- Email: [email protected]
Self-harm in bipolar disorder: Findings from a prospective clinical database
Journal of Affective Disorders 173 (2015) 113–119
Contents lists available at ScienceDirect
Journal of Affective Disorders journal homepage: www.elsevier.com/locate/jad
Research report
Self-harm in bipolar disorder: Findings from a prospective clinical database Caroline Clements a,n, Steve Jones b, Richard Morriss c, Sarah Peters d, Jayne Cooper a, David While a, Navneet Kapur a a
Centre for Mental Health and Risk, Institute of Brain, Behaviour and Mental Health, The University of Manchester, UK Spectrum Centre for Mental Health Research, School of Health and Medicine, Lancaster University, UK c Department of Psychiatry and Community Mental Health, The University of Nottingham, UK d School of Psychological Sciences, The University of Manchester, UK b
art ic l e i nf o
a b s t r a c t
Article history: Received 29 May 2014 Received in revised form 6 October 2014 Accepted 7 October 2014 Available online 23 October 2014
Background: People with bipolar disorder may be at increased risk of suicidal behaviour but there are few prospective studies of self-harm in this group. Our aim was to describe the characteristics and outcome (in terms of repetition) for individuals with bipolar disorder who presented to hospital following self-harm. Method: A nested case-control study was carried out using a large prospective self-harm database (1997–2010) in Manchester, UK. Characteristics of bipolar cases and non-bipolar controls were compared using conditional logistic regression, and outcomes were assessed via survival analyses. Results: Bipolar cases (n¼ 103) were more likely to repeat self-harm than controls (n¼515): proportion with at least one repeat episode 58% vs. 25%, HR 3.08 (95% CI; 2.2–4.18). Previous self-harm, unemployment, contact with psychiatric services and sleep disturbance were all more common in cases than controls. Even after adjustment for known risk factors, the risk of repetition remained higher in the bipolar group (adjusted HR 1.68; 95% CI; 1.10–2.56). Limitations: The study covers cases from hospital sites in Manchester, UK, and therefore only includes selfharm that was serious enough to present at hospital emergency departments. Conclusion: People with bipolar disorder who self-harm have a higher risk of repetition than people who self-harm more generally. Adjusting for some known risk factors moderated, but did not abolish, this finding. Other factors, such as impulsivity, may also be important. & 2014 Elsevier B.V. All rights reserved.
Keywords: Self-harm Suicidal behaviour Bipolar disorder
1. Introduction People with bipolar disorder are known to be at increased risk of self-harm and suicide; up to 60% of people with bipolar disorder will self-harm at least once during their lifetime (Goodwin and Jamison, 2007; Baldessarini et al., 2006) with at least 5% of the bipolar population eventually dying by suicide (Nordentoft et al., 2011; Tondo et al., 2007). Studies have shown that previous self-harm is a particularly important risk factor for suicide in bipolar populations (Cassidy, 2011; Hawton et al., 2005; Rihmer, 2007), and approximately 60% of those who die by suicide have previously self-harmed (Oquendo et al., 2004; Clements et al., 2013). Given the high risk of n Correspondence to: Centre for Mental Health and Risk, Jean McFarlane Building, The University of Manchester, Oxford Road, Manchester, M20 2UT, UK. Tel.: þ 44 161 275 0735. E-mail address: [email protected] (C. Clements).
http://dx.doi.org/10.1016/j.jad.2014.10.012 0165-0327/& 2014 Elsevier B.V. All rights reserved.
self-harm in bipolar disorder and the association with more dangerous methods in these patients Simon et al., 2007), it is important that those most at risk can be identified early, and treated appropriately. Studies seeking to describe the characteristics of this high risk group could help to inform clinical practice as well as contributing to the research base on the prevention of suicidal behaviour in clinical populations. However, previous work has been subject to a number of methodological limitations; samples are often restricted to a single bipolar subtype, to small sample sizes, inpatient status, or geographical area (Hawton et al., 2005); retrospective patient self-report measures (Lopez et al., 2007); short follow-up times and lack of comparison groups (Goodwin and Jamison, 2007). These limitations have made it difficult to establish how bipolar patients differ from other people who have self-harmed (Hawton et al., 2005). The aims of this study were to describe the characteristics of people with bipolar disorder who self-harmed and to examine outcome in terms of repetition. We compared the
114
C. Clements et al. / Journal of Affective Disorders 173 (2015) 113–119
bipolar sample to a matched sample of non-bipolar controls using data from a large prospective database of hospital attendances for self-harm.
2. Method 2.1. The Manchester Self-Harm Project The analyses used data from the Manchester Self-Harm (MaSH) Project (further information is available at: 〈http://www.bbmh.man chester.ac.uk/cmhr/research/centreforsuicideprevention/MaSH〉). The project is a collaboration, between the University of Manchester, local mental health service providers, and the general hospitals that serve the population of the City of Manchester. Established monitoring methods are used to identify consecutive episodes of self-harm, presenting to the study hospitals. These include detailed searches of emergency department (ED) computer systems, using inclusive search terms (e.g. ‘neck injury’, ‘psychiatric’, ‘left department’), in order to maximise capture of possible episodes of self-harm. The nomenclature used in research on suicidal behaviour and self-harm varies between countries, research studies, and over time, and has been widely discussed within the literature (Silverman et al., 2007a, 2007b; Skegg, 2005). In this study we used the same definition of self-harm as the MaSH project (‘any intentional self-poisoning or self-injury, irrespective of motivation,’ Hawton et al., 2003). Motivations for acts of self-harm can be multiple, fluctuating, and difficult to identify even for the patient and reliably differentiating between those with or without suicidal intent can be problematic (Kapur et al., 2013a). Suicidal intent can change over time, even within the same self-harm episode and people who self-harm with apparently low intent are still at increased risk of repeat self-harm and suicide in the future (Kapur et al., 2013a). We therefore included all types of self-harm, regardless of whether the patient intended to die or not, or medical seriousness. This is also in line with the definition used by the National Institute for Health and Care Excellence (NICE) in clinical guidance on the management of self-harm in the UK (NICE, 2004, 2011). Sociodemographic data (age, sex, postcode and ethnicity) and details on method of self-harm are collected from ED notes, for all episodes, by research clerks. Detailed clinical information is then collected for patients who were assessed by mental health specialists and/or by ED doctors (including previous self-harm, drug or alcohol misuse, past psychiatric treatment, current mental state assessment, a risk assessment and clinical follow-up arrangements), as well as precipitating factors and circumstances of the act. Case ascertainment is essentially complete. Our data covered all self-harm presentations between September 1997 and December 2010. Levels of data completeness for individual variables are also high (Kapur et al., 2013b). Although the MaSH database records each presentation separately, each episode is linked to an identification number, which represents individual patients, and makes it possible to identify those who present repeatedly. 2.2. The nested case-control sample The MaSH database is suited to a nested case-control design, as it represents a well defined source population (of those in the Manchester area who attend hospital following self-harm), of known size, where cases and controls can be selected from a matched and relevant risk set. This method makes optimal use of the data available, for what is a relative rarely recorded event (e.g. self-harm by people with a diagnosis of bipolar disorder).
Bipolar cases were any individual with bipolar disorder recorded in the diagnosis field of the MaSH database. Diagnosis is generated by the assessing clinician, based on current mental state, and/or the patient's medical records. This information is then extracted from these paper assessments and electronic records and entered onto the MaSH database into broad diagnostic categories. The coding for a diagnosis of bipolar disorder on the MaSH database does not therefore distinguish between bipolar subtypes (e.g. bipolar I, bipolar II, rapid cycling etc.), and all diagnoses of bipolar disorder are included together in the analyses. Individuals with any conflicting diagnosis recorded in any other presentations were excluded from the bipolar group (i.e. three individuals were excluded at this point due to a conflicting diagnosis of schizophrenia). Cases and controls were matched only on date of self-harm. Inclusion of all demographic and clinical variables was considered important in order to identify all systematic differences between bipolar and non-bipolar self-harmers. Where there were less than five controls available for a particular date, controls were selected from the following day, and so on, until there were a total of five controls for each case. Matching by date also reduces the impact of any time critical factors that could influence results (e.g. changes in diagnostic criteria, treatment, or assessment protocols). It is common to use multiple controls for each case in nested methodologies, in order to optimise statistical power (Goldstein and Zhang, 2009). In this study we used five controls for each case and power calculations indicated this would give 99% power to detect a 30% difference at a 0.05 level of significance. 2.3. Ethics statement Systematic monitoring of self-harm in Manchester is conducted as part of the local NHS clinical audit programme. It is fully compliant with the UK Data Protection Act 1998. The MaSH Project's use of patient identifiable information without individual patient's consent was supported under the Section 251 of the NHS Act 2006 and approved by the National Information Governance Board for Health and Social Care. The local NHS research ethics committee ratified the project as an audit. 2.4. Analysis All statistical analyses were conducted using Stata/IC 12.1 (StataCorp LP, USA). 2.4.1. Characteristics of self-harm in bipolar disorder Simple descriptive statistics were used to describe the demographic and clinical characteristics of cases and controls. For proportions, the denominator was the number of valid responses for that item, ‘unknown’ responses were removed from each analysis. Variables were examined individually, in a series of univariate conditional logistic regressions, controlling for age and gender (an examination of the distribution of the age variable indicated it was not normally distributed; therefore, age groups were used instead of a continuous age variable). All variables that were significant (at a p value of r0.05) from the initial analyses were entered simultaneously into a multivariate conditional logistic regression, to investigate important discriminators between cases and control. 2.4.2. Repetition of self-harm The MaSH database allows repeat presentations to be linked to an individual patient, thereby providing a robust outcome measure of repeat self-harm, over a variable follow-up period. To make full use of these data we carried out survival analyses. First, the number of
C. Clements et al. / Journal of Affective Disorders 173 (2015) 113–119
repeat presentations in the case-control sample was summarised, and overall differences between groups were examined in a chi-square test of association. The data were then used to plot Kaplan–Meier survival curves, to describe and compare the cumulative survival characteristics of the two study populations. A log-rank test was used to compare the between-groups survival distributions. Preliminary tests indicated these data would be suitable for use in a cox regression model, to examine key explanatory variables, and their relationship to survival, or in this instance, occurrence of repeat self-harm. To that end, ‘diagnosis of bipolar disorder’ was entered into a model with other key variables known to be associated with self-harm (i.e. method of self-harm, previous self-harm, and receiving psychiatric care), and adjusted for age and gender.
