Journal of Affective Disorders 134 (2011) 333340

Contents lists available at ScienceDirect

Journal of Affective Disorders

j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / j a d

Research report

Prospective cohort study of suicide attempters aged 70 and above: One-year outcomes
Stefan Wiktorsson a,, Thomas Marlow a, Bo Runeson b, Ingmar Skoog a, Margda Waern a
a Neuropsychiatric Epidemiology Unit, Department of Psychiatry and Neurochemistry, Institute of Neuroscience and Physiology, Sahlgrenska Academy at the University of Gothenburg, Sweden b Division of Psychiatry, Department of Clinical Neuroscience, Karolinska Institute, Stockholm, Sweden

a r t i c l e

i n f o

a b s t r a c t
Background: Most elderly persons who attempt suicide suffer from depression. This study aimed to investigate one-year outcomes in suicide attempters aged 70+, and to identify predictors of these outcomes. Methods: 101 persons (mean age 80) who were hospitalized after a suicide attempt were interviewed at baseline and followed for one year by record linkage. Face-to-face interviews were carried out with 71% of those who were alive after one year (60 out of 85). Outcome measures included major/minor depression, MontgomeryAsberg Depression Rating Scale (MADRS) score, repeat non-fatal/fatal suicidal behavior and all-cause mortality. Results: One half (52%) of all those who were interviewed scored b 10 on the MADRS at follow-up. Among those with major depression at baseline, two thirds (26 out of 39) no longer fullled criteria for this disorder. Factors associated with non-remission of major depression (MADRS 10) included higher baseline depression and anxiety scores, higher suicide intent and lower Sense of Coherence. There were two suicides and six non-fatal repeat attempts. The relative risk of death (any cause) was 2.53 (95% CI = 1.454.10, p b 0.001). Limitations: This is a naturalistic study; participants received non-uniform treatment as usual. The proportion with repeat suicidal behavior was lower than anticipated and the study was thus underpowered with regard to this outcome. Conclusions: Half of the surviving attempters were free from depressive symptoms at one-year follow-up and there were relatively few repeat attempts. However, all cause mortality remained high in this elderly cohort. 2011 Elsevier B.V. All rights reserved.

Article history: Received 1 April 2011 Received in revised form 9 June 2011 Accepted 9 June 2011 Available online 6 July 2011 Keywords: Elderly Suicide attempt Depression Prospective study One-year mortality Antidepressant treatment

1. Introduction Populations are aging worldwide, and the past decade has seen the expansion of the literature on late life suicidal behavior. Psychological autopsy studies (Beautrais, 2002, Chiu et al., 2004, Conwell et al., 2002, Harwood et al., 2001, Waern et al., 2002) and controlled studies focusing on hospitalized suicide attempters (Liu and Chiu, 2009, Tsoh et al., 2005, Wiktorsson et al., 2010,

Corresponding author at: Neuropsychiatric Epidemiology Unit, Neuropsychiatry, SU/Mlndal, Wallinsgatan 6, SE 43141 Mlndal, Sweden. Tel.: +46 31 3438644; fax: +46 31 7760403. E-mail address: stefan.wiktorsson@neuro.gu.se (S. Wiktorsson). 0165-0327/$ see front matter 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.jad.2011.06.010

Yang et al., 2001) have generated risk estimates for a number of correlates of late life suicidal behavior. Estimates of population attributable risk associated with depression ranged from 67% (Tsoh et al., 2005) to 74% (Beautrais, 2002). The identication and treatment of depression are recognized as a central strategy for late life suicide prevention (Conwell et al., 2002). Considering the fact that high rates of suicide are observed worldwide for this age group (Hawton and van Heeringen, 2009), it is remarkable that the literature remains rather limited with regard to clinical outcomes for elderly people who seek hospital treatment in connection with attempted suicide. Several studies that were carried out during the 1990s (De Leo et al., 2002, Hepple and Quinton, 1997, Lebret et al., 2006)

334

S. Wiktorsson et al. / Journal of Affective Disorders 134 (2011) 333340

demonstrated high overall mortality and risk of repeat fatal/ non-fatal suicidal behavior. Only one (Hepple and Quinton, 1997) reported mental illness outcomes, noting that one third of those with a psychiatric diagnosis at index attempt were mentally healthy at one-year follow-up. Participants in that study were aged 65 and above and were referred to geriatric liaison services during 19891992. The past two decades have seen advances in the management of late life depression and antidepressants that are well-tolerated by the elderly are now widely prescribed (Carlsten and Waern, 2009). One could therefore anticipate better outcomes for more recent cohorts of elderly suicide attempters. We have previously reported results from a study of consecutive suicide attempt patients aged 70 years and above and a population comparison group (Wiktorsson et al., 2010). Over two thirds of the attempters fullled DSM-IV criteria for major depression and almost one quarter had minor depression. The current study aimed to investigate one-year outcomes in the suicide attempter cohort with respect to affective psychopathology, repeat suicidal behavior and overall mortality. A second objective was to identify predictors for these outcomes. 2. Methods 2.1. Cases Cases were recruited during 20032006 from emergency departments at ve hospitals in western Sweden. According to Statistics Sweden, the study area has a total population of 1.5 million of which 187,500 (12.5%) were aged over 70 years in 2005. A suicide attempt was dened as a situation in which a person has performed an actual or seemingly lifethreatening behavior with the intent of jeopardizing his life, or to give the appearance of such an intent, but which has not resulted in death(Beck et al., 1972). Of the 103 persons who took part in the baseline study (Wiktorsson et al., 2010), all but two consented to the release of medical records. Those

