If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Send correspondence and reprint requests to Mattias Jonson, M.D., Department of Psychiatry and Neurochemistry, University of Gothenburg Institute of Neuroscience and Physiology, Wallinsgatan 6, 431 41 Mölndal, Sweden.
Center for Ageing and Health (Age Cap), Department of Psychiatry and Neurochemistry (MJ, RS, MMF, IS,MW), Gothenburg University, Gothenburg, SwedenNeuropsychiatric Epidemiology Unit, Department of Psychiatry and Neurochemistry, Sahlgrenska Academy, (MJ, RS, MMF,MW), University of Gothenburg, SwedenRegion Västra Götaland, Sahlgrenska University Hospital, Department of Affective Psychiatry, (MJ) Gothenburg, Sweden
Center for Ageing and Health (Age Cap), Department of Psychiatry and Neurochemistry (MJ, RS, MMF, IS,MW), Gothenburg University, Gothenburg, SwedenNeuropsychiatric Epidemiology Unit, Department of Psychiatry and Neurochemistry, Sahlgrenska Academy, (MJ, RS, MMF,MW), University of Gothenburg, SwedenRegion Västra Götaland, Sahlgrenska University Hospital, Department of Cognition and Old Age Psychiatry, (RS, IS), Gothenburg, Sweden
Center for Ageing and Health (Age Cap), Department of Psychiatry and Neurochemistry (MJ, RS, MMF, IS,MW), Gothenburg University, Gothenburg, SwedenNeuropsychiatric Epidemiology Unit, Department of Psychiatry and Neurochemistry, Sahlgrenska Academy, (MJ, RS, MMF,MW), University of Gothenburg, Sweden
Center for Ageing and Health (Age Cap), Department of Psychiatry and Neurochemistry (MJ, RS, MMF, IS,MW), Gothenburg University, Gothenburg, SwedenNeuropsychiatric Epidemiology Unit, Department of Psychiatry and Neurochemistry, Sahlgrenska Academy, (MJ, RS, MMF,MW), University of Gothenburg, SwedenRegion Västra Götaland, Sahlgrenska University Hospital, Department of Cognition and Old Age Psychiatry, (RS, IS), Gothenburg, Sweden
Center for Ageing and Health (Age Cap), Department of Psychiatry and Neurochemistry (MJ, RS, MMF, IS,MW), Gothenburg University, Gothenburg, SwedenNeuropsychiatric Epidemiology Unit, Department of Psychiatry and Neurochemistry, Sahlgrenska Academy, (MJ, RS, MMF,MW), University of Gothenburg, SwedenRegion Västra Götaland, Sahlgrenska University Hospital, Psychosis Department, (MW), Gothenburg, Sweden
What is the primary question addressed by this study?
Are different levels of passive and active suicidal ideation differently associated with all-cause mortality in older adults?
•
What is the main finding of this study?
A wish to die, but neither life-weariness nor active suicidal ideation was associated with increased all-cause mortality.
•
What is the meaning of the finding?
A wish to die may be a more important marker of mortality risk than other levels of suicidal ideation in older adults.
ABSTRACT
Objectives
To investigate potential differences in the strength of associations between different levels of passive and active suicidal ideation and all-cause mortality in older adults.
Design
Prospective cohort study.
Setting
Population-based samples of older adults in Gothenburg, Sweden.
Participants
Older adults aged 79 and above who participated in any wave of the Gothenburg H70 Birth Cohort Studies or the Prospective Population Study of Women between 1986 and 2015 (n = 2,438; 1,737 women, 701 men; mean age 86.6).
Measurements
Most intense level of passive or active suicidal ideation during the past month: life-weariness, wish to die, or active suicidal ideation. The outcome was all-cause mortality over 3 years.
Results
During follow-up, 672 participants (27.6%) died. After adjustments for sex, age, and year of examination, participants who reported a wish to die (HR 2.01; 95% CI 1.55–2.60) as the most intense level of ideation, but not participants who reported life-weariness (HR 1.40; 95% CI 0.88–2.21) or active suicidal ideation (HR 1.10; 95% CI 0.69–1.76) were at increased risk of all-cause mortality. Reporting a wish to die remained associated with mortality in a fully adjusted model, including somatic conditions, dementia, depression, and loneliness (HR 1.70; 95% CI 1.27–2.26).