3. Results The study included 618 individuals – 103 bipolar cases, and 515 non-bipolar controls. Cases with a recorded diagnosis of bipolar disorder made a total of 269 presentations over the study period, representing around 0.7% of all self-harm presentations. Nearly 60% (n ¼60) of bipolar cases had at least one repeat presentation for self-harm recorded. The average period of follow up after the initial self-harm episode was 7.7 years. 3.1. Case-control analyses 3.1.1. Characteristics of self-harm in bipolar disorder The results of the conditional logistic regression are shown in Table 1, along with the frequencies of demographic characteristics, clinical characteristics, and precipitants of self-harm, in cases and controls. In the univariate analyses the bipolar cases were more likely to be, female, aged between 45 and 64, married or living with a partner, unemployed or registered as long-term sick. Cases were more likely to have sleep disturbance prior to the self-harm act, nearly three times more likely to have had previous episodes of self-harm, and around seven times more likely to have current or previous psychiatric treatment. The act of self-harm was more than twice as likely to be considered a direct response to psychiatric symptoms in the bipolar group, than in the nonbipolar control group. Bipolar cases were less likely to be, aged under 24 years, to have current alcohol use, to have suicidal plans, and to have relationship problems as a precipitant of the self-harm presentation. Table 2 shows the results of the simultaneous entry multivariate conditional logistic regression analyses, incorporating variables that were significant in the univariate analyses. The multivariate model showed that bipolar cases were around three times more likely than non-bipolar controls to be, unemployed, to have received previous psychiatric treatment, and to have sleep disturbance. But, cases were less likely to have experienced relationship problems as a precipitant to the episode of self-harm, to have current alcohol use, or to have suicide plans (i.e. an acknowledgement by the patient, of ongoing intent to die, usually accompanied by some consideration of what method might be used). 3.2. Repetition of self-harm The overall repetition rate (that is, the proportion of the sample with at least one repeat episode at any time during follow up was 58% (n ¼60) for cases and 25% for controls (n ¼129). A chi-square test showed a significant association between diagnosis of bipolar disorder and repeated self-harm (χ² ¼44.58, p o0.001).
115
Fig. 1 shows time to first repeat for cases and controls as a Kaplan–Meier plot. Overall, a higher proportion of cases had a re-presentation for self-harm during the study period compared to controls (log rank χ² ¼50.32; po 0.001). However, there appeared to be very little difference between the groups in terms of time to first repeat. Average time to first repeat was 658 days for cases and 548 days for controls, the difference was not significant (t¼ 0.90; d.f. ¼ 187; p ¼0.37). 3.2.1. Cox regression Diagnosis of bipolar disorder was significant in the cox regression model, with a hazard ratio of 3.08 (95% CI; 2.2–4.18; p o0.01), adjusted for age and gender, indicating that individuals with bipolar disorder were over three times more likely to repeat self-harm than individuals without bipolar disorder. A post-hoc sensitivity analysis, restricting the control sample to those with a formal psychiatric diagnosis (n ¼95), showed the hazard ratio for repeat self-harm in those with bipolar disorder (compared to those with other formal diagnoses) was somewhat reduced but remained elevated (HR, 2.40; 95% CI; 1.56–3.70; po 0.01). Table 3 gives the results of the multivariate model. The inclusion of additional explanatory variables did reduce the value of the hazard ratio for diagnosis of bipolar disorder, indicating that these variables also had an influence on the repetition of suicidal behaviour. However, the presence of bipolar disorder still maintained significance at the 0.05 level in the adjusted analysis (HR, 1.68; 95% CI; 1.10–2.56; po 0.02), showing that repetition may well have been partly explained by the presence of this variable. Previous self-harm was the strongest predictor (HR, 15.28; 95% CI; 6.17–37.84; po 0.01) in the model, and current psychiatric treatment was also significant (HR, 1.80; 95% CI; 1.20–2.70; p o0.01). 3.2.2. Suicide deaths Mortality outcome data was available for 79% (n ¼491) of the case-control sample (mortality data were not available for the earlier years of the study because systematic data linkage was only carried out from 2001 onwards). In terms of fatal repetition, 4 (3.9%) of the bipolar cases died by suicide during the follow up period (receiving a suicide or undetermined verdict at Coroner's Inquest; Linsley et al., 2001) compared to 7 (1.4%) controls. Suicide mortality appeared higher in the bipolar group but, probably because of small numbers, this difference failed to reach statistical significance (χ²¼3.1, p ¼0.08). 4. Discussion 4.1. Main findings This study used a nested case-control method to describe the characteristics of people with bipolar disorder, who presented to EDs in Manchester, UK, following self-harm. Univariate analyses showed some differences in characteristics between cases and controls, most notably in relation to female gender, age between 45 and 64 years, unemployment, registered as long term sick, a history of previous self-harm, current or previous psychiatric care. In subsequent multivariate analysis, unemployment, previous psychiatric treatment, and sleep disturbance were more common in the bipolar cases, while current alcohol use, suicidal plans and relationship problems (as a precipitant of self-harm) were significantly more common among the non-bipolar control group. Repeat self-harm was significantly associated with a diagnosis of bipolar disorder, with a much larger proportion of cases repeating at least once during the 13 year study period, compared to controls (58% vs. 25%). Although there was no significant difference in time to first repeat between the two groups, survival
116
C. Clements et al. / Journal of Affective Disorders 173 (2015) 113–119
Table 1 Socio-demographic and clinical characteristics of bipolar cases and non-bipolar controls who had self-harmed. Odds ratios and p values calculated using univariate conditional logistic regression analysis. Variable
Cases N¼ 103 n (%)
Controlsa N ¼ 515 n (%)
Odds ratio (95% CI)
Female Age 24 and under 25 to 44 45 and 64 65 and over Ethnicity White Black Indian/Bangladeshi/Pakistani Other ethnicity Marital status Single Separated/divorce Married/partner Widowed Employment status Employed Unemployed Registered sick Retired Student Homemaker/carer Other employment Living arrangements Homeless Lives alone Lives with family Lives with friends Other Method of self-harm Self-poisoning, drugs Self-poisoning, other Self-injury Other History of self-harm Previous self-harm Self-harm in previous 12 months Psychiatric care Current psychiatric treatment Previous psychiatric treatment Clinical characteristics Current alcohol use Current substance misuse Feeling depressed Looking depressed Sleep disturbance Appetite disturbance Hopelessness Suicidal thoughts Suicide plans Hallucinations Precipitants of self-harm Relationship problems partner/friend Relationship problems parents/siblings Relationship problems others Bullying/intimidation Bereavement Housing Work Legal Health problems Miscarriage Money Direct response to psychiatric symptoms Other psychiatric problems Abuse Due to alcohol abuse Due to substance abuse Other problems
72(69.90)
283(54.95)
1.771(1.117–2.809)
21(20.39) 49(47.57) 29(28.16) 4(3.88)
192(37.35) 252(49.03) 59(11.48) 11(2.14)
0.416(0.245–0.708) 0.968(0.630–1.485) 2.979(1.756–5.052) 1.787(0.513–6.227)
90(88.24) 3(2.94) 8(7.84) 1(0.98)
400(88.30) 11(2.43) 33(7.28) 5(1.10)
0.878(0.434–1.777) 1.015(0.245–4.203) 1.438(0.624–3.316) 1.177(0.127–10.906)
0.717 0.984 0.394 0.886
42(44.21) 13(13.68) 38(40.00) 2(2.11)
271(57.78) 72(15.35) 116(24.73) 10(2.13)
0.856(0.508–1.444) 0.546(0.266–1.121) 1.746(1.059–2.878) 0.407(0.069–2.408)
0.560 0.099 0.029n 0.322
16(16.67) 45(46.88) 20(20.83) 5(5.21) 6(6.25) 3(3.13) 1(1.04)
115(28.19) 141(34.56) 46(11.27) 16(3.92) 54(13.24) 25(6.13) 11(2.70)
0.555(0.304–1.010) 1.890(1.154–3.095) 2.028(1.047–3.930) 0.229(0.025–2.102) 0.614(0.230–1.642) 0.395(0.113–1.377) 0.230(0.028–1.872)
0.054 0.011n 0.036n 0.193 0.331 0.145 0.169
1(1.08) 18(19.35) 48(51.61) 9(20.45) 9(9.68)
8(2.04) 93(23.72) 193(49.23) 35(79.55) 38(9.69)
0.609(0.070–5.314) 0.665(0.361–1.224) 1.230(0.760–1.990) 1.303(0.570–2.980) 1.073(0.483–2.385)
0.653 0.190 0.400 0.531 0.863
84(81.55) 2(1.94) 14(13.59) 3(2.91)
414(80.54) 4(0.78) 75(14.59) 21(4.09)
0.889(0.506–1.561) 2.210(0.330–14.800) 1.061(0.568–1.981) 1.025(0.284–3.694)
0.682 0.414 0.853 0.970
71(73.96) 33(39.29)
187(48.95) 112(36.96)
2.983(1.726–5.153) 1.043(0.567–1.918)
o 0.001n 0.892
75(79.79) 78(87.64)
135(34.88) 172(44.44)
7.111(3.854–13.119) 6.889(3.480–13.637)
o 0.001n o 0.001n
14(16.28) 7(8.14) 58(64.44) 46(52.27) 61(67.78) 42(46.67) 32(37.21) 33(37.08) 7(7.95) 15(16.85)
111(29.52) 49(13.21) 251(65.54) 177(46.34) 203(53.14) 160(41.67) 155(40.90) 144(37.50) 70(18.47) 34(8.88)