who died during the one-year observation period (n = 16) did not differ from those who survived regarding a number of sociodemographic and clinical characteristics at baseline (see items listed in the upper half of Table 2). Nor did these characteristics differ in those who did and did not take part in the follow-up interview. Mean SOC score was lower in those who died during the observation period compared to those who survived (Table 1). Table 1 shows further that those who chose not to take part in the one-year follow-up were physically healthier at baseline and had lower suicide intent scores than those who participated. In all, 71% of those who were alive (60 out of 85) took part in the follow-up interview (32 women and 28 men, mean age 81.1 years). Patient ow for the study in its entirety is shown in Fig. 1. 2.2. Interview Follow-up interviews were carried out by the same clinical psychologist (SW) who had performed the baseline interviews. The median time between the suicide attempt and the baseline interview was 11 days (Wiktorsson et al., 2010) and the median time between the suicide attempt and the follow-up interview was 391 days. Follow-up interviews were performed in participants' homes (n= 48), at nursing homes (n= 9), on psychiatric wards (n = 2) and at a psychiatric outpatient departments (n= 1). 2.3. Instruments Identical instruments were employed at baseline and oneyear follow-up. The MontgomeryAsberg Depression Rating Scale (MADRS) (Montgomery and Asberg, 1979) was used to rate depressive symptoms during the month preceding the index attempt and the month prior to the follow-up interview. This scale includes ten items scored 06, with 6 indicating the most severe level of symptoms, yielding a maximum score of 60. A slightly modied version of the Brief Scale for Anxiety (BSA) (Tyrer et al., 1984) was used to investigate anxiety

Table 1 Baseline characteristics of elderly suicide attempters (70+) (n = 101) by one-year follow-up status. Died before follow-up Deceased (n = 16) Alive at follow-up All living (n = 85) Participants (n = 60) Nonparticipants (n = 25) Mean 80.0 7.4 13.9 27.6 9.7 135.2 25.8 SD 5.4 3.3 5.2 12.0 7.3 17.4 2.5 Test resultb t = 0.443, 83 df, p = 0.659 t = 2.840, 83 df, p = 0.006 t = 2.286, 77 df, p = 0.025 t = 1.152, 81 df, p = 0.253 t = 1.276, 81 df, p = 0.205 t = 0.868, 72 df, p = 0.368 t = 0.000, 74 df, p = 1.000

Baseline characteristics Age The Cumulative Illness Rating Scale for Geriatrics (CIRS-G) The Suicide Intent Scalec (SIS) The MontgomeryAsberg Depression Rating Scaled (MADRS) The Brief Scale for Anxietye (BSA) Sense of Coherencef (SOC) Mini Mental State Examinationg (MMSE)
a b c d e f g

Mean 79.7 10.1 16.9 31.3 9.6 117.0 26.2

SD 4.9 3.7 2.3 11.1 5.8 29.4 3.1

Test resulta t = 0.092, 99 df, p = 0.927 t = 0.666, 99 df, p = 0.507 t = 0.771, 90 df, p = 0.443 t = 1.911, 96 df, p = 0.059 t = 0.659, 95 df, p = 0.511 t = 2.182, 85 df, p = 0.032 t = 0.336, 86 df, p = 0.738

Mean 79.6 9.3 15.8 25.4 9.0 131.4 25.8

SD 5.4 4.2 4.8 11.0 5.8 20.4 3.1

Mean 79.4 10.1 16.5 24.5 8.7 130.3 25.8

SD 5.5 4.4 4.5 10.5 5.1 21.3 3.3

T-test was used to compare continuous variables between those who died during the observation period and those who survived. T-test was used to compare baseline scores in participants and non-participants. Missing data for 3 deceased, 3 participants and 3 non-participants. Missing data for 1 deceased and 3 non-participants. Missing data for 2 deceased and 2 non-participants. Missing data for 3 deceased, 3 participants and 8 non-participants. Missing data for 4 deceased, 3 participants and 6 non-participants.

