DOI: https://doi.org/10.21203/rs.2.19129/v1
Background
This study aims to investigate suicide risk within one year of receiving a diagnosis of cognitive impairment in older adults without mental disorders.
Methods
This study used National Health Insurance Service-Senior Cohort data on older adults with newly diagnosed cognitive impairment including Alzheimer’s disease, vascular dementia, other/unspecified dementia, and mild cognitive impairment from 2004 to 2012. We selected 41,195 older adults without cognitive impairment through 1:1 propensity score matching using age, gender, Charlson Comorbidity Index, and index year, with follow-up throughout 2013. We eliminated subjects with mental disorders and estimated adjusted hazard ratios (AHR) of suicide deaths within one year after diagnosis using the Cox proportional hazards models.
Results
We identified 49 suicide deaths during the first year after cognitive impairment diagnosis. The proportion of observed suicide deaths was the highest within one year after cognitive impairment diagnosis (48.5% of total); older adults with cognitive impairment were at a higher suicide risk than those without cognitive impairment (AHR, 1.89; 95% confidence interval [CI], 1.18–3.04). Subjects with Alzheimer’s disease and other/unspecified dementia were at greater suicide risk than those without cognitive impairment (AHR, 1.94, 1.94; 95% CI, 1.12–3.38, 1.05–3.58). Suicide risk in female and young-old adults (60–74 years) with cognitive impairment was higher than in the comparison group (AHR, 2.61, 5.13; 95% CI, 1.29–5.28, 1.48–17.82).
Conclusions
Older patients with cognitive impairment were at increased suicide risk within one year of diagnosis. Early intervention for suicide prevention should be provided to older adults with cognitive impairment.
Suicide is one of the most significant public health issues among older adults [1, 2]. Although suicide attempts are more frequent in young people in most countries, the mortality rate of suicide is higher among older adults [3, 4]. In 2016, South Korea was ranked 10th globally by suicide rate (20.2 per 100,000 people); the rate among older adults (≥ 60 years) in the same year was 53.3 per 100,000 population, which was the highest in the Organization for Economic Co-operation and Development countries [5, 6]. Considering the rapid growth rate (mean, 4.1% annually) and higher proportion of the elderly in the total population (14.3% in 2018) as compared to the global situation (8.2%) [7], there is a need for continued research on the causes of suicide in South Korean older adults.
Cognitive impairment is one of the most common health concerns among older adults [8] and a review of global figures on dementia prevalence reported an age-standardized prevalence for individuals over the age of 60 of between 5% and 7% [9]. Cognitive impairment includes mild cognitive impairment and diverse types of dementia, and is correlated with a higher risk of death and disability [10]. Cognitive impairment results in a significant reduction in the quality of life [8], and the prevalence rate of depressive disorders among patients with cognitive impairment is also high [11, 12]. This distress may play a role in the development of suicidal ideation and attempts.
The evidence for suicide risk among patients with cognitive impairment is unclear [13]. Most studies report that death by suicide is infrequent among patients with dementia [14–17], whereas a Danish study indicated that such patients have a higher suicide risk than others [18]. However, the participants in the Danish study were those with a dementia diagnosis during full-time admission in somatic or psychiatric hospitals. Psychiatric disorders are strongly associated with suicide risk [19] and it is necessary to identify suicide risk by cognitive impairment itself.
Further, little is known about suicide risk in patients with mild cognitive impairment, which is an intermediate cognitive state between normal aging and dementia—a definition implying neither a specific outcome nor etiology [20]. Suicide risk in mild cognitive impairment may be different from dementia because severe cognitive impairment, such as advanced dementia, could protect against suicide deaths by decreasing the capacity to achieve a suicide plan [21].
Although a recent study showed nonsignificant results for the relationship between cognitive impairment and suicide death [22], the data were obtained from few hospitals, limiting the generalizability of the results. The present study is based on prior literature and uses nationally representative data to present evidence for suicide risk following cognitive impairment diagnosis among older adults. We focused on suicide risk within one year of a diagnosis of cognitive impairment because the risk of suicide deaths may be high in the short term after diagnosis. Previous studies reporting that suicide risk is highest within one year after clinical diagnosis support our hypothesis [23–27].