0.457(0.237–0.881) 0.790(0.316–1.977) 1.029(0.735–1.440) 1.007(0.611–1.659) 1.737(1.022–2.951) 1.154(0.706–1.886) 0.637(0.377–1.076) 0.859(0.511–1.444) 0.362(0.146–0.898) 1.865(0.928–3.745)
0.019n 0.615 0.869 0.977 0.041n 0.567 0.092 0.567 0.028n 0.080
23(26.14) 15(17.05) 13(14.77) 2(2.27) 8(9.09) 12(13.64) 8(9.09) 2(2.27) 5(5.68) 1(1.14) 10(11.36) 31(35.23) 2(6.67) 5(5.68) 1(3.33) 1(3.33) 8(9.09)
161(42.82) 71(18.88) 38(10.11) 20(5.32) 34(9.04) 36(9.57) 42(11.17) 13(3.46) 34(9.04) 4(1.07) 44(11.70) 53(14.10) 8(7.84) 22(5.85) 20(19.61) 6(5.88) 49(13.03)
0.477(0.268–0.849) 0.905(0.461–1.777) 1.577(0.738–3.369) 0.463(0.102–2.106) 0.808(0.322–2.027) 1.860(0.853–4.060) 1.028(0.450–2.350) 0.807(0.168–3.868) 0.707(0.248–2.019) 1.175(0.112–12.364) 1.397(0.630–3.096) 2.319(1.322–4.068) 3.156(0.436–22.832) 1.108(0.336–3.656) 0.176(0.021–1.494) 1.472(0.098–22.087) 0.612(0.259–1.447)
0.012n 0.772 0.240 0.319 0.649 0.119 0.948 0.788 0.518 0.893 0.410 0.003n 0.255 0.867 0.111 0.779 0.264
n
p Value 0.015n 0.001n 0.880 o 0.001n 0.362
Significant at pr 0.05 Diagnostic breakdown of controls: n¼ 420 (82%) no formal psychiatric diagnosis recorded; n¼ 40 (8%) depressive disorders; n¼ 15 (3%) adjustment/stress disorders; n¼ 15 (3%) alcohol misuse; n ¼25 (5%) other psychiatric disorders. a
C. Clements et al. / Journal of Affective Disorders 173 (2015) 113–119
Table 2 A simultaneous entry multivariate conditional logistic regression comparing the presence of key characteristics in people with bipolar disorder who self-harm, and people without bipolar disorder who self-harm. Variable
OR(95% CI)
p
Female 25 to 44 45 and 64 65 and over Married/partner Unemployed Long term sick Relationship problems partner/friend Direct response to psychiatric symptoms Previous self-harm Current psychiatric treatment Previous psychiatric treatment Current alcohol use Sleep disturbance Suicide plans
1.830(0.834–4.018) 1.026(0.381–2.767) 0.931(0.280–3.099) 1.192(0.111–12.814) 2.060(0.878–4.833) 3.081(1.210–7.849) 2.180(0.682–6.960) 0.305(0.116–0.802) 0.639(0.238–1.721) 2.253(0.927–5.476) 2.176(0.769–6.157) 2.955(1.049–8.327) 0.310(0.100–0.957) 2.945(1.287–6.738) 0.245(0.070–0.859)
0.132 0.959 0.907 0.885 0.097 0.018n 0.188 0.016n 0.376 0.073 0.143 0.040n 0.042n 0.011n 0.028n
n
Significant at pr 0.05.
Kaplan-Meier survival estimates
100%
75%
50%
25%
0% 0
1000
2000
3000
4000
5000
Days to first repeat Other cases
Bipolar cases
Fig. 1. Kaplan–Meier survival curves, showing time to first repeat presentation for self-harm, for cases and controls.
Table 3 Results of a simultaneous entry multivariate cox regression, examining the association between bipolar disorder and repeat self-harm. Key risk factors, known to increase risk of self-harm, have been included in the model as co-variates. Variable
Hazard ratio(95% CI)
p
Bipolar disorder 25 to 44 45 and 64 65 and over Female Poisoning (vs. injury) Previous self-harm Current psychiatric care
1.675(1.098–2.556) 0.886(0.592–1.326) 0.645(0.364–1.143) 0.990(0.357–2.744) 1.154(0.743–1.790) 0.503(0.292–0.867) 15.284(6.174–37.838) 1.797(1.195–2.703)
0.017n 0.557 0.133 0.985 0.524 0.013n o0.001n 0.005n
n
Significant at pr 0.05
analysis showed that a diagnosis of bipolar disorder did have an independent effect on the risk of repetition, even when other known predictors (e.g. previous self-harm) were included in the model.
117
have shown a greater gender equality in self-harm (Cassidy, 2011). Our results showed that females were over represented among bipolar cases, even in comparison to ‘other’ self-harmers – a group that is already known to be disproportionately female (Hawton et al., 2003). One possible explanation is that, although bipolar disorder itself is unrelated to gender, women are more likely to seek help, and therefore, more likely to be in contact with services and given an appropriate diagnosis. Another explanation is that, while self-harm methods used by people with bipolar disorder tend to be more potentially lethal in general (Cassidy, 2011), there is evidence that methods may be particularly dangerous in men with bipolar disorder, with a higher case fatality (D’Ambrosio et al., 2012). Although not significant in the multivariate analysis (due to the presence of ‘previous psychiatric treatment' in the model which may be acting as a confounder), previous self-harm was significant in the univariate analysis, and the difference in proportions between cases and controls was large, at 74% vs. 49%. This is in line with a number of previous studies on both lethal and non-lethal self-harm in bipolar disorder (Hawton et al., 2005; Rihmer, 2007). It might also be of note that, ‘direct response to psychiatric symptoms’ was the only significant predictor within the precipitants of self-harm variables, and it was more than twice as common in the bipolar cases. If this is considered along with the tendency for the bipolar cases to be in current or previous contact with psychiatric services, and to have previously self-harmed, this could be indicative of a group who are more seriously unwell, and who might have had a longer duration of illness. There is some evidence that those with longer durations of untreated illness/ incorrect diagnosis, greater illness severity, and illness burden, have a higher frequency of suicidal behaviour and poorer outcomes (Baldessarini et al., 2006; Tondo et al., 2007). By contrast those without bipolar disorder were more likely to report relationship problems as a precipitant. While depression and hopelessness were present, to some extent, across all cases and controls, as might be expected (Hawton et al., 2005), these were not significant predictors of repetition. Of the clinical characteristics described in this dataset, only sleep disturbance seemed to be a robust independent predictor in bipolar disorder (Rihmer, 2007), and indeed, was almost three times more likely to be present in bipolar cases than non-bipolar controls. Previous work looking at the association of drug and/or alcohol use with suicidal behaviour in bipolar disorder has produced contradictory results (Cassidy, 2011). We found that alcohol use was, in fact, recorded less frequently in cases than controls. This is somewhat surprising given that alcohol and substance use is recognised as common in people with bipolar disorder (Goodwin and Jamison, 2007). However, there is evidence that alcohol/ substance use may not be not strongly related to suicidal behaviour in bipolar disorder (Hawton et al., 2005), and it may be that illness related factors play a more important role in this group. For example, suicidal plans were much less common in bipolar cases, which supports the hypothesis that impulsivity, a characteristic trait in bipolar patients, may be of particular importance (Tondo and Baldessarini, 2005). Impulsivity, driven by positive or negative affect, in the presence of other factors recognised associated with self-harm and suicide more generally, may be especially important in precipitating episodes of self-harm, and may be a key area for future research. 4.3. Methodological limitations
4.2. Findings in relation to previous research Our study supports work that has shown that women with bipolar disorder have a higher rate of suicidal behaviour than men (Tondo et al., 2003), and contrasts with the findings of studies that
People with bipolar disorder are at high risk of self-harm (Goodwin and Jamison, 2007; Baldessarini et al., 2006), and it is possible that some people in the control group could have undiagnosed bipolar disorder. The number of cases with a
118
C. Clements et al. / Journal of Affective Disorders 173 (2015) 113–119
diagnosis of bipolar disorder was small in comparison to the size of the overall dataset (less than 1% of total presentations). This suggests that psychiatric diagnoses in general may be underreported on the MaSH database. This is in line with previous reviews, which suggest that only a small proportion of self-harm patients receive a psychiatric diagnosis when they present to hospitals (Effective Health Care Bulletin, 1998). This may be due to time pressures within the emergency department, the need for rapid assessment and turnover of self-harm patients, or reluctance to assign a diagnosis on the basis of one assessment. Diagnosis is assigned by the clinician at assessment, and reflects mental state at the time, as well as historical information taken from the patient and medical records. Where a specific diagnosis is unclear the field will remain empty, and any diagnosis given is, therefore, more likely to be based on clear evidence, which may favour those with more pronounced symptoms and more severe illness. Misdiagnosis could also be a problem within the bipolar disorder cases, where there is known to be difficulty distinguishing bipolar disorder from borderline personality disorder—which is also associated with high risk of self-harm. However, this is true of most studies of bipolar disorder that rely on past records and clinical judgement, and verifying diagnosis is beyond the scope of the current study. The effect of bias from such missed or misallocated bipolar disorder cases/controls would be to minimise between-group differences described in this study, and the fact that the differences seen in our findings are consistent with previous