S. Wiktorsson et al. / Journal of Affective Disorders 134 (2011) 333340

335

Attempters fulfilling inclusion criteria (n = 140)

Discharged without study information (n = 7)

Invited to participate at baseline (n = 133)

Accepted but died before the interview (n = 2)

Declined participation (n = 28)

Participants at baseline study (n = 103)

Consent to record release (n = 101)

Died before follow-up (n = 16)

Untraceable (n = 2)

Eligible for follow-up (n = 83)

Declined participation (n = 23)

Participants at follow-up study (n = 60)

Fig. 1. One-year prospective study of hospitalized elderly suicide attempters (70+). Flow of participants.

symptoms during the month prior to the index attempt. The original BSA comprises 10 items rated 06, with 6 corresponding to the most severe level of symptoms. For the purpose of this study we used all items (inner feelings, hostile feelings, hypochondriasis, worrying over tries, reduced sleep, autonomic disturbances (reported and observed), aches and pains, and muscular tension) with the exception of the phobia item, yielding a maximum total score of 54. As sleep disturbances have been shown to be related to suicidal behavior (Agargun et al., 1997, Fawcett et al., 1990, Sjostrom et al., 2007) we

constructed a categorical sleep variable. A person who scored 3 on the MADRS item for reduced sleep was considered to have sleep problems. A single item (Do you think your situation is hopeless?) from the Geriatric Depression Scale (GDS) (Yesavage et al., 1982) was used to evaluate hopelessness. Perceived loneliness was examined with a single question (Do you feel lonely?). Suicide intent at the time of the index attempt was measured using the Suicide Intent Scale (SIS) (Beck et al., 1974). This scale comprises eight objective items and seven subjective items surrounding circumstances of the attempt.

336

S. Wiktorsson et al. / Journal of Affective Disorders 134 (2011) 333340

Items are scored from 0 (low intent) to 2 (high intent) yielding a maximum possible score of 30. Methods were denoted as non-violent (overdose, poisoning) or violent (hanging, cutting, drowning and other violent methods) (Conwell et al., 1990). The Cumulative Illness Rating Scale for Geriatrics (CIRS-G) (Miller et al., 1992) was used to rate physical illness/disability. A score ranging from 0 (no pathology) to 4 (extremely severe illness/impairment) was generated for each organ system. For the purpose of this study, a person was considered to have a serious physical illness/disability if scoring 3 or 4 in any of the 13 (non-psychiatric) organ categories. Mini Mental State Examination (MMSE) (Folstein et al., 1983) was administered to investigate cognitive capacity. The Sense of Coherence Scale (SOC) (Antonovsky, 1987) was used to examine to what extent participants found their lives meaningful, manageable and comprehensible. The Swedish version of the 29 item SOC scale, which has high validity and reliability (Langius et al., 1992), was administered. The answers were provided using a 7 point response scale with a maximum score of 203. A high score corresponds to strong SOC. 2.4. Diagnostics The Comprehensive Psychopathological Rating Scale (CPRS) (Asberg et al., 1978) was used to investigate psychopathology the month prior to the suicide attempt. An algorithm (see Appendix) based on CPRS and in accordance with the Diagnostic and Statistical Manual of Mental Disorder, fourth edition (DSM-IV) was used to diagnose major (Skoog et al., 1993) and minor depression (Wiktorsson et al., 2010) in accordance with DSM-IV research criteria. 2.5. Case records and Cause of Death Register Interview data and case records from primary care, psychiatric clinics, hospital emergency departments and geriatric departments were reviewed by the rst and last authors. Records were perused for evidence of previous psychiatric treatment and suicide attempts, physical illness/ disability, treatment with antidepressants, mood stabilizers, hypnotics, ECT, psychosocial treatments and evidence of new episodes of self-harm during the one-year observation period. Cases were linked to the Swedish Cause of Death Register using the individual's unique personal identier. Death certicates for deaths occurring during the one-year observation period were provided by the National Board of Health and Welfare. Data from Statistics Sweden were utilized to estimate the expected one-year mortality rate for an age- and sex matched general population sample. 2.6. Statistical analyses A person was considered to be in remission if they no longer fullled criteria for major depression and the MADRS score was b 10 (Licht-Strunk et al., 2009). MADRS data was missing for one participant at follow-up. A paired t-test was used to test differences in MADRS score at baseline and follow-up. Proportions were compared with Fisher's exact test (FET) and the t-test was used to test differences between groups regarding continuous variables. Multivariate logistic regressions were used to determine how baseline MADRS score, as a confounder,