Therefore, we examine suicide deaths during the first year after a principal diagnosis of cognitive impairment (Alzheimer’s disease, vascular dementia, other dementia, and mild cognitive impairment) among older adults. We also explore suicide risk in older adults with cognitive impairment according to gender and age group.
This study used the National Health Insurance Service-Senior Cohort (NHIS-SC) data supplied by the Korean National Health Insurance Service. The NHIS data covered a 10% random sample of the elderly population aged ≥ 60 years (5.5 million) in 2002. Unless disqualified because of emigration or death, participants were followed up for 12 years until 2013. Information for each individual’s cause of death was extracted from Statistics Korea and linked with our database. The NHIS-SC data include demographics and clinical information regarding diagnoses, treatments, and prescribed drugs for all medical institution visits.
Among 528,655 participants from the NHIS-SC between 2004 and 2012, we divided study subjects into older adults with cognitive impairment (n = 78,885) and those without cognitive impairment (n = 449,770). We defined older adults with cognitive impairment as those who were first diagnosed with mild cognitive impairment (F067) or dementia (F00–F03, F051, G30, G31) based on the International Classification of Diseases 10th Revision (ICD-10) code from January 1, 2004, to December 31, 2012. Dementia was further classified as Alzheimer’s disease (ICD-10 codes F00, G30), vascular dementia (ICD-10 codes F01), and other/unspecified dementia (ICD-10 codes F02, F03, F051, G31). Among subjects with cognitive impairment, we excluded older adults with cognitive impairment between 2002 and 2003 to include newly diagnosed cognitive impairment (n = 2,697). We also excluded subjects with mental disorders (n = 34,993) to identify suicide risk by cognitive impairment itself.
For the purpose of comparison, we selected the elderly without cognitive impairment (n = 469,770) from 2004 to 2012; subjects with cognitive impairment between 2002 and 2003 (n = 1,608) were excluded. We also excluded subjects with mental disorders (n = 135,056). We conducted a 1:1 propensity score matching using age, gender, Charlson Comorbidity Index, and index year as matching variables. The subjects were divided into young-old (60–74 years) and old-old adults (> 75 years) based on a previous study [28, 29]. The year of first diagnosis of cognitive impairment in the study period was defined as the index year, and the comparison group was matched with patients with cognitive impairment in the same index year. The Charlson Comorbidity Index corresponds to the sum of the weighted scores applied to diverse main medical conditions [30] and is measured by screening the year before the participants’ index year. After the propensity score was estimated, patients without cognitive impairment were matched with patients with cognitive impairment with the same (or similar) propensity scores. Finally, this study included patients with (n = 41,195) and without cognitive impairment (n = 41,195) (supplementary Fig. 1).
Suicide death, defined by the ICD-10 code X60–X84, within one year of the index date was the dependent variable. The index date of patients with cognitive impairment was the date of first diagnosis of cognitive impairment in the study period, and matched patients without cognitive impairment were allocated the same index date. The survival length was measured in days, and all subjects were followed up with until suicide, withdrawal from the medical security system (National Health Insurance subscribers or Medical Aid beneficiaries), death from other causes, or end of 2013, whichever occurred first.
Potential confounding factors were residential area, household income, disability, and insurance type; information regarding these factors was based on the index year of the subjects. Household income was classified as (1) low (< 40th percentile), (2) middle (41st–80th percentile), or (3) high (81st–100th percentile). Residential area was classified as metropolitan (capital), urban (local government where > 1 million people live), or rural (otherwise). Insurance type was classified as either National Health Insurance or Medical Aid, which are public medical support policies assisting poor individuals.
The subjects’ baseline characteristics were summarized by cognitive impairment. The statistical significance of the proportional differences in characteristics between patients with cognitive impairment and the comparison group was tested using Pearson’s χ2 test. We presented average person-years from the index date to the date of the final event and number of suicides within one year after discharge per group. The starting point of person-years was the index date for the first diagnosis. Based on person-years at risk and number of suicide deaths, we calculated the number of suicide deaths per 100,000 person-years. This study utilized the Cox proportional hazards models to evaluate suicide risk within one year after index date. The results were reported as estimates of adjusted hazard ratio (AHR) with a 95% confidence interval (CI). Events other than suicide death within one year from the index date were censored in the survival analysis.