work is reassuring. The control sample contained data on 95 (18%) individuals who had a psychiatric diagnosis, including: depressive illnesses (n ¼43, 8%), anxiety and stress disorders (n ¼20, 4%), alcohol misuse/ dependence (n ¼18, 3%), and personality disorder (n ¼ 3, 0.6%). It was not our intention to compare self-harm between diagnostic groups in this study. Instead, we aimed to investigate factors which distinguish self-harm in bipolar disorder from self-harm in general, which will include a proportion of people with psychiatric diagnoses. We believe that our study has potentially important implications for clinicians when they encounter a person with bipolar disorder who has self-harmed. As the majority of controls (over 80%) had no recorded psychiatric diagnosis it could be that some of the findings were not specific to those with bipolar disorder but reflect wider risks for individuals with mental illness. However, a post-hoc sensitivity analysis showed that the hazard ratio for repeat self-harm in those with bipolar disorder compared to controls with a psychiatric diagnosis remained elevated with a hazard ratio of 2.40 (compared to a hazard ratio of 3.08 for the whole sample). Another factor that might have reduced between-group differences was the presence of co-morbidities among bipolar disorder cases. However, only 11 (11%) of cases had a co-morbid psychiatric diagnosis recorded on the MaSH database; depressive illnesses were most common (n¼ 5, 5%), followed by anxiety and stress disorders (n ¼3, 3%), alcohol misuse (n ¼2, 2%), and learning difficulties (n ¼1, 1%). Nineteen (20% of the controls with a psychiatric diagnosis, 4% of all controls) were recorded as having co-morbid psychiatric diagnoses. The dataset used is geographically restricted to one city in England, with particular economic and socio-demographic characteristics, and it is possible that it may not be valid to extrapolate findings to other areas. Data also only cover self-harm acts that are serious enough to present at hospital and these may be a systematically different group to less medically-serious forms of selfharm. However, this would be true across both cases and controls and does not reduce the importance of the findings for those who do present to hospital following self-harm. There are also a number of psychological characteristics that are of current
research interest in relation to self-harm and suicide in bipolar disorder, such as temperament (Pompili et al., 2012; Sarısoy et al., 2012), and impulsivity (Mahon et al., 2012; Watkins & Meyer, 2013), which we acknowledge are important, but did not have sufficient detail within the data to explore in this study. 4.4. Clinical implications In this study we found that repetition of self-harm was particularly common in individuals with bipolar disorder – nearly six out of ten repeated during the follow up period. We found that people with bipolar disorder who self-harmed, typically had a number of high risk characteristics (e.g. previous self-harm, history of psychiatric care). However, these characteristics, recognised as risk factors in general and psychiatric self-harm populations, did not comprehensively explain the increased risk associated with a diagnosis of bipolar disorder. Previous work suggests that impulsivity is associated with self-harm in bipolar disorder (Tondo and Baldessarini, 2005; Swann et al., 2005), and this may explain why bipolar disorder cases in this study were less likely to have ‘suicidal plans’, even though the proportion of those with suicidal thoughts was similar between groups. Sleep disturbance was common and an independent predictor of self-harm in the bipolar group. This may be a marker of changes in illness, agitation or mood-state, as found in mixed depressive episodes and known to be associated with self-harm in bipolar (Valtonen et al., 2008), and depressive symptoms were evident in nearly 65% of the bipolar cases. We were not able to assess psychiatric comorbidity, severity, or duration of illness, in this study. However, given the frequency of previous self-harm and current/previous psychiatric care in the bipolar cases, it is likely that those with bipolar disorder who self-harm, have a more chronic and complicated course of illness than those with bipolar disorder who do not self-harm. The key precipitant that distinguished the bipolar group was ‘a direct response to psychiatric symptoms’ which indicates that illness-related factors, may play a larger role in acts of selfharm in bipolar disorder, than any specific negative life events. Clinicians need to be aware of the higher risk of self-harm in bipolar disorder, and the link between previous self-harm and suicide (Clements et al., 2013). Of course these risks might not be restricted to those with bipolar disorder and those with other psychiatric disorders may also be at higher risk. The risk is not restricted to the short term – the average time to repetition approached 2 years in this study. Those who repeat are also at higher risk for completed suicide (Zahl and Hawton, 2004). Sixteen per cent (n¼ 13) of bipolar cases who were in previous contact with psychiatric services, were recorded as not in current contact with services. As risk appears to be enduring, continued contact with services may be appropriate. The impulsivity of self-harm in bipolar disorder may make it more difficult to predict when patients are most at risk. However, good quality monitoring and management of symptoms could help. Patients, who report sleep disturbance and/or changes in mood, might be considered to be at increased risk of engaging in suicidal behaviours, and treated accordingly. The use of lithium, which has been shown to have specific benefits for prevention of suicide (Tondo et al., 2003; Cipriani et al., 2013), and other mood stabilisers, may be beneficial. Psychological approaches, such as cognitive behavioural therapies, which have been shown to have a positive impact on suicidal behaviour in general, might also have an important role in prevention (Tarrier et al., 2008). Studies that could draw out differences in risk of self-harm between subtypes of bipolar disorder would be valuable, along with more detailed work to draw out and describe the influence of risk factors in the period that precedes acts of self-harm. Qualitative studies would allow for closer examination of antecedents and
C. Clements et al. / Journal of Affective Disorders 173 (2015) 113–119
the experience of them at an individual level could help to identify when and how services might be able to intervene appropriately to reduce self-harm in this potentially high risk population. Role of funding source This article presents independent research funded by the National Institute for Health Research (NIHR) under its Programme Grants for Applied Research programme (Reference number RP-PG-0407-10389). The views expressed are those of the author(s) and not necessarily those of the NHS, the NIHR or the Department of Health.
Conflict of interest There are no conflicts of interest to declare.
Acknowledgements We would like to thank Sarah Steeg, Iain Donaldson and the rest of the MaSH team for their help in organising and accessing data for this study.
References Baldessarini, R.J., Pompili, M., Tondo, L., 2006. Suicide in bipolar disorder: risks and management. CNS Spectr. 11, 465–471. Cassidy, F., 2011. Risk factors of attempted suicide in bipolar disorder. Suicide Life Threat Behav. 41, 6–11. Cipriani, A., Hawton, K., Stockton, S., Geddes, J., 2013. Lithium in the prevention of suicide in mood disorders: updated systematic review and meta-analysis. BMJ 346, f3646. Clements, C., Morriss, R., Jones, S., Peters, S., Roberts, C., Kapur, N., 2013. Suicide in bipolar disorder in a national English sample, 1996–2009: frequency, trends, and characteristics. Psychol. Med. 43, 2593–2602. D'Ambrosio, V., Salvi, V., Bogetto, F., Maina, G., 2012. Serum lipids, metabolic syndrome and lifetime suicide attempts in patients with bipolar disorder. Prog. Neuropsychopharmacol. Biol. Psychiatry 37, 136–140. Goldstein, L., Zhang, H., 2009. Efficiency of the maximum likelihood estimator for nested case control sampling. Bernoulli 15, 569–597. Goodwin, F.K., Jamison, K.R., 2007. Manic-Depressive Illness: Bipolar Disorders and Recurrent Depression. Oxford University Press, New York, NY. Hawton, K., Harriss, L., Hall, S., Simkin, S., Bale, E., Bond, A., 2003. Deliberate selfharm in Oxford, 1990–2000: a time of change in patient characteristics. Psychol. Med. 33, 987–996. Hawton, K., Sutton, L., Haw, C., Sinclair, J., Harris, L., 2005. Suicide and attempted suicide in bipolar disorder: a systematic review of risk factors. J. Clin. Psychiatry 66, 693–704. Kapur, N., Cooper, J., O'Connor, R., Hawton, K., 2013a. Non-suicidal self-injury v. attempted suicide: new diagnosis or false dichotomy? Br. J. Psychiatry 202, 326–328. Kapur, N., Steeg, S., Webb, R., Haigh, M., Bergen, H., Hawton, K., et al., 2013b. Does clinical management improve outcomes following self-harm? Results from the Multicentre Study of Self-Harm in England. PLoS ONE 8, e70434. http://dx.doi. org/10.1371/journal.pone.0070434. Linsley, K.R., Schapira, K., Kelly, T.P., 2001. Open verdict vs. suicide—importance to research. Br. J. Psychiatry 178, 465–468.