Conclusion
In older adults, reporting a wish to die appears to be more strongly associated with all-cause mortality than either life-weariness or active suicidal ideation
from life-weariness (feeling that life is not worth living), to a wish to die, to thoughts of actively ending one's life (active suicidal ideation). Different forms of suicidal ideation have been associated with all-cause mortality. A wish to die or active suicidal ideation was associated with a 35% increase in mortality in an Australian population sample of older adults.
The increased risk was attenuated to 23% after adjusting for demographics, physical and mental health at baseline. Similar results were found in a recent Irish study on adults over the age of 50.
In a clinical sample of middle-aged veterans from the United States, the higher all-cause mortality among suicide ideators was not explained by differences in physical or mental health, nor by differential suicide rates.
Studies that have shown associations between suicidal ideation and mortality have defined suicidal ideation in various ways: Life-weariness, a wish to die, or active suicidal ideation,
found that 85-year-olds with life-weariness or a wish to die had increased mortality but not those with active ideation, but only 15 participants reported active ideation in that study. Raue et al.
found no difference in mortality between participants with a wish to die and participants with a suicidal desire, but only 13 participants reported a suicidal desire. Shiner et al.
found a non-significant tendency of higher all-cause mortality with increasing intensity of suicidal ideation, but only 23 participants reported the most intense level.
As previous research is inconclusive, further investigation of the association between suicidal ideation and mortality might shed more light on mechanisms. While it is possible that there is a dose-response relationship, with higher mortality as the severity of suicidal ideation increases, various types of ideation might be differentially associated with mortality, in a non-ordinal manner. We aimed to investigate this in a large, population-representative sample of older adults. A secondary aim was to examine possible confounders. We hypothesized that the highest mortality would be observed among those with the most severe suicidal ideation. Further, we hypothesized that differential mortality rates would be partly accounted for by demographic and psychosocial factors, and mental, and physical health.
METHODS
Sample
We drew participants, 79 years and older, from all examination waves of the Gothenburg H70 Birth Cohort Studies and the Prospective Population study of Women between 1986 and 2015 (Fig. 1). Examination procedures have been extensively described previously.
The prospective population study of women in gothenburg, sweden, 1968-69 to 1992-93. A 24-year follow-up study with special reference to participation, representativeness, and mortality.
Briefly, we invited persons living in both private households and in institutions, who were born on specific days each month, yielding a systematic population sample. Participation rates varied between 57.9% and 73.4% across waves.
Eleven examination waves, including 3,264 unique participants aged 79 years or older, took place between 1986 and 2015. For participants who had more than one examination (e.g. examined at ages 79 and 85) we utilized data from the first examination. Three hundred eighty-three participants lacked data on the Paykel questions at their baseline examination, and a further 443 lacked data on covariates, leaving 2,438 participants (1,737 women, 701 men). Participants were born between 1901 and 1930 and were between 79 and 101 years of age (mean 86.6, SD: 6.1) at baseline. To yield a similar follow-up time for all participants, we censored follow-up at 3 years past their first examination. Death dates were retrieved on 04-20-2021 from the Swedish Tax Agency.