First, we estimated AHR and 95% CI for suicide risk among older adults with cognitive impairment compared to those without cognitive impairment. Next, we analyzed suicide risk by subtypes of cognitive impairment (mild cognitive impairment, Alzheimer’s disease, vascular dementia, and other/unspecified dementia) compared to those without cognitive impairment. Finally, we measured suicide risk in older adults with cognitive impairment according to gender and age group compared to those without cognitive impairment. All dataset extraction and statistical analyses were conducted using the SAS 9.4 software, and proportional hazards assumptions were evaluated statistically and satisfied for all models.
Table 1 indicates the subjects’ baseline characteristics between 2004 and 2012. Prior to propensity score matching, all confounding factors between older adults with and without cognitive impairment were significantly different. After matching, approximately 69.3% of the subjects were females, and 27.9% and 72.1% were young-old and old-old adults, respectively, in both older adults with cognitive impairment and matched comparison groups. In both groups, 62.8% of subjects had comorbidities. The proportion of low household income in older patients with cognitive impairment was higher than in the matched group (41.1% vs. 35.7%). The proportion of individuals receiving Medical Aid in older adults with cognitive impairment was higher than in the non-cognitive impairment group (18.3% vs. 10.3%). Older adults with cognitive impairment were less likely to live in a metropolitan area compared to their non-cognitive impairment counterparts (14.0% vs. 18.1%). The proportion of disability in older adults with cognitive impairment was higher than in the non-cognitive impairment group (2.2% vs. 1.2%).
Figure 1 indicates that the number of suicides per period after cognitive impairment. Of 41,195 patients with cognitive impairment, 101 older adults died by suicide over the entire follow-up period. The median time from cognitive impairment diagnosis to suicide death was 1.08 years (range, 0.01–6.67 years), and the number of suicide deaths decreased progressively with time following cognitive impairment. In the first seven years after cognitive impairment diagnosis, 49 (48.5% of the total), 18 (17.8%), 17 (16.8%), 8 (7.9%), 3 (3.0%), 3 (3.0%), and 3 (3.0%) suicides occurred, respectively.
Table 2 shows the suicide risk within one year after a diagnosis of cognitive impairment. Model 1 indicates that older adults with cognitive impairment were at a higher suicide risk than the comparison group (AHR, 1.89; 95% CI, 1.18–3.04). Model 2 shows that suicide risk among older adults with Alzheimer’s disease (AHR, 1.94; 95% CI, 1.12–3.38) and other/unspecified dementia (AHR, 1.94; 95% CI, 1.05–3.58) was higher than in those without cognitive impairment. However, vascular dementia (AHR, 2.26; 95% CI, 0.93–5.49) and mild cognitive impairment (AHR, 0.96; 95% CI, 0.23–4.04) were not significantly correlated with suicide risk.
Table 3 shows suicide risk within one year after a cognitive impairment diagnosis according to gender and age group. In the female group, older adults with cognitive impairment were more likely to die by suicide compared to those without cognitive impairment (AHR, 2.61; 95% CI, 1.29–5.28). Among young-old adults, subjects with cognitive impairment were at a higher suicide risk compared to those without cognitive impairment (AHR, 5.13; 95% CI, 1.48–17.82). However, male (AHR, 1.41; 95% CI, 0.74–2.70) and old-old adults (AHR, 1.50; 95% CI, 0.88–2.53) with cognitive impairment were not significantly associated with suicide risk compared to those without cognitive impairment.
This study presents three main findings. First, suicide risk in older adults who had received a diagnosis of cognitive impairment in the preceding year was significantly higher than in those without cognitive impairment. Second, regarding the subtypes of cognitive impairment, older adults with Alzheimer’s disease and other/unspecified dementia had a higher suicide risk compared to the non-cognitive impairment group. Finally, suicide risk in female and young-old adults with cognitive impairment was higher than in the comparison group.