119
Lopez, V.A., Detera-Wadleigh, S., Cardona, I., Kassem, L., McMahon, F.J., 2007. Nested association between genetic variation in tryptophan hydroxylase II, bipolar affective disorder, and suicide attempt. Biol. Psychiatry 61, 181–186. Mahon, K., Burdick, K.E., Wu, J., Ardekani, B.A., Szeszko, P.R., 2012. Relationship between suicidality and impulsivity in bipolar I disorder: a diffusion tensor imaging study. Bipolar Disord. 14, 80–89. NICE, 2004. Self-harm: the short-term physical and psychological management and secondary prevention of self-harm in primary and secondary care, NICE Clinical Guideline 16. Available at: 〈www.nice.org.uk/CG16〉. NICE, 2011. Self-harm: longer-term management, NICE Clinical Guideline 133. Available at 〈www.nice.org.uk/CG133〉. Nordentoft, M., Mortensen, P., Pedersen, C., 2011. Absolute risk of suicide after first hospital contact in mental disorder. Arch. Gen. Psychiatry 68, 1058–1064. Oquendo, M.A., Galfalvy, H., Russo, S., Ellis, S.P., Grunebaum, M.F., Burke, A., et al., 2004. Prospective study of clinical predictors of suicidal acts after a major depressive episode in patients with major depressive disorder or bipolar disorder. Am. J. Psychiatry 161, 1433–1441. Pompili, M., Innamorati, M., Rihmer, Z., Gonda, X., Serafini, G., Akiskal, H., et al., 2012. Cyclothymic-depressive-anxious temperament pattern is related to suicide risk in 346 patients with major mood disorders. J. Affect. Disord. 136, 405–411. Rihmer, Z., 2007. Suicide risk in mood disorders. Curr. Opin. Psychiatry 20, 17–22. Sarısoy, G., Kaçar, O.F., Pazvantoğlu, O., Oztürk, A., Korkmaz, I.Z., Kocamanoğlu, B., et al., 2012. Temperament and character traits in patients with bipolar disorder and associations with attempted suicide. Compr. Psychiatry 53, 1096–1102. Silverman, M.M., Berman, A.L., Sanddal, N.D., O'Carroll, P.W., Joiner, T.E., 2007a. Rebuilding the tower of Babel: a revised nomenclature for the study of suicide and suicidal behaviors. Part 1: Background, rationale, and methodology. Suicide Life Threat Behav. 37, 248–263. Silverman, M.M., Berman, A.L., Sanddal, N.D., O'carroll, P.W., Joiner, T.E., 2007b. Rebuilding the tower of Babel: a revised nomenclature for the study of suicide and suicidal behaviours. Part 2: Suicide-related ideations, communications, and behaviors. Suicide Life Threat Behav. 37, 264–277. Simon, G.E., Hunkler, E., Fireman, B., Lee, J.Y., Savarino, J., 2007. Risk of suicide attempt and suicide death in patients treated for bipolar disorder. Bipolar Disord. 9, 526–530. Skegg, K., 2005. Self-harm. Lancet 366, 1471–1483. Swann, A.C., Dougherty, D.M., Pazzaglia, P.J., Pham, M., Steinberg, J.L., Moeller, F.C., 2005. Increased impulsivity associated with severity of suicide attempt history in patients with bipolar disorder. Am. J. Psychiatry 162, 1680–1687. Tarrier, N., Taylor, K., Gooding, P., 2008. Cognitive-behavioural interventions to reduce suicide behaviour. Behav. Modif. 32, 77–108. Tondo, L., Baldessarini, R., 2005. Suicidal risk in bipolar disorder. Clin. Neuropsychiatry 2, 55–65. Tondo, L., Isacsson, G., Baldessarini, R., 2003. Suicidal behaviour in bipolar disorder. CNS Drugs 17, 491–511. Tondo, L., Lepri, B., Baldessarini, R.J., 2007. Suicidal risk among 2826 Sardinian major affective disorder patients. Acta Psychiatr. Scand. 116, 419–428. Valtonen, H.M., Suominen, K., Haukka, J., Mantere, O., Leppamaki, S., Arvilommi, P., et al., 2008. Differences in incidence of suicide attempts during phases of bipolar I and II disorders. Bipolar Disord. 10, 588–596. Watkins, H.B., Meyer, T.D., 2013. Is there an empirical link between impulsivity and suicidality in bipolar disorders? A review of the current literature and the potential psychological implications of the relationship. Bipolar Disord. 15, 542–558. Zahl, D.L., Hawton, K., 2004. Repetition of deliberate self-harm and subsequent suicide risk: long-term follow-up study of 11,583 patients. Br. J. Psychiatry 185, 70–75.
Contents lists available at ScienceDirect
Journal of Affective Disorders journal homepage: www.elsevier.com/locate/jad
Research report
Self-harm in bipolar disorder: Findings from a prospective clinical database Caroline Clements a,n, Steve Jones b, Richard Morriss c, Sarah Peters d, Jayne Cooper a, David While a, Navneet Kapur a a
Centre for Mental Health and Risk, Institute of Brain, Behaviour and Mental Health, The University of Manchester, UK Spectrum Centre for Mental Health Research, School of Health and Medicine, Lancaster University, UK c Department of Psychiatry and Community Mental Health, The University of Nottingham, UK d School of Psychological Sciences, The University of Manchester, UK b
art ic l e i nf o
a b s t r a c t
Article history: Received 29 May 2014 Received in revised form 6 October 2014 Accepted 7 October 2014 Available online 23 October 2014
Background: People with bipolar disorder may be at increased risk of suicidal behaviour but there are few prospective studies of self-harm in this group. Our aim was to describe the characteristics and outcome (in terms of repetition) for individuals with bipolar disorder who presented to hospital following self-harm. Method: A nested case-control study was carried out using a large prospective self-harm database (1997–2010) in Manchester, UK. Characteristics of bipolar cases and non-bipolar controls were compared using conditional logistic regression, and outcomes were assessed via survival analyses. Results: Bipolar cases (n¼ 103) were more likely to repeat self-harm than controls (n¼515): proportion with at least one repeat episode 58% vs. 25%, HR 3.08 (95% CI; 2.2–4.18). Previous self-harm, unemployment, contact with psychiatric services and sleep disturbance were all more common in cases than controls. Even after adjustment for known risk factors, the risk of repetition remained higher in the bipolar group (adjusted HR 1.68; 95% CI; 1.10–2.56). Limitations: The study covers cases from hospital sites in Manchester, UK, and therefore only includes selfharm that was serious enough to present at hospital emergency departments. Conclusion: People with bipolar disorder who self-harm have a higher risk of repetition than people who self-harm more generally. Adjusting for some known risk factors moderated, but did not abolish, this finding. Other factors, such as impulsivity, may also be important. & 2014 Elsevier B.V. All rights reserved.
Keywords: Self-harm Suicidal behaviour Bipolar disorder
1. Introduction People with bipolar disorder are known to be at increased risk of self-harm and suicide; up to 60% of people with bipolar disorder will self-harm at least once during their lifetime (Goodwin and Jamison, 2007; Baldessarini et al., 2006) with at least 5% of the bipolar population eventually dying by suicide (Nordentoft et al., 2011; Tondo et al., 2007). Studies have shown that previous self-harm is a particularly important risk factor for suicide in bipolar populations (Cassidy, 2011; Hawton et al., 2005; Rihmer, 2007), and approximately 60% of those who die by suicide have previously self-harmed (Oquendo et al., 2004; Clements et al., 2013). Given the high risk of n Correspondence to: Centre for Mental Health and Risk, Jean McFarlane Building, The University of Manchester, Oxford Road, Manchester, M20 2UT, UK. Tel.: þ 44 161 275 0735. E-mail address: [email protected] (C. Clements).
http://dx.doi.org/10.1016/j.jad.2014.10.012 0165-0327/& 2014 Elsevier B.V. All rights reserved.
self-harm in bipolar disorder and the association with more dangerous methods in these patients Simon et al., 2007), it is important that those most at risk can be identified early, and treated appropriately. Studies seeking to describe the characteristics of this high risk group could help to inform clinical practice as well as contributing to the research base on the prevention of suicidal behaviour in clinical populations. However, previous work has been subject to a number of methodological limitations; samples are often restricted to a single bipolar subtype, to small sample sizes, inpatient status, or geographical area (Hawton et al., 2005); retrospective patient self-report measures (Lopez et al., 2007); short follow-up times and lack of comparison groups (Goodwin and Jamison, 2007). These limitations have made it difficult to establish how bipolar patients differ from other people who have self-harmed (Hawton et al., 2005). The aims of this study were to describe the characteristics of people with bipolar disorder who self-harmed and to examine outcome in terms of repetition. We compared the
114
C. Clements et al. / Journal of Affective Disorders 173 (2015) 113–119
bipolar sample to a matched sample of non-bipolar controls using data from a large prospective database of hospital attendances for self-harm.
2. Method 2.1. The Manchester Self-Harm Project The analyses used data from the Manchester Self-Harm (MaSH) Project (further information is available at: 〈http://www.bbmh.man chester.ac.uk/cmhr/research/centreforsuicideprevention/MaSH〉). The project is a collaboration, between the University of Manchester, local mental health service providers, and the general hospitals that serve the population of the City of Manchester. Established monitoring methods are used to identify consecutive episodes of self-harm, presenting to the study hospitals. These include detailed searches of emergency department (ED) computer systems, using inclusive search terms (e.g. ‘neck injury’, ‘psychiatric’, ‘left department’), in order to maximise capture of possible episodes of self-harm. The nomenclature used in research on suicidal behaviour and self-harm varies between countries, research studies, and over time, and has been widely discussed within the literature (Silverman et al., 2007a, 2007b; Skegg, 2005). In this study we used the same definition of self-harm as the MaSH project (‘any intentional self-poisoning or self-injury, irrespective of motivation,’ Hawton et al., 2003). Motivations for acts of self-harm can be multiple, fluctuating, and difficult to identify even for the patient and reliably differentiating between those with or without suicidal intent can be problematic (Kapur et al., 2013a). Suicidal intent can change over time, even within the same self-harm episode and people who self-harm with apparently low intent are still at increased risk of repeat self-harm and suicide in the future (Kapur et al., 2013a). We therefore included all types of self-harm, regardless of whether the patient intended to die or not, or medical seriousness. This is also in line with the definition used by the National Institute for Health and Care Excellence (NICE) in clinical guidance on the management of self-harm in the UK (NICE, 2004, 2011). Sociodemographic data (age, sex, postcode and ethnicity) and details on method of self-harm are collected from ED notes, for all episodes, by research clerks. Detailed clinical information is then collected for patients who were assessed by mental health specialists and/or by ED doctors (including previous self-harm, drug or alcohol misuse, past psychiatric treatment, current mental state assessment, a risk assessment and clinical follow-up arrangements), as well as precipitating factors and circumstances of the act. Case ascertainment is essentially complete. Our data covered all self-harm presentations between September 1997 and December 2010. Levels of data completeness for individual variables are also high (Kapur et al., 2013b). Although the MaSH database records each presentation separately, each episode is linked to an identification number, which represents individual patients, and makes it possible to identify those who present repeatedly. 2.2. The nested case-control sample The MaSH database is suited to a nested case-control design, as it represents a well defined source population (of those in the Manchester area who attend hospital following self-harm), of known size, where cases and controls can be selected from a matched and relevant risk set. This method makes optimal use of the data available, for what is a relative rarely recorded event (e.g. self-harm by people with a diagnosis of bipolar disorder).