which consists of five questions of increasing intensity that assess passive suicidal ideation of two forms (life-weariness and a wish to die) as well as active suicidal ideation: 1) Have you ever felt that life was not worth living? 2) Have you ever wished that you were dead – for instance, that you could go to sleep and not wake up? 3) Have you ever thought of taking your own life, even if you would not really do it? 4) Have you ever reached the point where you seriously considered taking your own life, or perhaps made plans on how you would go about doing it? 5) Have you ever made an attempt to take your own life? In case of a positive response to any of these questions, the participant was asked about the last time this occurred (past week, past month, past year, or longer than a year ago). For the main analysis, we categorized responses into three levels based on the most intense passive or active suicidal ideation experienced during the past month: Life-weariness (positive response to question 1, but not to question 2–5), wish to die (positive response to question 2, but not to question 3–5), active suicidal ideation (positive response to any of questions 3–5). For a sensitivity analysis, we categorized responses in the same manner, for any point in life (any of: past week, past month, past year, or longer than a year ago). The suicide item on the Montgomery Åsberg Depression Rating Scale (MADRS)
was also applied. The response options for that item are: 0) Enjoys life or takes it as it comes. 2) Weary of life. Only fleeting suicidal thoughts. 4) Probably better off dead. Suicidal thoughts are common, and suicide is considered a possible solution, but without specific plans or intention. 6) Explicit plans for suicide when there is an opportunity. Active preparations for suicide. Levels 1, 3, and 5 are midpoints between the other options. We categorized levels 0–1 as no suicidal ideation, 2–3 as passive ideation, and 4–6 as active ideation. Loneliness was assessed by the interviewer asking whether the participant felt lonely. Responses “never”, “seldom” and “sometimes” were regarded as not lonely while “often” was regarded as lonely. As this question was not included in two of the examination waves, for these waves we instead examined responses to the following two questions: Have you felt lonelier during the past 5 years due to functional decline or deaths of relatives? Do you find it hard to be alone? A participant with an affirmative response to either of these questions was regarded as lonely.
Information on somatic conditions (myocardial infarction, fracture of the hip, diabetes) was retrieved during the interviews by the interviewer asking the participant whether a nurse or a doctor ever had informed them that they had any of those conditions. In case of missing data on somatic conditions from the participant, information from close informants was used. Cause of death data were retrieved from the Swedish Cause of Death Register. We considered the underlying cause and categorized based on Swedish versions of ICD-9
codes: Cancer (140-208 or C00-C97); Respiratory disorders (460-519 or J00-J99); Ischaemic heart disease (410-414 or I20-I25); Stroke (430-438 or I60-I69); Dementia (290-294/330-337 or F00-F03); Suicide (including uncertain suicides) (E950-E959/E980-E989 or X60-X84/Y10-Y34).
Statistical Analysis
We used Pearson χ2 to test for group differences in proportions. We employed the Spearman rank-order correlation coefficient to test correlation between ordinal data. p-values were two-tailed and considered significant <0.05. We used Kaplan-Meier curves to plot survival by level of passive/active suicide ideation and Cox regression models to calculate Hazard Ratios (HRs) with 95% confidence intervals (CI) for mortality. Passive/active suicidal ideation was entered as a categorical variable with no passive/active suicidal ideation as reference. We conducted three models: 1) including sex, age, and examination year, 2) adding somatic conditions (history of myocardial infarction, history of hip fracture, diabetes) and dementia, and 3) adding also loneliness, and depression status (no/minor/major) as a categorical variable with no depression as reference. The proportional hazards assumption was tested by calculating Shoenfeld residuals; no obvious violations were found. In sensitivity analyses, we reran Cox regression models a-c using highest level of lifetime ideation. Further, we reran model a (using past month ideation) stratified by age category, and stratified by decade of examination. Finally, we ran models a-c above replacing levels based on the Paykel questions with levels based on the MADRS suicide item. Four participants had missing data on the MADRS suicide item and were excluded in this analysis. All statistical analyses were performed with IBM SPSS Statistics for Windows (Version 25), except for Shoenfeld residuals which were calculated in StataCorp. 2015. Stata Statistical Software: Release 14. College Station, TX: StataCorp LP.
RESULTS
Of all participants, 240 (9.8%) reported some type of passive or active suicidal ideation during the past month. A wish to die was most frequently the highest level reported (Table 1). Within three years of the baseline examination, 672 participants (27.6%) had died. Of those, 102 (15.2%) had reported any level of passive/active suicidal ideation at the baseline examination. Passive/active suicidal ideation tended to be more prevalent in older age categories, in earlier examinations, among women, among participants with a history of hip fracture, among participants with dementia, among those with minor or major depression, and among those who felt lonely (Table 1). The most common cause of death was ischaemic heart disease (22.5%), followed by cancer (14.4%), stroke (10.4%), respiratory disorders (7.9%), and dementia (4.5%). No participant died by suicide. Causes of death did not differ significantly by level of passive/active suicidal ideation (Pearson χ2 (df: 15): p = 0.876) (Supplementary material, Table S1).