inuenced the relationship between remission and SIS, BSA and SOC scores. A test based on the Poisson distribution was used to test the difference between observed and expected one-year mortality. All exploratory and formal statistical tests were carried out using SPSS for Windows (version 15, SPSS Inc, Chicago, IL, USA). All tests were two-tailed and p values b 0.05 were considered statistically signicant. 2.7. Ethical approval All participants received oral and written information about the study. They were assured of their right to withdraw from the study at any time, and written consent was obtained. The Research Ethics Committee at the University of Gothenburg approved the study. 3. Results 3.1. Depression outcomes Two thirds (26 out of 39) of those with major depression at index attempt no longer fullled criteria for this disorder at follow-up. Two persons with bipolar disorder took part in the follow-up interview and both had ongoing episodes of major depression observed at both index attempt and follow-up. Fourteen of those who participated in the follow-up had fullled DSM-IV research criteria for minor depression at index attempt. Neither major nor minor depression could be observed in 62% (n = 33) of these at follow-up. Slightly more than half of those interviewed after one year (31 out of 59) scored b 10 on the MADRS. The mean MADRS score for all participants was 11.0, range = 032, (SD = 9.1), which was signicantly lower than at index attempt, mean = 24.3, range = 348, (SD = 10.5), (t = 9.5, df = 58, p b 0.001). For the group with major depression at the index attempt, the mean MADRS score at follow-up was 12.6, range = 032 (SD = 9.7). This can be compared with a mean score of 30.2 at index attempt, range = 1948 (SD = 6.9), (t= 12.9, df= 38, p b 0.001). Table 2 focuses on persons with major depression at baseline. Full remission was observed at one year follow-up in 43.6%. Factors associated with non-remission included higher scores on MADRS, BSA, SIS and lower scores on the SOC scale at the time of the index attempt. The subjective subscale of the Suicide Intent Scale was related to non-remission (mean score= 11.5, vs. 9.3, t = 3.078, df = 35, p = 0.004), while the objective part was not (mean score= 6.3 vs. 4.8, t = 1.668, df= 36, p = 0.104). As the associations shown above might be confounded by depression, we created multivariate logistic regression models to determine whether associations remained after adjustment for MADR score. Subjective SIS subscore remained an independent predictor (Odds ratio; OR = 0.69, 95% CI = 0.480.99, p = 0.046). Neither anxiety symptom burden nor Sense of Coherence was independently associated with non-remission after adjusting for baseline MADRS score (results not shown). 3.2. Repeated suicidal behavior There were two suicides and both occurred within one month of the index attempt. One man took a lethal overdose while on short term leave from a geropsychiatric unit and the

S. Wiktorsson et al. / Journal of Affective Disorders 134 (2011) 333340 Table 2 Remission vs. non remission by baseline characteristics for elderly suicide attempters (70+) with major depression at baseline (n = 39). Remission, MADRS score 10 Yes (n = 17) Baseline characteristics Women Living alone Education, mandatory only History of psychiatric treatment Previous suicide attempt(s) ADb prescription at index attempt Reduced sleepc Serious physical illnessd Hopelessnesse Loneliness Violent attemptf n 9 13 12 11 7 11 6 10 8 9 5 Mean Age The Cumulative Illness Rating Scale for Geriatrics (CIRS-G) The Suicide Intent Scaleh (SIS) The Montgomery-Asberg Depression Rating Scale (MADRS) The Brief Scale for Anxiety (BSA) Sense of Coherencei (SOC) Mini Mental State Examinationj (MMSE)
a b c d e f g h i j

337

No (n = 22) (%) (53) (77) (71) (65) (41) (65) (35) (59) (47) (53) (29) SD 5.9 4.5 4.1 5.2 3.7 24.2 2.4 n 12 17 14 13 7 14 12 16 16 15 6 Mean 79.7 10.5 17.8 33.0 12.3 117.7 24.9 (%) (55) (77) (64) (59) (32) (64) (55) (73) (73) (68) (27) SD 5.1 4.0 4.0 6.9 4.6 16.8 3.5

Test resulta p 1.000 1.000 0.740 0.753 0.738 1.000 0.334 0.497 0.184 0.508 1.000 Test resultg t = 0.153, t = 0.373, t = 2.788, t = 3.119, t = 2.957, t = 2.324, t = 1.626, 37 37 35 37 37 35 36 df, df, df, df, df, df, df, p = 0.879 p = 0.711 p = 0.009 p = 0.004 p = 0.005 p = 0.026 p = 0.113

79.4 9.9 14.1 26.7 8.2 133.4 26.5

Fisher's exact test was used to compare proportions. Antidepressant. 3 on the MADRS item, reduced sleep. 3 in at least one of the CIRS-G somatic organ categories. One item from the Geriatric Depression Scale (GDS). Hanging, cutting, drowning and other violent methods. T-test was used to compare continuous variables. Missing data for 2 participants. Missing data for 2 participants. Missing data for 1 participant.

other took his life by hanging shortly after discharge. A further six persons made non-fatal attempts during the one-year observation period. Sociodemographic and clinical characteristics listed in Table 2 were also applied in an analysis comparing repeaters and non-repeaters. None of those with fatal/non-fatal repeats was married or cohabiting. This can be compared with 38% of the non-repeaters (p= 0.048, FET). No other signicant baseline differences could be shown between those who did and did not repeat (results not shown). SIS score could not be shown to predict repeat suicidal behavior (repeaters, mean = 14.8, vs. non-repeaters, mean = 16.1, t = 0.78, df= 90, p = 0.438). 3.3. One year mortality Fourteen natural deaths were recorded during the one-year observation period. Primary causes included: cardiac disease (n= 8), pulmonary disease (n= 2), cancer (n= 2), and other causes (n = 2). All cause one-year mortality was 15.8% (n= 16). This can be compared to the expected one-year mortality (6.3%) for a general population in Sweden in 2005 with this sex and age distribution, corresponding to a relative risk of 2.53 (95% CI= 1.454.10, p = 0.001). 3.4. Treatment Fifty-two of the 60 followed-up participants had received psychiatric in-patient care in connection with the index