In most previous studies, suicide risk in dementia or cognitive impairment was nonsignificant, which differs from our findings [14–17, 22]. Margda et al. (2002) showed that dementia was not associated with suicide in individuals ≥ 65 years [14]. Harris and Barraclough (1997) showed that suicide risk is increased in all mental disorders except mental retardation and dementia [15]. Wiktorsson et al. (2010) found that factors correlated with suicide attempts were living alone, being unmarried, history of psychiatric treatment, low educational level, and prior suicide attempts; no relationship with dementia was observed [16]. Randall et al. (2014) reported that all mental disorders (depression, anxiety, substance use, and schizophrenia) except dementia were independently associated with suicide death [17]. An et al. (2019) added up evidence for nonsignificant suicide risk associated with cognitive impairment [22]. The most remarkable difference between our findings and those from previous studies pertains to the monitoring period for suicide risk after diagnosis. Previous studies implied that those who committed suicide might have had more preserved insight and less cognitive impairment at the time of death. In other words, better cognitive states enabling suicide execution and sustained insight into the diseases might increase the likelihood of suicide death in patients with dementia [31]. By contrast, as per our results, suicide risk in patients with mild cognitive impairment was not associated with suicide deaths. We suggest that the early period after diagnosis is an important factor in the prediction of suicide risk in older adults with cognitive impairment.
Contrary to the nonsignificant results presented in most previous studies, Erlangsen et al. (2008) showed that adults with dementia are at a greater suicide risk compared to those without dementia [18]. However, the results cannot be generalized to older adults without mental disorders. Our finding implies that diagnosis of cognitive impairment, not psychiatric disorders, increases the risk of suicide.
In our results regarding the subtypes of cognitive impairment, older adults with Alzheimer’s disease dementia were at a higher suicide risk, whereas vascular dementia was not significantly associated with suicide risk compared to the non-cognitive impairment group. Alzheimer’s disease and vascular dementia share many clinical signs and symptoms. Both are characterized by cognitive decline, functional deterioration, and neuropsychiatric symptoms that may present as behavioral alterations [32]. However, some differences have been observed. Whereas memory and language function deficits prevail in Alzheimer’s disease, executive frontal lobe cognitive functions, such as attention, planning, and speed of mental processing, are more impaired in vascular dementia [33]. Motor function is not as greatly affected by Alzheimer’s disease [34, 35] as by vascular dementia [36, 37]. We suggest that patients with Alzheimer’s disease who have relatively less impaired ability to plan and execute suicide are at higher risk for suicide death. In South Korean clinical circumstances, if specialists cannot clearly identify dementia, they tend to diagnose patients with other/unspecified dementia, which may include other psychiatric disorders. We find it difficult to elucidate the mechanisms underlying the reason why older adults with other/unspecified dementia are at an increased risk for suicide, suggesting the need for further research.
This study demonstrated that suicide risk in female and young-old adults with cognitive impairment was higher than in the comparison group. Previous studies in the general population reported that male and old-old adults were at a higher suicide risk [28, 38]. These results imply that the characteristics that influence suicide in older adults with dementia differ from those that are significant in the general elderly population, and suicide prevention strategies for older adults with cognitive impairment should be specifically tailored to their characteristics.
This study has several limitations. First, we could not determine cognitive impairment severity from the medical records. A previous study revealed that severity of cognitive impairment measured by Clinical Dementia Rating scale was not associated with suicide risk [22]. Despite nonsignificant results, future research is warranted to explore suicide risk by specific period considering the severity of cognitive impairment. Second, although lifetime personal history of suicide attempts is a risk factor for death by suicide, our database had no information on individuals who attempted suicide [39]. Finally, our study included only the South Korean elderly population. Therefore, caution needs to be exercised when generalizing the results to populations with dissimilar backgrounds (e.g., young populations or those in countries with lower suicide rates than South Korea).
This is the first longitudinal cohort study to explore suicide risk within one year after cognitive impairment diagnosis by using nationally representative data. This study indicated that older adults with cognitive impairment are at an increased risk of suicide. Given the high suicide risk immediately after cognitive impairment diagnosis, early intervention for suicide prevention should be provided to older adults with cognitive impairment.
AHR:adjusted hazard ratio; CI:confidence interval; ICD-10:International Classification of Diseases 10th Revision; NHIS-SC:National Health Insurance Service-Senior Cohort
Ethics approval and consent to participate
The IRB of Yonsei University exempted this study from the need for ethical approval (7001988-201909-HR-686-01E). Consent to participate is not applicable.
Consent for publication
Not applicable.
