients with depression had a 10.0-and 3.09-fold higher SDO risk than did controls and COPD patients without depression, respectively. Contrary to the belief that COPD moderately increases SA and SDO risks in patients with depression, this large-scale study revealed enhancement of SA and SDO risks in COPD patients with coexisting depression. Furthermore, regardless of the confounding effects of socioeconomic factors such as monthly income, urbanization level, and occupation category, age less than 50 years is a specifically high risk factor for SA in COPD patients that interact with depression. Moreover, no differences were observed in SA risk between male and female patients in this study when considering the overwhelming enhancement effects of depression on COPD. We usually evaluate suicidality using self-reported passive and active death-related thoughts collected through a questionnaire with a binary (yes/no) response. Using this method, suicidal ideation can be investigated; however, valid methods for assessing suicidal behavior, SA, and suicidal death risk are unavailable. Moreover, in the United Kingdom, 1 of 5 adults considers suicide; however, only one in 15 attempts suicide (6). Another study reported that only a small proportion of individuals who exhibit suicidal ideation actually attempt suicide (16). Our study showed proportionally extreme increases in SA and SDO risks resulting from the effects of comorbid depression on COPD. Compared with the annual SA prevalence of 0.4%-0.5% in CI, confidence interval; HR, hazard ratio; PY, person-years. # Incidence rate per 10,000 person-years. $ Multivariable analysis included age, monthly income, urbanization level, occupation category, and comorbidities of schizophrenia, alcohol-related illness, anxiety, mental disorders, and insomnia. † New Taiwan Dollar (NTD), 1 NTD is equal to 0.03 USD. ‡ The urbanization level was divided into 4 levels based on the population density of the residential area, where level 1 was the most urbanized and level 4 was the least urbanized. & Other occupation categories included those who were primarily retired, unemployed, and low-income people. *P < .05, **P < .01, ***P < .001.
Western countries (2,3), the prevalence in our patients with COPD and depression was lower than expected. However, as a nonviolent method of SA, SDO is frequently noted in Taiwanese patients with COPD and depression. Although more than 95% of the people who attempt SDO could eventually survive in Europe (17), SDO causes considerable burden in Asian countries; hence, SA through SDO should be effectively prevented (18,19). Our results implied that COPD patients could be dangerously abuse prescribed medicines and tranquilizers for attempting suicide through overdose, especially when these patients have comorbid depression. This study reiterates that medicines and tranquilizers used for self-poisoning are highly accessible in Taiwan, because of the restricted access to firearms in Taiwan, similar to most Asian countries. Therefore, careful and long-term accurate assessment should be performed before prescribing multiple medicines or tranquilizers to COPD patients with depression. Moreover, rather than focusing on their suicidal ideation only, awareness of the relationships of different physical disorders with depression would help identify patients at a high SA risk, thereby preventing suicidal death. However, depression mediates the relationship between value strain, deprivation strain, aspiration strain, coping strain, and suicidality (8). Psychological strain may cause suicidal behaviors or impulsive suicide in patients with a lack of social support/ coping capacities (20), poor functional status (21,22), and disorders of subjective perception including anxiety, panic, or other mental disorders (23,24). Those might partially explain the enhancement of depression on suicidal behaviors in COPD patients. Nonetheless, the actual interactions between COPD and depression on suicidality warrant additional investigational studies. Given the neuro-psychological pathology of COPD, considerable biological evidence has demonstrated that chronic inflammation, such as that in COPD, may increase the pathologies over the blood-brain barrier, glutamate regulation, microglia activation, and autoimmune response (25,26). These can affect patients' physical endurance, sleep, chronic cognition, and associated psychosocial sequelae, thereby contributing to patients' suicidality. Furthermore, oxidative stress in COPD could decrease serotonin synthesis owing to the effects of hypoxemia on tryptophan hydroxylase, and oxidative stress New Taiwan Dollar (NTD), 1 NTD is equal to 0.03 USD. ‡ The urbanization level was divided into 4 levels based on the population density of the residential area, where level 1 was the most urbanized and level 4 was the least urbanized. & Other occupation categories included those who were primarily retired, unemployed, and low-income people. § Individuals with schizophrenia, depression, alcohol-related illness, anxiety, mental disorders, and insomnia were classified into the comorbidity group. ***P < .001. could cause depression and suicidality (27). These findings revealed a notably bidirectional relationship between depression and COPD that increases suicidality in affected patients.
Our findings proved that depression is an extreme factor enhancing suicidal behaviors in patients with COPD. However, this study has some limitations. First, this was a retrospective study that used NHIRD claims data with hospitalization records related to COPD and depression during the study period. Although the accuracy of inpatient data is assured because of severe penalty associated with providing incorrect data, identifying SA, SDO, COPD, and depression based on the ICD-9-CM coding system is still associated with a possibility of underdiagnosis and underestimation. Minor SA and SDO in patients with COPD might not require medical services, resulting in the underestimation of SA and SDO risks. Second, we could not directly contact patients to determine the severity of COPD and depression, their actual taken medications, or treatments for their disorders because patients' identities were anonymized in the NHIRD. These details may confound their suicidality. Finally, even though our study design controlled for numerous confounders to the best of our knowledge, some unmeasured or unknown confounders may remain. Further large-scale studies are necessary to gain a comprehensive understanding of suicidality in patients with COPD with and without depression; such studies would help establish an effective prevention system in Taiwan and worldwide.

CONCLUSIONS
COPD and depression increase the risk of suicidality individually and synergistically. COPD patients with depression have increased SA and SDO risks. Specifically, COPD patients with depression aged less than 50 years have high SA risk. Our findings provide crucial information for suicide prevention for clinicians and the governments of Taiwan and other Asian countries. In COPD patients with depression, long-term medicines or tranquilizers should be prescribed and dispensed with caution.

DATA AVAILABILITY STATEMENT
The datasets generated for this study are available on request to the corresponding authors.

ETHICS STATEMENT
The studies involving human participants were reviewed and approved by the research Ethics Committee of China Medical University and Hospital in Taiwan approved the study (CMUH104-REC2-115-CR4). Written informed consent for participation was not required for this study in accordance with the national legislation and the institutional requirements.

AUTHOR CONTRIBUTIONS
Study conception/design: TH, C-HK. Provision of study material and patients: C-HK. Collection and assembly of data, data analysis and interpretation, manuscript writing, and final approval of manuscript: all authors.