attempt. The median length of stay was 25 days. Six persons were treated with electroconvulsive therapy in connection with the index attempt. At discharge 94% of those who had received a research diagnosis of major or minor depression at index attempt had a prescription for at least one antidepressant (one antidepressant, n = 31, two antidepressants, n = 20). Twenty-six persons (43%) of the total group received psychosocial counseling with a psychiatric nurse, social worker or psychologist during the year following the index attempt. Almost half (48%) were seen by a psychiatrist after discharge and all but one had contact with a primary care physician. At follow-up 88% (35 out of 40) of those with major depression and 93% (13 out of 14) of those with minor depression at baseline had a prescription for at least one antidepressant. All but three of those who had an ongoing antidepressant prescription at follow-up reported that they took the drug as prescribed. Hypnotics were prescribed for 68% (41 out of 60) of the total group at follow-up.

4. Discussion Slightly over one half of those who took part in the interview at one-year follow up had MADRS scores indicative of full remission. Record linkage for the entire study group (N = 101) showed only two suicides during the year that followed the index attempt. Before ndings are addressed in detail, some parts of the method require discussion.

338

S. Wiktorsson et al. / Journal of Affective Disorders 134 (2011) 333340

An important strength of this study is the use of both categorical and dimensional depression outcomes. Affective psychopathology was measured in face-to-face interviews at baseline and follow-up with the MADRS, which is a widely used, validated instrument that has been shown to retain its relevance also for elderly clinical populations (van der Laan et al., 2005). All interviews were carried out by the same psychologist, and we have previously reported high reliability between ratings of this psychologist and those of psychiatric research nurses and psychiatrists at our unit (Wiktorsson et al., 2010). Another important strength is the high age cut-off for inclusion. While most studies on late life suicidal behavior tend to include also persons in their 60s and in some cases even those in their 50s, the current study focuses specically on elderly persons, with a mean age of 80 years at index attempt. A major limitation is related to attrition, which is a common problem in prospective clinical studies that focus on persons who self-harm (Haw et al., 2001). However, a response rate of more than 70% is satisfactory in this age group. We carried out a detailed attrition analysis that showed that survivors had lower physical illness burden and lower suicide intent suggesting a healthy drop-out effect. While it is unclear how this may have affected results, it cannot explain the relatively favorable prognosis observed in this study. Another limitation is the fact that the study included only one follow-up measurement point. The duration of remission from depression cannot therefore be determined. Further, there were no objective measures of adherence meaning it is possible that some persons may not have taken their antidepressants as prescribed, even though they reported that they did. Another important limitation is the size of the cohort. The number of persons with repeat suicidal behavior within one year after index attempt was lower than anticipated based on the literature (De Leo et al., 2002), and the study was underpowered for analysis of this outcome. Finally, this is a naturalistic study and all persons received treatment as usual. An intervention based on a collaborative care model would be of great interest to test for this patient group. Most of the participants had major depression at the time of the index attempt, and such a model has been shown to reduce suicidal ideation and improve depression outcome in elderly with this diagnosis (Alexopoulos et al., 2009). Two thirds of those with major depression at index attempt no longer fullled criteria for this disorder at one-year followup, and half of those who had fullled DSM-IV research criteria for minor depression at baseline had neither major nor minor depression. While affective psychopathology was signicantly reduced, the mean MADRS score at follow-up (11.0) was still above that (6.2) observed in a general population sample with similar mean age (Vercelletto et al., 2002). Severity of depression at baseline predicted non-remission at follow-up, in line with results from a study on sequential treatment of late-life depression (Kok et al., 2009). High anxiety score at baseline was also associated with non-remission at follow-up. Pre-existing anxiety symptoms have been shown to be independently related to poorer antidepressant treatment response in late-life depression (Cohen et al., 2009) suggesting that these symptoms require special attention. The subjective part of the Suicide Intent Scale predicted non-remission