Availability of data and materials
The Korean National Health Service System requires that all researchers agree not to share, report, or update the information. Every request concerning the study itself and data must be addressed to the relevant authors who have signed the Korean National Health Service System data release agreement. Certain researchers can demand access to the data directly from South Korea’s National Health Insurance Program.
Competing interests
The authors declare that they have no competing interests.
Funding
We gratefully acknowledge the research support by a grant from the National Research Foundation of Korea (grant number: 2019R1A2C1003259). This work was supported (in part) by the Yonsei University Research Fund (Post Doc. Researcher Supporting Program) of 2019 (project no.: 2019-12-0129).
Authors’ contributions
JWC, KSL, and EH designed the study. JWC and KSL performed the literature review and interpretation for data analysis. JWC analyzed the data. JWC, KSL, and EH wrote the draft. All authors read and approved the final manuscript.
Acknowledgments
The authors would like to thank Enago ([email protected]) for English language editing.
Table 1. Subjects’ general characteristics (2004–2012) |
|
|
|
|
|
||||
Variables |
Before Propensity Score Matching |
After Propensity Score Matchinga |
|||||||
With cognitive impairment |
Without cognitive impairment |
p-value |
Without cognitive impairment |
p-value |
|||||
N |
% |
N |
% |
N |
% |
||||
Total |
41,195 |
11.6 |
313,106 |
88.4 |
|
41,195 |
11.6 |
|
|
Gender |
|
|
|
|
< .001 |
|
|
1.000 |
|
|
Male |
12,633 |
30.7 |
143,552 |
45.8 |
|
12,633 |
30.7 |
|
|
Female |
28,562 |
69.3 |
169,554 |
54.2 |
|
28,562 |
69.3 |
|
Age |
|
|
|
|
< .001 |
|
|
1.000 |
|
|
60–74 |
11,501 |
27.9 |
236,676 |
75.6 |
|
11,501 |
27.9 |
|
|
≥ 75 |
29,694 |
72.1 |
76,430 |
24.4 |
|
29,694 |
72.1 |
|
Index year |
|
|
|
|
< .001 |
|
|
1.000 |
|
|
2004 |
1,676 |
4.1 |
268,093 |
85.6 |
|
1,676 |
4.1 |
|
|
2005 |
2,535 |
6.2 |
30,362 |
9.7 |
|
2,535 |
6.2 |
|
|
2006 |
3,245 |
7.9 |
8,446 |
2.7 |
|
3,245 |
7.9 |
|
|
2007 |
3,927 |
9.5 |
2,791 |
0.9 |
|
3,927 |
9.5 |
|
|
2008 |
4,614 |
11.2 |
1,482 |
0.5 |
|
4,614 |
11.2 |
|
|
2009 |
5,331 |
12.9 |
912 |
0.3 |
|
5,331 |
12.9 |
|
|
2010 |
5,996 |
14.6 |
543 |
0.2 |
|
5,996 |
14.6 |
|
|
2011 |
6,599 |
16.0 |
286 |
0.1 |
|
6,599 |
16.0 |
|
|
2012 |
7,272 |
17.7 |
191 |
0.1 |
|
7,272 |
17.7 |
|
Charlson Comorbidity Index |
|
|
|
|
< .001 |
|
|
1.000 |
|
|
0 |
15,317 |
37.2 |
200,683 |
64.1 |
|
15,317 |
37.2 |
|
|
1 |
11,445 |
27.8 |
64,128 |
20.5 |
|
11,445 |
27.8 |
|
|
2 |
7,471 |
18.1 |
28,845 |
9.2 |
|
7,471 |
18.1 |
|
|
≥ 3 |
6,962 |
16.9 |
19,450 |
6.2 |
|
6,962 |
16.9 |
|
Household income |
|
|
|
|
< .001 |
|
|
< .001 |
|
|
Low |
16,933 |
41.1 |
112,747 |
36.0 |
|
14,712 |
35.7 |
|
|
Middle |
10,992 |
26.7 |
107,542 |
34.3 |
|
12,052 |
29.3 |
|
|
High |
13,270 |
32.2 |
92,817 |
29.6 |
|
14,431 |
35.0 |
|
Insurance type |
|
|
|
|