TABLE 1Characteristics of Participants at Baseline Examination in Population-Based Samples of Older Adults, by Most Intense Level of Passive/Active Suicidal Ideation Reported During the Past Month
As illustrated in Figure 2, participants who reported a wish to die had the lowest survival. Adjusted for age, sex, and examination year (Table 2, model a), only a wish to die was associated with increased mortality. The hazard ratio for a wish to die compared to no ideation was slightly attenuated after including somatic conditions and dementia (model b). In the final model (c) that included also major and minor depression and loneliness, the hazard ratio was further attenuated, although still significantly elevated.
FIGURE 2Kaplan-Meier plot of survival by most intense level of passive/active suicidal ideation reported during the past month, in population-based samples of Swedish older adults (n = 2,438).
TABLE 2Cox Regression Models for Time to Death From all Causes Within 3 Years, in a Population Sample of Older Adults (n = 2,438), by Most Intense Level of Passive/Active Suicidal Ideation Reported During the Past Month
Participants
Events
Model a: Including Sex, Age, and Year of Examination.
Replacing level of ideation during the past month with highest level during lifetime in the Cox regression (Supplementary material, Table S2), attenuated the hazard ratios for wish to die somewhat, but did not change the overall pattern. In Cox regressions by age category (Supplementary material, Table S3), a wish to die was associated with mortality in age categories 85–94 and 95+, but not in those under 85 years. Both life-weariness and a wish to die were associated with mortality in those examined in the 1980s. Only a wish to die was associated with mortality in those examined in the 2000s and the 2010s. Active suicidal ideation was not associated with mortality in any decade or age category (Supplementary material, Table S4).
As responses on the MADRS suicide item and most intense level of ideation based on the Paykel questions differed somewhat (Spearman correlation coefficient 0.69, p <0.001), we reran Cox regression models a-c replacing the exposure variable with one based on the MADRS question. Passive suicidal ideation (MADRS levels 2–3) was associated with mortality in models a–c with successively attenuated hazard ratios (Supplementary material, Table S5). Active suicidal ideation (MADRS levels 4–6) was not associated with mortality in any model.
CONCLUSION
Contrary to our hypothesis of a dose-response relationship, only a wish to die was associated with all-cause mortality, after taking age, sex, and examination year into account. Although the hazard ratio was attenuated, a wish to die was still associated with a 70% increased risk of all-cause mortality compared to no ideation in the fully adjusted model including somatic conditions, dementia, depression, and loneliness. Results were corroborated by sensitivity analyses using an alternate measure of suicidal ideation, indicating that passive suicidal ideation was associated with mortality, while active suicidal ideation was not. Results were also corroborated by examining lifetime levels of suicidal ideation, with a wish to die conferring increased risk of mortality.
Our findings might seem somewhat counterintuitive as passive suicidal ideation is often considered less severe than active suicidal ideation, for example in suicide assessment scales.
The Columbia-Suicide Severity Rating Scale: initial validity and internal consistency findings from three multisite studies with adolescents and adults.
found that among older adults with major depression, clinical presentation did not differ in those with passive versus active suicidal ideation. Baca-Garcia et al.
found, in a study on U.S. adults, that desire for death and suicidal ideation were associated with similar risk of lifetime suicide attempt. Studies on Swedish,
have emphasized the importance of passive suicidal ideation as it seldom occurs in the absence of previous active suicidal ideation or current anxiety or depression. Further, a recent systematic review and meta-analysis found that passive and active suicidal ideation were associated with similar psychiatric morbidity and sociodemographic characteristics.