independent of baseline depressive symptoms. In contrast, physical illness burden ratings were identical in those with and without remission. This was somewhat unexpected, as concomitant physical illness has been shown to affect outcome in late life depression (Lyness et al., 2002, Tuma, 2000). As anticipated, Sense of Coherence at index attempt was low in comparison with a healthy elderly cohort (Wiesmann and Hannich, 2008). This is, to the best of our knowledge, the rst study to demonstrate an association between low Sense of Coherence and non-remission in suicidal seniors. However, our ndings suggest that this association may be mediated by depression. We have recently reported a strong association between major depression and SOC in this sample (Mellqvist et al., 2011), and SOC did not predict non-remission in the current study after the model was adjusted for MADRS score. Two percent died by suicide during the follow-up period. As Swedish late-life suicide rates are at an intermediate level in an international perspective, we anticipated a proportion more in line with that (13%) observed in the European multicenter study (De Leo et al., 2002). Six percent of the elderly attempters repeated non-lethal self-harm during the one-year observation period, a gure also lower than the 11.1% reported from the multicenter study (De Leo et al., 2002). Marital status predicted repeat fatal/non-fatal suicidal behavior in our study. Hepple and Quinton showed that divorced elderly suicide attempters were more likely to commit suicide, but they could show no difference in marital status in non-fatal re-attempters (Hepple and Quinton, 1997). Suicide intent score could not be shown to predict repeat attempt in our study. Scores were numerically higher in non-repeaters, as was also demonstrated in the European multicenter study (De Leo et al., 2002). Overall one-year mortality was more than double the population rate, suggesting that elderly who attempt suicide are physically more fragile than their peers. The proportion that died (16%) was however lower than that (24%) previously reported from the British study on 100 attempters aged 65 and above who were referred to geriatric liaison services during 19891992 (Hepple and Quinton, 1997). Our lower gure was somewhat unexpected considering the higher age of our participants. However, it should be pointed out that none of those in the current study died as a direct result of the attempt, which was the case for ve persons in the British study. Period changes in geriatric health care might have affected results, but ndings are difcult to interpret as previous studies do not show relative risks of mortality. Paralleling previous ndings from our study focusing on completed suicide in the same Region (Waern et al., 1996), the proportion who were already prescribed depression treatment at the time of the suicidal act was relatively high. It has been demonstrated that suicidal older people may require adjunctive medication (Szanto et al., 2007). For attempters in the current study, pharmacological treatments were intensied and augmented after the index attempt. At follow-up, 87% of all of the participants were on at least one antidepressant, and a quarter of those who had major or minor depression at baseline had more than one antidepressant. These drugs are readily available at low cost to all persons in Sweden, which might in part explain the nding that reported adherence was high. However, as non-adherence has been reported in over a fourth of older adults undergoing antidepressant treatment in a US study (Bosworth et al., 2008), the latter needs replication with a

S. Wiktorsson et al. / Journal of Affective Disorders 134 (2011) 333340

339

structured instrument or an objective monitoring procedure. Psychosocial treatments were initiated in over 40% of the cases after the index attempt in the current study. There are several studies, albeit involving somewhat younger cohorts, showing that specic psychotherapies (Van Schaik et al., 2006, 2007) are useful in the treatment of late-life depression. As this is a naturalistic study it was not possible to tease out the contributions of specic treatment elements. Findings from our study cannot be extrapolated to other settings. Differences in health care delivery systems, prescribing patterns and the availability of psychiatric services would be expected to affect one-year outcomes. Further, rates of attempted and completed suicide vary in different geographical settings. Attempted suicide rates for persons 70 years and older in the study area were 68/100,000 for men and 44/100,000 for women in 2005. The corresponding gures for completed suicide were 41/100,000 and 10/100,000, respectively. With the exception of the attempt rate in women, which was somewhat lower in the study area, our regional gures are similar to Swedish national rates for this age group. In conclusion, the one-year prognosis for elderly suicide attempters in the current study was relatively favorable. While we cannot conclude that advancements in the clinical management of the suicidal elderly may have contributed to this result, this is one possible explanation. Larger prospective multicenter studies will be required to elucidate this. High depression and anxiety scores, high suicide intent and low Sense of Coherence at baseline were all related to non-remission. Controlled intervention trials that also address existential issues for elders who self-harm are needed.
Role of funding source The study was supported by grants from the Swedish Research Council 11267, 20058460, 825-2007-7462, Swedish Council for Working Life and Social Research (no 20012835, 20012646, 20030234, 20040150, 2004 0145, 20060596, 20060020, 20081229, 20081111), the Alzheimer's Association Stephanie B. Overstreet Scholars (IIRG-00-2159), the Bank of Sweden Tercentenary Foundation, the Sderstrm-Knig Foundation, the Thuring Foundation, the Hjalmar Svensson Foundation, the Organon Foundation, the Axel Linder Foundation and the Wilhelm and Martina Lundgren Foundation. The funding sources had no role in the study design, the data collection, the analyses and interpretation of data, the writing of the report or in the decision of submit the paper for publication. Conict of interests Mr. Wiktorsson, Mr. Marlow, Dr. Runeson, Dr. Skoog, and Dr. Waern disclose no conict of interest. Acknowledgments The authors wish to thank all study participants and hospital staff. We would also like to thank Kristoffer Bckman, Erik Joas and Valter Sund for consultation regarding statistical issues.