< .001 |
|
|
< .001 |
|
|
National Health Insurance |
33,642 |
81.7 |
290,684 |
92.8 |
|
36,965 |
89.7 |
|
|
Medical Aid |
7,553 |
18.3 |
22,422 |
7.2 |
|
4,230 |
10.3 |
|
Residential area |
|
|
|
|
< .001 |
|
|
< .001 |
|
|
Metropolitan |
5,771 |
14.0 |
57,212 |
18.3 |
|
7,460 |
18.1 |
|
|
Urban |
9,405 |
22.8 |
67,248 |
21.5 |
|
8,782 |
21.3 |
|
|
Rural |
26,019 |
63.2 |
188,646 |
60.2 |
|
24,953 |
60.6 |
|
Disability |
|
|
|
|
< .001 |
|
|
< .001 |
|
|
Yes |
914 |
2.2 |
2,358 |
0.8 |
|
479 |
1.2 |
|
No |
40,281 |
97.8 |
310,748 |
99.2 |
40,716 |
98.8 |
|||
aAfter propensity score matching for gender, age, Charlson Comorbidity Index, and year of cognitive impairment diagnosis (index year) |
Table 2. Suicide risk within one year after a diagnosis of cognitive impairment |
|
|
|
|
|
|
|
|||
Model |
Variables |
Number of subjects |
Person-years |
Number of suicides |
Number of suicide deaths per 100,000 person-years |
HRa |
95% CI |
p-value |
|
|
|
||||||||||
Model 1 |
Non-cognitive impairment |
41,195 |
170,106 |
27 |
15.9 |
1.00 |
|
|
||
Cognitive impairment |
41,195 |
121,652 |
49 |
40.3 |
1.89 |
1.18 |
3.04 |
0.008 |
|
|
Model 2 |
Non-cognitive impairment |
41,195 |
170,106 |
27 |
15.9 |
1.00 |
|
|||
Alzheimer’s disease |
19,803 |
57,278 |
24 |
41.9 |
1.94 |
1.12 |
3.38 |
0.019 |
|
|
Vascular dementia |
4,218 |
13,269 |
6 |
45.2 |
2.26 |
0.93 |
5.49 |
0.071 |
|
|
Other/unspecified dementia |
14,045 |
42,208 |
17 |
40.3 |
1.94 |
1.05 |
3.58 |
0.033 |
|
|
Mild cognitive impairment |
3,129 |
8,897 |
2 |
22.5 |
0.96 |
0.23 |
4.04 |
0.957 |
|
|
Note. HR, hazard ratio; CI, confidence interval |
|
|
|
|
|
|
|
|
||
aHRs were estimated after adjusting for mental disorders, household income, insurance type, residential area, and disability |
|
|
|
Table 3. Suicide risk within one year of a cognitive impairment diagnosis by gender and age group |
|
|
|
|
|
|
|||||
Category |
Variables |
Number of subjects |
Person-years |
Number of suicides |
Number of suicide deaths per 100,000 person-years |
HRa |
95% CI |
p-value |
|
||
|
|||||||||||
Gender |
Male |
Non-cognitive impairment |
12,633 |
50,701 |
16 |
31.6 |
1.00 |
|
|
||
Cognitive impairment |
12,633 |
35,145 |
22 |
62.6 |
1.41 |
0.74 |
2.70 |
0.301 |
|
||
Female |
Non-cognitive impairment |
28,562 |
119,406 |
11 |
9.2 |
1.00 |
|
||||
Cognitive impairment |
28,562 |
86,507 |
27 |
31.2 |
2.61 |
1.29 |
5.28 |
0.008 |
|
||
Age group |
Young-old adults (60–74 years) |
Non-cognitive impairment |
11,501 |
56,193 |
3 |
5.3 |
1.00 |
|
|||
Cognitive impairment |
11,501 |
44,537 |
15 |
33.7 |
5.13 |
1.48 |
17.82 |
0.010 |
|
||
Old-old adults (≥75 years) |
Non-cognitive impairment |
29,694 |
113,913 |
24 |
21.1 |
1.00 |
|
||||
Cognitive impairment |
29,694 |
77,115 |
34 |
44.1 |
1.50 |
0.88 |
2.53 |
0.135 |
|
||
Note. HR, hazard ratio; CI, confidence interval |
|
|
|
|
|
|
|
|
|
||
aHRs were estimated after adjusting for household income, insurance type, residential area, and disability |
|
|
|
|
|