Bipolar cases were any individual with bipolar disorder recorded in the diagnosis field of the MaSH database. Diagnosis is generated by the assessing clinician, based on current mental state, and/or the patient's medical records. This information is then extracted from these paper assessments and electronic records and entered onto the MaSH database into broad diagnostic categories. The coding for a diagnosis of bipolar disorder on the MaSH database does not therefore distinguish between bipolar subtypes (e.g. bipolar I, bipolar II, rapid cycling etc.), and all diagnoses of bipolar disorder are included together in the analyses. Individuals with any conflicting diagnosis recorded in any other presentations were excluded from the bipolar group (i.e. three individuals were excluded at this point due to a conflicting diagnosis of schizophrenia). Cases and controls were matched only on date of self-harm. Inclusion of all demographic and clinical variables was considered important in order to identify all systematic differences between bipolar and non-bipolar self-harmers. Where there were less than five controls available for a particular date, controls were selected from the following day, and so on, until there were a total of five controls for each case. Matching by date also reduces the impact of any time critical factors that could influence results (e.g. changes in diagnostic criteria, treatment, or assessment protocols). It is common to use multiple controls for each case in nested methodologies, in order to optimise statistical power (Goldstein and Zhang, 2009). In this study we used five controls for each case and power calculations indicated this would give 99% power to detect a 30% difference at a 0.05 level of significance. 2.3. Ethics statement Systematic monitoring of self-harm in Manchester is conducted as part of the local NHS clinical audit programme. It is fully compliant with the UK Data Protection Act 1998. The MaSH Project's use of patient identifiable information without individual patient's consent was supported under the Section 251 of the NHS Act 2006 and approved by the National Information Governance Board for Health and Social Care. The local NHS research ethics committee ratified the project as an audit. 2.4. Analysis All statistical analyses were conducted using Stata/IC 12.1 (StataCorp LP, USA). 2.4.1. Characteristics of self-harm in bipolar disorder Simple descriptive statistics were used to describe the demographic and clinical characteristics of cases and controls. For proportions, the denominator was the number of valid responses for that item, ‘unknown’ responses were removed from each analysis. Variables were examined individually, in a series of univariate conditional logistic regressions, controlling for age and gender (an examination of the distribution of the age variable indicated it was not normally distributed; therefore, age groups were used instead of a continuous age variable). All variables that were significant (at a p value of r0.05) from the initial analyses were entered simultaneously into a multivariate conditional logistic regression, to investigate important discriminators between cases and control. 2.4.2. Repetition of self-harm The MaSH database allows repeat presentations to be linked to an individual patient, thereby providing a robust outcome measure of repeat self-harm, over a variable follow-up period. To make full use of these data we carried out survival analyses. First, the number of
C. Clements et al. / Journal of Affective Disorders 173 (2015) 113–119
repeat presentations in the case-control sample was summarised, and overall differences between groups were examined in a chi-square test of association. The data were then used to plot Kaplan–Meier survival curves, to describe and compare the cumulative survival characteristics of the two study populations. A log-rank test was used to compare the between-groups survival distributions. Preliminary tests indicated these data would be suitable for use in a cox regression model, to examine key explanatory variables, and their relationship to survival, or in this instance, occurrence of repeat self-harm. To that end, ‘diagnosis of bipolar disorder’ was entered into a model with other key variables known to be associated with self-harm (i.e. method of self-harm, previous self-harm, and receiving psychiatric care), and adjusted for age and gender.
3. Results The study included 618 individuals – 103 bipolar cases, and 515 non-bipolar controls. Cases with a recorded diagnosis of bipolar disorder made a total of 269 presentations over the study period, representing around 0.7% of all self-harm presentations. Nearly 60% (n ¼60) of bipolar cases had at least one repeat presentation for self-harm recorded. The average period of follow up after the initial self-harm episode was 7.7 years. 3.1. Case-control analyses 3.1.1. Characteristics of self-harm in bipolar disorder The results of the conditional logistic regression are shown in Table 1, along with the frequencies of demographic characteristics, clinical characteristics, and precipitants of self-harm, in cases and controls. In the univariate analyses the bipolar cases were more likely to be, female, aged between 45 and 64, married or living with a partner, unemployed or registered as long-term sick. Cases were more likely to have sleep disturbance prior to the self-harm act, nearly three times more likely to have had previous episodes of self-harm, and around seven times more likely to have current or previous psychiatric treatment. The act of self-harm was more than twice as likely to be considered a direct response to psychiatric symptoms in the bipolar group, than in the nonbipolar control group. Bipolar cases were less likely to be, aged under 24 years, to have current alcohol use, to have suicidal plans, and to have relationship problems as a precipitant of the self-harm presentation. Table 2 shows the results of the simultaneous entry multivariate conditional logistic regression analyses, incorporating variables that were significant in the univariate analyses. The multivariate model showed that bipolar cases were around three times more likely than non-bipolar controls to be, unemployed, to have received previous psychiatric treatment, and to have sleep disturbance. But, cases were less likely to have experienced relationship problems as a precipitant to the episode of self-harm, to have current alcohol use, or to have suicide plans (i.e. an acknowledgement by the patient, of ongoing intent to die, usually accompanied by some consideration of what method might be used). 3.2. Repetition of self-harm The overall repetition rate (that is, the proportion of the sample with at least one repeat episode at any time during follow up was 58% (n ¼60) for cases and 25% for controls (n ¼129). A chi-square test showed a significant association between diagnosis of bipolar disorder and repeated self-harm (χ² ¼44.58, p o0.001).
115
Fig. 1 shows time to first repeat for cases and controls as a Kaplan–Meier plot. Overall, a higher proportion of cases had a re-presentation for self-harm during the study period compared to controls (log rank χ² ¼50.32; po 0.001). However, there appeared to be very little difference between the groups in terms of time to first repeat. Average time to first repeat was 658 days for cases and 548 days for controls, the difference was not significant (t¼ 0.90; d.f. ¼ 187; p ¼0.37). 3.2.1. Cox regression Diagnosis of bipolar disorder was significant in the cox regression model, with a hazard ratio of 3.08 (95% CI; 2.2–4.18; p o0.01), adjusted for age and gender, indicating that individuals with bipolar disorder were over three times more likely to repeat self-harm than individuals without bipolar disorder. A post-hoc sensitivity analysis, restricting the control sample to those with a formal psychiatric diagnosis (n ¼95), showed the hazard ratio for repeat self-harm in those with bipolar disorder (compared to those with other formal diagnoses) was somewhat reduced but remained elevated (HR, 2.40; 95% CI; 1.56–3.70; po 0.01). Table 3 gives the results of the multivariate model. The inclusion of additional explanatory variables did reduce the value of the hazard ratio for diagnosis of bipolar disorder, indicating that these variables also had an influence on the repetition of suicidal behaviour. However, the presence of bipolar disorder still maintained significance at the 0.05 level in the adjusted analysis (HR, 1.68; 95% CI; 1.10–2.56; po 0.02), showing that repetition may well have been partly explained by the presence of this variable. Previous self-harm was the strongest predictor (HR, 15.28; 95% CI; 6.17–37.84; po 0.01) in the model, and current psychiatric treatment was also significant (HR, 1.80; 95% CI; 1.20–2.70; p o0.01). 