References
Agargun, M.Y., Kara, H., Solmaz, M., 1997. Sleep disturbances and suicidal behavior in patients with major depression. J. Clin. Psychiatry 58, 249251. Alexopoulos, G.S., Reynolds III, C.F., Bruce, M.L., Katz, I.R., Raue, P.J., Mulsant, B.H., Oslin, D.W., Ten Have, T., 2009. Reducing suicidal ideation and depression in older primary care patients: 24-month outcomes of the PROSPECT study. Am. J. Psychiatry 166, 882890. Antonovsky, A., 1987. Unraveling The Mystery of Health How People Manage Stress and Stay Well. Jossey-Bass Publishers, San Francisco.

Asberg, M., Montgomery, S.A., Perris, C., Schalling, D., Sedvall, G., 1978. A comprehensive psychopathological rating scale. Acta Psychiatr. Scand. Suppl. 527. Beautrais, A.L., 2002. A case control study of suicide and attempted suicide in older adults. Suicide Life Threat. Behav. 32, 19. Beck, A.T., Davis, J.H., Frederick, C.J., Perlin, S., Pokorny, A.D., Schulman, R.E., Seiden, R.H., Wittlin, B.J., 1972. Classication and nomenclature. In: RESNICK, H.L.P.A.H., B.C. (Eds.), Suicide Prevention in the 70s. U.S Government Printing Ofce, Washington, D.C. The prediction of suicide. In: Beck, A.T., R. H. L. P., Lettieri, D.J. (Eds.), Development of suicidal scale., Bowie. Charles Press, Maryland. Bosworth, H.B., Voils, C.I., Potter, G.G., Steffens, D.C., 2008. The effects of antidepressant medication adherence as well as psychosocial and clinical factors on depression outcome among older adults. Int. J. Geriatr. Psychiatry 23, 129134. Carlsten, A., Waern, M., 2009. Are sedatives and hypnotics associated with increased suicide risk of suicide in the elderly? BMC Geriatr. 9, 20. Chiu, H.F., Yip, P.S., Chi, I., Chan, S., Tsoh, J., Kwan, C.W., Li, S.F., Conwell, Y., Caine, E., 2004. Elderly suicide in Hong Konga case-controlled psychological autopsy study. Acta Psychiatr. Scand. 109, 299305. Cohen, A., Gilman, S.E., Houck, P.R., Szanto, K., Reynolds III, C.F., 2009. Socioeconomic status and anxiety as predictors of antidepressant treatment response and suicidal ideation in older adults. Soc. Psychiatry Psychiatr. Epidemiol. 44, 272277. Conwell, Y., Rotenberg, M., Caine, E.D., 1990. Completed suicide at age 50 and over. J. Am. Geriatr. Soc. 38, 640644. Conwell, Y., Duberstein, P.R., Caine, E.D., 2002. Risk factors for suicide in later life. Biol. Psychiatry 52, 193204. De Leo, D., Padoani, W., Lonnqvist, J., Kerkhof, A.J., Bille-Brahe, U., Michel, K., Salander-Renberg, E., Schmidtke, A., Wasserman, D., Caon, F., Scocco, P., 2002. Repetition of suicidal behaviour in elderly Europeans: a prospective longitudinal study. J. Affect. Disord. 72, 291295. Fawcett, J., Scheftner, W.A., Fogg, L., Clark, D.C., Young, M.A., Hedeker, D., Gibbons, R., 1990. Time-related predictors of suicide in major affective disorder. Am. J. Psychiatry 147, 11891194. Folstein, M.F., Robins, L.N., Helzer, J.E., 1983. The Mini-mental state examination. Arch. Gen. Psychiatry 40, 812. Harwood, D., Hawton, K., Hope, T., Jacoby, R., 2001. Psychiatric disorder and personality factors associated with suicide in older people: a descriptive and case-control study. Int. J. Geriatr. Psychiatry 16, 155165. Haw, C., Hawton, K., Houston, K., Townsend, E., 2001. Psychiatric and personality disorders in deliberate self-harm patients. Br. J. Psychiatry 178, 4854. Hawton, K., Van Heeringen, K., 2009. Suicide. Lancet 373, 13721381. Hepple, J., Quinton, C., 1997. One hundred cases of attempted suicide in the elderly. Br. J. Psychiatry 171, 4246. Kok, R.M., Nolen, W.A., Heeren, T.J., 2009. Outcome of late-life depression after 3 years of sequential treatment. Acta Psychiatr. Scand. 119, 274281. Langius, A., Bjorvell, H., Antonovsky, A., 1992. The sense of coherence concept and its relation to personality traits in Swedish samples. Scand. J. Caring Sci. 6, 165171. Lebret, S., Perret-Vaille, E., Mulliez, A., Gerbaud, L., Jalenques, I., 2006. Elderly suicide attempters: characteristics and outcome. Int. J. Geriatr. Psychiatry 21, 10521059. Licht-Strunk, E., Van Marwijk, H.W., Hoekstra, T., Twisk, J.W., De Haan, M., Beekman, A.T., 2009. Outcome of depression in later life in primary care: longitudinal cohort study with three years' follow-up. BMJ 338, a3079. Liu, I.C., Chiu, C.H., 2009. Case-control study of suicide attempts in the elderly. Int. Psychogeriatr. 21, 896902. Lyness, J.M., Caine, E.D., King, D.A., Conwell, Y., Duberstein, P.R., Cox, C., 2002. Depressive disorders and symptoms in older primary care patients: oneyear outcomes. Am. J. Geriatr. Psychiatry 10, 275282. Mellqvist, M., Wiktorsson, S., Joas, E., Ostling, S., Skoog, I., Waern, M., 2011. Sense of coherence in elderly suicide attempters: the impact of social and health-related factors. Int. Psychogeriatr. 18. Miller, M.D., Paradis, C.F., Houck, P.R., Mazumdar, S., Stack, J.A., Rifai, A.H., Mulsant, B., Reynolds III, C.F., 1992. Rating chronic medical illness burden in geropsychiatric practice and research: application of the Cumulative Illness Rating Scale. Psychiatry Res. 41, 237248. Montgomery, S.A., Asberg, M., 1979. A new depression scale designed to be sensitive to change. Br. J. Psychiatry 134, 382389. Sjostrom, N., Waern, M., Hetta, J., 2007. Nightmares and sleep disturbances in relation to suicidality in suicide attempters. Sleep 30, 9195. Skoog, I., Nilsson, L., Palmertz, B., Andreasson, L.A., Svanborg, A., 1993. A population-based study of dementia in 85-year-olds. N. Engl. J. Med. 328, 153158. Szanto, K., Mulsant, B.H., Houck, P.R., Dew, M.A., Dombrovski, A., Pollock, B.G., Reynolds III, C.F., 2007. Emergence, persistence, and resolution of suicidal ideation during treatment of depression in old age. J. Affect. Disord. 98, 153161.