Characterizing the phenomenology of passive suicidal ideation: a systematic review and meta-analysis of its prevalence, psychiatric comorbidity, correlates, and comparisons with active suicidal ideation.
There are several possible explanations for the increased mortality among persons who wish to die. Firstly, we note that no participant died by suicide. Perhaps the most intuitive explanation is that those who wish to die also have poor physical health. This partly explained the excess mortality in the previously mentioned Australian
studies. Mirroring those results, we could see that the hazard ratio for a wish to die decreased, although only very slightly, after including somatic conditions in the model. There is of course a possibility of unmeasured and residual confounding, as our measurements of somatic health were limited. Another explanation is that a wish to die can be a symptom of depression, which is consistently associated with mortality,
A study from the United States on older adults found that among those with minor or major depression, those who also wished to die had higher mortality, if receiving treatment as usual. This difference was not observed in those who were treated at centers randomized to having depression care managers who helped physicians optimize depression treatment.
This supports an explanation that some of the excess mortality in older adults with a wish to die may be due to inadequate treatment of depression. However, also those who wished to die in the absence of depression in that study were at increased risk of mortality, controlling for functional disability and smoking status. This is in line with our results where a wish to die was associated with mortality also after controlling for minor and major depression.
A further possibility is that a wish to die is associated with mortality due to associations with social determinants of health, such as social support and feelings of belonging.
The role of perceived and objective social connectedness on risk for suicidal thoughts and behavior in late-life and their moderating effect on cognitive deficits.
However, a wish to die remained significant in the fully adjusted model. This remaining increased risk must then either be due to unmeasured or residual confounding, or due to one or several mechanisms, behavioral or biological, through which a wish to die shortens life.
Concerning causality between psychological states (such as a wish to die) and physical illness, Engel
suggested a psychological state he called the “giving up, given up”-complex, which may increase the risk of physical illness. This was based on observations that such a state often preceded physical illness. However, the direction of causality is not known, it is possible that undiagnosed physical illness presents as a negative psychological state. A Danish study suggested that mental stress may play a causal role in excess mortality based on a population-based study on bereavement. They found an increased risk of mortality, particularly during the first month after bereavement (HR 2.50; 95% CI 2.37–2.63), after controlling for demographic factors and physical and mental disorders.
As a wish to die may be perceived as stressful, this finding to some extent supports causality between a wish to die and death, although we cannot say if the mechanism is behavioral or biological. A further possibility is internalized ageism, which has been associated with premature mortality,
It could be that a wish to die is an indicator of internalized ageism. If so, this has clear implications for intervention.
We cannot say why the mechanisms discussed above may only, or more strongly operate for those with a wish to die, and not for those with other types of suicidal ideation. Cause of death did not differ by type of ideation, but the study was underpowered for such an analysis. Although we adjusted for some measurements of physical health, it is still possible that persons with a wish to die are more physically ill than persons with life-weariness or active suicidal ideation.
Another question is whether our results are clinically meaningful. The 70% increased mortality in the present study for persons with a wish to die can be compared to the 23% increased risk for older adults with a wish to die or suicidal ideation in the previously mentioned Australian study.
The authors of that study found that the population attributable mortality risk for suicidal ideation/a wish to die was similar to that of depression. As we found an even higher hazard ratio for a wish to die, we argue that our results are clinically meaningful. Older adults who report a wish to die require careful clinical assessment, for instance regarding poor self-care or demoralization even if they do not present with the typical major depressive picture.
Strengths and Limitations
Strengths of the study include the large and representative sample of the general older adult population in Sweden, the good response rate, and the comprehensive assessment of passive and active suicidal ideation.