3.2.2. Suicide deaths Mortality outcome data was available for 79% (n ¼491) of the case-control sample (mortality data were not available for the earlier years of the study because systematic data linkage was only carried out from 2001 onwards). In terms of fatal repetition, 4 (3.9%) of the bipolar cases died by suicide during the follow up period (receiving a suicide or undetermined verdict at Coroner's Inquest; Linsley et al., 2001) compared to 7 (1.4%) controls. Suicide mortality appeared higher in the bipolar group but, probably because of small numbers, this difference failed to reach statistical significance (χ²¼3.1, p ¼0.08). 4. Discussion 4.1. Main findings This study used a nested case-control method to describe the characteristics of people with bipolar disorder, who presented to EDs in Manchester, UK, following self-harm. Univariate analyses showed some differences in characteristics between cases and controls, most notably in relation to female gender, age between 45 and 64 years, unemployment, registered as long term sick, a history of previous self-harm, current or previous psychiatric care. In subsequent multivariate analysis, unemployment, previous psychiatric treatment, and sleep disturbance were more common in the bipolar cases, while current alcohol use, suicidal plans and relationship problems (as a precipitant of self-harm) were significantly more common among the non-bipolar control group. Repeat self-harm was significantly associated with a diagnosis of bipolar disorder, with a much larger proportion of cases repeating at least once during the 13 year study period, compared to controls (58% vs. 25%). Although there was no significant difference in time to first repeat between the two groups, survival
116
C. Clements et al. / Journal of Affective Disorders 173 (2015) 113–119
Table 1 Socio-demographic and clinical characteristics of bipolar cases and non-bipolar controls who had self-harmed. Odds ratios and p values calculated using univariate conditional logistic regression analysis. Variable
Cases N¼ 103 n (%)
Controlsa N ¼ 515 n (%)
Odds ratio (95% CI)
Female Age 24 and under 25 to 44 45 and 64 65 and over Ethnicity White Black Indian/Bangladeshi/Pakistani Other ethnicity Marital status Single Separated/divorce Married/partner Widowed Employment status Employed Unemployed Registered sick Retired Student Homemaker/carer Other employment Living arrangements Homeless Lives alone Lives with family Lives with friends Other Method of self-harm Self-poisoning, drugs Self-poisoning, other Self-injury Other History of self-harm Previous self-harm Self-harm in previous 12 months Psychiatric care Current psychiatric treatment Previous psychiatric treatment Clinical characteristics Current alcohol use Current substance misuse Feeling depressed Looking depressed Sleep disturbance Appetite disturbance Hopelessness Suicidal thoughts Suicide plans Hallucinations Precipitants of self-harm Relationship problems partner/friend Relationship problems parents/siblings Relationship problems others Bullying/intimidation Bereavement Housing Work Legal Health problems Miscarriage Money Direct response to psychiatric symptoms Other psychiatric problems Abuse Due to alcohol abuse Due to substance abuse Other problems
72(69.90)
283(54.95)
1.771(1.117–2.809)
21(20.39) 49(47.57) 29(28.16) 4(3.88)
192(37.35) 252(49.03) 59(11.48) 11(2.14)
0.416(0.245–0.708) 0.968(0.630–1.485) 2.979(1.756–5.052) 1.787(0.513–6.227)
90(88.24) 3(2.94) 8(7.84) 1(0.98)
400(88.30) 11(2.43) 33(7.28) 5(1.10)
0.878(0.434–1.777) 1.015(0.245–4.203) 1.438(0.624–3.316) 1.177(0.127–10.906)
0.717 0.984 0.394 0.886
42(44.21) 13(13.68) 38(40.00) 2(2.11)
271(57.78) 72(15.35) 116(24.73) 10(2.13)
0.856(0.508–1.444) 0.546(0.266–1.121) 1.746(1.059–2.878) 0.407(0.069–2.408)
0.560 0.099 0.029n 0.322
16(16.67) 45(46.88) 20(20.83) 5(5.21) 6(6.25) 3(3.13) 1(1.04)
115(28.19) 141(34.56) 46(11.27) 16(3.92) 54(13.24) 25(6.13) 11(2.70)
0.555(0.304–1.010) 1.890(1.154–3.095) 2.028(1.047–3.930) 0.229(0.025–2.102) 0.614(0.230–1.642) 0.395(0.113–1.377) 0.230(0.028–1.872)
0.054 0.011n 0.036n 0.193 0.331 0.145 0.169
1(1.08) 18(19.35) 48(51.61) 9(20.45) 9(9.68)
8(2.04) 93(23.72) 193(49.23) 35(79.55) 38(9.69)
0.609(0.070–5.314) 0.665(0.361–1.224) 1.230(0.760–1.990) 1.303(0.570–2.980) 1.073(0.483–2.385)
0.653 0.190 0.400 0.531 0.863
84(81.55) 2(1.94) 14(13.59) 3(2.91)
414(80.54) 4(0.78) 75(14.59) 21(4.09)
0.889(0.506–1.561) 2.210(0.330–14.800) 1.061(0.568–1.981) 1.025(0.284–3.694)
0.682 0.414 0.853 0.970
71(73.96) 33(39.29)
187(48.95) 112(36.96)
2.983(1.726–5.153) 1.043(0.567–1.918)
o 0.001n 0.892
75(79.79) 78(87.64)
135(34.88) 172(44.44)
7.111(3.854–13.119) 6.889(3.480–13.637)
o 0.001n o 0.001n
14(16.28) 7(8.14) 58(64.44) 46(52.27) 61(67.78) 42(46.67) 32(37.21) 33(37.08) 7(7.95) 15(16.85)
111(29.52) 49(13.21) 251(65.54) 177(46.34) 203(53.14) 160(41.67) 155(40.90) 144(37.50) 70(18.47) 34(8.88)
0.457(0.237–0.881) 0.790(0.316–1.977) 1.029(0.735–1.440) 1.007(0.611–1.659) 1.737(1.022–2.951) 1.154(0.706–1.886) 0.637(0.377–1.076) 0.859(0.511–1.444) 0.362(0.146–0.898) 1.865(0.928–3.745)
0.019n 0.615 0.869 0.977 0.041n 0.567 0.092 0.567 0.028n 0.080
23(26.14) 15(17.05) 13(14.77) 2(2.27) 8(9.09) 12(13.64) 8(9.09) 2(2.27) 5(5.68) 1(1.14) 10(11.36) 31(35.23) 2(6.67) 5(5.68) 1(3.33) 1(3.33) 8(9.09)
161(42.82) 71(18.88) 38(10.11) 20(5.32) 34(9.04) 36(9.57) 42(11.17) 13(3.46) 34(9.04) 4(1.07) 44(11.70) 53(14.10) 8(7.84) 22(5.85) 20(19.61) 6(5.88) 49(13.03)
0.477(0.268–0.849) 0.905(0.461–1.777) 1.577(0.738–3.369) 0.463(0.102–2.106) 0.808(0.322–2.027) 1.860(0.853–4.060) 1.028(0.450–2.350) 0.807(0.168–3.868) 0.707(0.248–2.019) 1.175(0.112–12.364) 1.397(0.630–3.096) 2.319(1.322–4.068) 3.156(0.436–22.832) 1.108(0.336–3.656) 0.176(0.021–1.494) 1.472(0.098–22.087) 0.612(0.259–1.447)
0.012n 0.772 0.240 0.319 0.649 0.119 0.948 0.788 0.518 0.893 0.410 0.003n 0.255 0.867 0.111 0.779 0.264
n
p Value 0.015n 0.001n 0.880 o 0.001n 0.362
Significant at pr 0.05 Diagnostic breakdown of controls: n¼ 420 (82%) no formal psychiatric diagnosis recorded; n¼ 40 (8%) depressive disorders; n¼ 15 (3%) adjustment/stress disorders; n¼ 15 (3%) alcohol misuse; n ¼25 (5%) other psychiatric disorders. a
C. Clements et al. / Journal of Affective Disorders 173 (2015) 113–119
Table 2 A simultaneous entry multivariate conditional logistic regression comparing the presence of key characteristics in people with bipolar disorder who self-harm, and people without bipolar disorder who self-harm. Variable
OR(95% CI)
p
Female 25 to 44 45 and 64 65 and over Married/partner Unemployed Long term sick Relationship problems partner/friend Direct response to psychiatric symptoms Previous self-harm Current psychiatric treatment Previous psychiatric treatment Current alcohol use Sleep disturbance Suicide plans
1.830(0.834–4.018) 1.026(0.381–2.767) 0.931(0.280–3.099) 1.192(0.111–12.814) 2.060(0.878–4.833) 3.081(1.210–7.849) 2.180(0.682–6.960) 0.305(0.116–0.802) 0.639(0.238–1.721) 2.253(0.927–5.476) 2.176(0.769–6.157) 2.955(1.049–8.327) 0.310(0.100–0.957) 2.945(1.287–6.738) 0.245(0.070–0.859)
0.132 0.959 0.907 0.885 0.097 0.018n 0.188 0.016n 0.376 0.073 0.143 0.040n 0.042n 0.011n 0.028n
n
Significant at pr 0.05.
Kaplan-Meier survival estimates
100%
75%
50%
25%
0% 0
1000
2000
3000
4000
5000
Days to first repeat Other cases
Bipolar cases
Fig. 1. Kaplan–Meier survival curves, showing time to first repeat presentation for self-harm, for cases and controls.
Table 3 Results of a simultaneous entry multivariate cox regression, examining the association between bipolar disorder and repeat self-harm. Key risk factors, known to increase risk of self-harm, have been included in the model as co-variates. Variable
Hazard ratio(95% CI)
p
Bipolar disorder 25 to 44 45 and 64 65 and over Female Poisoning (vs. injury) Previous self-harm Current psychiatric care
1.675(1.098–2.556) 0.886(0.592–1.326) 0.645(0.364–1.143) 0.990(0.357–2.744) 1.154(0.743–1.790) 0.503(0.292–0.867) 15.284(6.174–37.838) 1.797(1.195–2.703)
0.017n 0.557 0.133 0.985 0.524 0.013n o0.001n 0.005n
n
Significant at pr 0.05
analysis showed that a diagnosis of bipolar disorder did have an independent effect on the risk of repetition, even when other known predictors (e.g. previous self-harm) were included in the model.