340

S. Wiktorsson et al. / Journal of Affective Disorders 134 (2011) 333340 Vercelletto, M., Martinez, F., Lanier, S., Magne, C., Jaulin, P., Bourin, M., 2002. Negative symptoms, depression and Alzheimer's disease. Int. J. Geriatr. Psychiatry 17, 383387. Waern, M., Beskow, J., Runeson, B., Skoog, I., 1996. High rate of antidepressant treatment in elderly people who commit suicide. BMJ 313, 1118. Waern, M., Runeson, B.S., Allebeck, P., Beskow, J., Rubenowitz, E., Skoog, I., Wilhelmsson, K., 2002. Mental disorder in elderly suicides: a casecontrol study. Am. J. Psychiatry 159, 450455. Wiesmann, U., Hannich, H.J., 2008. A salutogenic view on subjective wellbeing in active elderly persons. Aging Ment. Health 12, 5665. Wiktorsson, S., Runeson, B., Skoog, I., Ostling, S., Waern, M., 2010. Attempted suicide in the elderly: characteristics of suicide attempters 70 years and older and a general population comparison group. Am. J. Geriatr. Psychiatry 18, 5767. Yang, C.H., Tsai, S.J., Chang, J.W., Hwang, J.P., 2001. Characteristics of Chinese suicide attempters admitted to a geropsychiatric unit. Int. J. Geriatr. Psychiatry 16, 10331036. Yesavage, J.A., Brink, T.L., Rose, T.L., Lum, O., Huang, V., Adey, M., Leirer, V.O., 1982. Development and validation of a geriatric depression screening scale: a preliminary report. J. Psychiatr. Res. 17, 3749.

Tsoh, J., Chiu, H.F., Duberstein, P.R., Chan, S.S., Chi, I., Yip, P.S., Conwell, Y., 2005. Attempted suicide in elderly Chinese persons: a multi-group, controlled study. Am. J. Geriatr. Psychiatry 13, 562571. Tuma, T.A., 2000. Outcome of hospital-treated depression at 4.5 years. An elderly and a younger adult cohort compared. Br. J. Psychiatry 176, 224228. Tyrer, P., Owen, R.T., Cicchetti, D.V., 1984. The brief scale for anxiety: a subdivision of the comprehensive psychopathological rating scale. J. Neurol. Neurosurg. Psychiatry 47, 970975. Van Der Laan, N.C., Schimmel, A., Heeren, T.J., 2005. The applicability and the inter-rater reliability of the Comprehensive Psychopathological Rating Scale in an elderly clinical population. Int. J. Geriatr. Psychiatry 20, 3540. Van Schaik, A., Van Marwijk, H., Ader, H., Van Dyck, R., De Haan, M., Penninx, B., Van Der Kooij, K., Van Hout, H., Beekman, A., 2006. Interpersonal psychotherapy for elderly patients in primary care. Am. J. Geriatr. Psychiatry 14, 777786. Van Schaik, D.J., Van Marwijk, H.W., Beekman, A.T., De Haan, M., Van Dyck, R., 2007. Interpersonal psychotherapy (IPT) for late-life depression in general practice: uptake and satisfaction by patients, therapists and physicians. BMC Fam. Pract. 8, 52.