Some limitations should be mentioned. First, fewer participants reported life-weariness and active suicidal ideation as highest level than who reported a wish to die as highest level. This might have underpowered the analyses in these groups. Although the difference in mortality between a wish to die and active ideation was fairly robust, the confidence intervals for life-weariness and a wish to die had a larger overlap, in line with the sensitivity analysis employing the MADRS question, indicating a higher risk for passive ideation. Second, the frailest participants, with the highest risk of death, are more likely to have not participated, based on the previously published slightly higher mortality among non-participants in some,
of the samples included in this study. This may have led to an under-estimation of the association between passive/active suicidal ideation and mortality, as we expect this group to have a higher prevalence of suicidal ideation. Third, some of the data emanate from interviews carried out several decades ago and might therefore not be generalizable to today's older adults. We note, however, that the association between death wishes and mortality was observed across examination decades. Fourth, loneliness was based on different questions in some examination waves, which might have affected analyses of that factor. Fifth, depression diagnoses were not specific for geriatric depression. This may have underestimated the contribution of depression on mortality. Sixth, as we only had cross-sectional data, we were unable to investigate the potentially mediating effect of health-related behavior. It is possible that those who wish to die make less healthy decisions regarding lifestyle, seek less healthcare and adhere less to treatment. Also due to the cross-sectional nature of our data, we cannot examine how changes in suicidal ideation and depression over time might influence mortality. However, the sensitivity analysis using lifetime ideation instead of past month yielded similar results.
AUTHOR CONTRIBUTIONS
MJ, RS, KAVO and MW designed this substudy. MW was main supervisor. IS designed the original population study and supervised data collection. MJ and RS carried out the data analyses. MJ prepared the first draft of the manuscript. All authors contributed to data interpretation and manuscript preparation, and all have approved the final manuscript.
DATA STATEMENT
The data has previously been presented as a digital poster, to conference attendees, at the International Summit on Suicide Research in October 25-27, 2021. It has also been presented orally, to conference attendees, at the Swedish Psychiatry Conference in Stockholm, March 17, 2022.
DISCLOSURES
MW has received Textbook royalties from Liber Förlag and Studentlitteratur (three textbooks). She has also provided consultation regarding Suicide assessment scales for Jansen Pharmaceuticals, provided local, regional, and national education for care professionals and social workers, and also educational activities through NGOs (Gothenburg University was remunerated for these services). For the remaining authors no conflicts of interest were declared.
The study was financed by grants from the Swedish state under the agreement between the Swedish government and the county councils, the ALF-agreement (ALF965812, ALF 716681, ALFGBG-942684, ALFGBG-715841, ALFGBG-965525, ALFGBG 147361, ALFGBG 433511), the Swedish Research Council (2012-5041, 2013-8717, 2013-02699, 2015-02830, 2016-01590, 2017-00639, 2019-01096), the Swedish Research Council for Health, Working Life and Welfare (Forte) (2012-1138, 2013-1202, AGECAP 2013-2300, 2013-2496, 2013-0475, 2016-07097, 2017-1604, 2018-00471), Konung Gustaf V:s och Drottning Victorias Frimurarestiftelse (Waern, Skoog), Hjärnfonden/Swedish Brain Foundation (FO2014-0207, FO2016-0214, FO2018-0214, FO2019-0163, FO2020-0235), Alzheimerfonden (AF-554461, AF-647651, AF-743701, AF-844671, AF-930868, AF940139, AF-968441), Eivind och Elsa K:son Sylvans stiftelse, the Alzheimer's Association (IIRG-09-131338), Stiftelsen Söderström-Königska Sjukhemmet, Stiftelsen för Gamla Tjänarinnor, Göteborg Center for person-centered care, Dahren's fund for schizophrenia research, and the Hjalmar Svensson Foundation.
The prospective population study of women in gothenburg, sweden, 1968-69 to 1992-93. A 24-year follow-up study with special reference to participation, representativeness, and mortality.
The Columbia-Suicide Severity Rating Scale: initial validity and internal consistency findings from three multisite studies with adolescents and adults.
Characterizing the phenomenology of passive suicidal ideation: a systematic review and meta-analysis of its prevalence, psychiatric comorbidity, correlates, and comparisons with active suicidal ideation.
The role of perceived and objective social connectedness on risk for suicidal thoughts and behavior in late-life and their moderating effect on cognitive deficits.
00543-7&id=gr1.jpg){kind=link}
00543-7&id=gr2.jpg){kind=link}