117
have shown a greater gender equality in self-harm (Cassidy, 2011). Our results showed that females were over represented among bipolar cases, even in comparison to ‘other’ self-harmers – a group that is already known to be disproportionately female (Hawton et al., 2003). One possible explanation is that, although bipolar disorder itself is unrelated to gender, women are more likely to seek help, and therefore, more likely to be in contact with services and given an appropriate diagnosis. Another explanation is that, while self-harm methods used by people with bipolar disorder tend to be more potentially lethal in general (Cassidy, 2011), there is evidence that methods may be particularly dangerous in men with bipolar disorder, with a higher case fatality (D’Ambrosio et al., 2012). Although not significant in the multivariate analysis (due to the presence of ‘previous psychiatric treatment' in the model which may be acting as a confounder), previous self-harm was significant in the univariate analysis, and the difference in proportions between cases and controls was large, at 74% vs. 49%. This is in line with a number of previous studies on both lethal and non-lethal self-harm in bipolar disorder (Hawton et al., 2005; Rihmer, 2007). It might also be of note that, ‘direct response to psychiatric symptoms’ was the only significant predictor within the precipitants of self-harm variables, and it was more than twice as common in the bipolar cases. If this is considered along with the tendency for the bipolar cases to be in current or previous contact with psychiatric services, and to have previously self-harmed, this could be indicative of a group who are more seriously unwell, and who might have had a longer duration of illness. There is some evidence that those with longer durations of untreated illness/ incorrect diagnosis, greater illness severity, and illness burden, have a higher frequency of suicidal behaviour and poorer outcomes (Baldessarini et al., 2006; Tondo et al., 2007). By contrast those without bipolar disorder were more likely to report relationship problems as a precipitant. While depression and hopelessness were present, to some extent, across all cases and controls, as might be expected (Hawton et al., 2005), these were not significant predictors of repetition. Of the clinical characteristics described in this dataset, only sleep disturbance seemed to be a robust independent predictor in bipolar disorder (Rihmer, 2007), and indeed, was almost three times more likely to be present in bipolar cases than non-bipolar controls. Previous work looking at the association of drug and/or alcohol use with suicidal behaviour in bipolar disorder has produced contradictory results (Cassidy, 2011). We found that alcohol use was, in fact, recorded less frequently in cases than controls. This is somewhat surprising given that alcohol and substance use is recognised as common in people with bipolar disorder (Goodwin and Jamison, 2007). However, there is evidence that alcohol/ substance use may not be not strongly related to suicidal behaviour in bipolar disorder (Hawton et al., 2005), and it may be that illness related factors play a more important role in this group. For example, suicidal plans were much less common in bipolar cases, which supports the hypothesis that impulsivity, a characteristic trait in bipolar patients, may be of particular importance (Tondo and Baldessarini, 2005). Impulsivity, driven by positive or negative affect, in the presence of other factors recognised associated with self-harm and suicide more generally, may be especially important in precipitating episodes of self-harm, and may be a key area for future research. 4.3. Methodological limitations
4.2. Findings in relation to previous research Our study supports work that has shown that women with bipolar disorder have a higher rate of suicidal behaviour than men (Tondo et al., 2003), and contrasts with the findings of studies that
People with bipolar disorder are at high risk of self-harm (Goodwin and Jamison, 2007; Baldessarini et al., 2006), and it is possible that some people in the control group could have undiagnosed bipolar disorder. The number of cases with a
118
C. Clements et al. / Journal of Affective Disorders 173 (2015) 113–119
diagnosis of bipolar disorder was small in comparison to the size of the overall dataset (less than 1% of total presentations). This suggests that psychiatric diagnoses in general may be underreported on the MaSH database. This is in line with previous reviews, which suggest that only a small proportion of self-harm patients receive a psychiatric diagnosis when they present to hospitals (Effective Health Care Bulletin, 1998). This may be due to time pressures within the emergency department, the need for rapid assessment and turnover of self-harm patients, or reluctance to assign a diagnosis on the basis of one assessment. Diagnosis is assigned by the clinician at assessment, and reflects mental state at the time, as well as historical information taken from the patient and medical records. Where a specific diagnosis is unclear the field will remain empty, and any diagnosis given is, therefore, more likely to be based on clear evidence, which may favour those with more pronounced symptoms and more severe illness. Misdiagnosis could also be a problem within the bipolar disorder cases, where there is known to be difficulty distinguishing bipolar disorder from borderline personality disorder—which is also associated with high risk of self-harm. However, this is true of most studies of bipolar disorder that rely on past records and clinical judgement, and verifying diagnosis is beyond the scope of the current study. The effect of bias from such missed or misallocated bipolar disorder cases/controls would be to minimise between-group differences described in this study, and the fact that the differences seen in our findings are consistent with previous work is reassuring. The control sample contained data on 95 (18%) individuals who had a psychiatric diagnosis, including: depressive illnesses (n ¼43, 8%), anxiety and stress disorders (n ¼20, 4%), alcohol misuse/ dependence (n ¼18, 3%), and personality disorder (n ¼ 3, 0.6%). It was not our intention to compare self-harm between diagnostic groups in this study. Instead, we aimed to investigate factors which distinguish self-harm in bipolar disorder from self-harm in general, which will include a proportion of people with psychiatric diagnoses. We believe that our study has potentially important implications for clinicians when they encounter a person with bipolar disorder who has self-harmed. As the majority of controls (over 80%) had no recorded psychiatric diagnosis it could be that some of the findings were not specific to those with bipolar disorder but reflect wider risks for individuals with mental illness. However, a post-hoc sensitivity analysis showed that the hazard ratio for repeat self-harm in those with bipolar disorder compared to controls with a psychiatric diagnosis remained elevated with a hazard ratio of 2.40 (compared to a hazard ratio of 3.08 for the whole sample). Another factor that might have reduced between-group differences was the presence of co-morbidities among bipolar disorder cases. However, only 11 (11%) of cases had a co-morbid psychiatric diagnosis recorded on the MaSH database; depressive illnesses were most common (n¼ 5, 5%), followed by anxiety and stress disorders (n ¼3, 3%), alcohol misuse (n ¼2, 2%), and learning difficulties (n ¼1, 1%). Nineteen (20% of the controls with a psychiatric diagnosis, 4% of all controls) were recorded as having co-morbid psychiatric diagnoses. The dataset used is geographically restricted to one city in England, with particular economic and socio-demographic characteristics, and it is possible that it may not be valid to extrapolate findings to other areas. Data also only cover self-harm acts that are serious enough to present at hospital and these may be a systematically different group to less medically-serious forms of selfharm. However, this would be true across both cases and controls and does not reduce the importance of the findings for those who do present to hospital following self-harm. There are also a number of psychological characteristics that are of current
research interest in relation to self-harm and suicide in bipolar disorder, such as temperament (Pompili et al., 2012; Sarısoy et al., 2012), and impulsivity (Mahon et al., 2012; Watkins & Meyer, 2013), which we acknowledge are important, but did not have sufficient detail within the data to explore in this study. 4.4. Clinical implications In this study we found that repetition of self-harm was particularly common in individuals with bipolar disorder – nearly six out of ten repeated during the follow up period. We found that people with bipolar disorder who self-harmed, typically had a number of high risk characteristics (e.g. previous self-harm, history of psychiatric care). However, these characteristics, recognised as risk factors in general and psychiatric self-harm populations, did not comprehensively explain the increased risk associated with a diagnosis of bipolar disorder. Previous work suggests that impulsivity is associated with self-harm in bipolar disorder (Tondo and Baldessarini, 2005; Swann et al., 2005), and this may explain why bipolar disorder cases in this study were less likely to have ‘suicidal plans’, even though the proportion of those with suicidal thoughts was similar between groups. Sleep disturbance was common and an independent predictor of self-harm in the bipolar group. This may be a marker of changes in illness, agitation or mood-state, as found in mixed depressive episodes and known to be associated with self-harm in bipolar (Valtonen et al., 2008), and depressive symptoms were evident in nearly 65% of the bipolar cases. We were not able to assess psychiatric comorbidity, severity, or duration of illness, in this study. However, given the frequency of previous self-harm and current/previous psychiatric care in the bipolar cases, it is likely that those with bipolar disorder who self-harm, have a more chronic and complicated course of illness than those with bipolar disorder who do not self-harm. The key precipitant that distinguished the bipolar group was ‘a direct response to psychiatric symptoms’ which indicates that illness-related factors, may play a larger role in acts of selfharm in bipolar disorder, than any specific negative life events. Clinicians need to be aware of the higher risk of self-harm in bipolar disorder, and the link between previous self-harm and suicide (Clements et al., 2013). Of course these risks might not be restricted to those with bipolar disorder and those with other psychiatric disorders may also be at higher risk. The risk is not restricted to the short term – the average time to repetition approached 2 years in this study. Those who repeat are also at higher risk for completed suicide (Zahl and Hawton, 2004). Sixteen per cent (n¼ 13) of bipolar cases who were in previous contact with psychiatric services, were recorded as not in current contact with services. As risk appears to be enduring, continued contact with services may be appropriate. The impulsivity of self-harm in bipolar disorder may make it more difficult to predict when patients are most at risk. However, good quality monitoring and management of symptoms could help. Patients, who report sleep disturbance and/or changes in mood, might be considered to be at increased risk of engaging in suicidal behaviours, and treated accordingly. The use of lithium, which has been shown to have specific benefits for prevention of suicide (Tondo et al., 2003; Cipriani et al., 2013), and other mood stabilisers, may be beneficial. Psychological approaches, such as cognitive behavioural therapies, which have been shown to have a positive impact on suicidal behaviour in general, might also have an important role in prevention (Tarrier et al., 2008). Studies that could draw out differences in risk of self-harm between subtypes of bipolar disorder would be valuable, along with more detailed work to draw out and describe the influence of risk factors in the period that precedes acts of self-harm. Qualitative studies would allow for closer examination of antecedents and
C. Clements et al. / Journal of Affective Disorders 173 (2015) 113–119
the experience of them at an individual level could help to identify when and how services might be able to intervene appropriately to reduce self-harm in this potentially high risk population. Role of funding source This article presents independent research funded by the National Institute for Health Research (NIHR) under its Programme Grants for Applied Research programme (Reference number RP-PG-0407-10389). The views expressed are those of the author(s) and not necessarily those of the NHS, the NIHR or the Department of Health.
Conflict of interest There are no conflicts of interest to declare.
Acknowledgements We would like to thank Sarah Steeg, Iain Donaldson and the rest of the MaSH team for their help in organising and accessing data for this study.
References Baldessarini, R.J., Pompili, M., Tondo, L., 2006. Suicide in bipolar disorder: risks and management. CNS Spectr. 11, 465–471. Cassidy, F., 2011. Risk factors of attempted suicide in bipolar disorder. Suicide Life Threat Behav. 41, 6–11. Cipriani, A., Hawton, K., Stockton, S., Geddes, J., 2013. Lithium in the prevention of suicide in mood disorders: updated systematic review and meta-analysis. BMJ 346, f3646. Clements, C., Morriss, R., Jones, S., Peters, S., Roberts, C., Kapur, N., 2013. Suicide in bipolar disorder in a national English sample, 1996–2009: frequency, trends, and characteristics. Psychol. Med. 43, 2593–2602. D'Ambrosio, V., Salvi, V., Bogetto, F., Maina, G., 2012. Serum lipids, metabolic syndrome and lifetime suicide attempts in patients with bipolar disorder. Prog. Neuropsychopharmacol. Biol. Psychiatry 37, 136–140. Goldstein, L., Zhang, H., 2009. Efficiency of the maximum likelihood estimator for nested case control sampling. Bernoulli 15, 569–597. Goodwin, F.K., Jamison, K.R., 2007. Manic-Depressive Illness: Bipolar Disorders and Recurrent Depression. Oxford University Press, New York, NY. Hawton, K., Harriss, L., Hall, S., Simkin, S., Bale, E., Bond, A., 2003. Deliberate selfharm in Oxford, 1990–2000: a time of change in patient characteristics. Psychol. Med. 33, 987–996. Hawton, K., Sutton, L., Haw, C., Sinclair, J., Harris, L., 2005. Suicide and attempted suicide in bipolar disorder: a systematic review of risk factors. J. Clin. Psychiatry 66, 693–704. Kapur, N., Cooper, J., O'Connor, R., Hawton, K., 2013a. Non-suicidal self-injury v. attempted suicide: new diagnosis or false dichotomy? Br. J. Psychiatry 202, 326–328. Kapur, N., Steeg, S., Webb, R., Haigh, M., Bergen, H., Hawton, K., et al., 2013b. Does clinical management improve outcomes following self-harm? Results from the Multicentre Study of Self-Harm in England. PLoS ONE 8, e70434. http://dx.doi. org/10.1371/journal.pone.0070434. Linsley, K.R., Schapira, K., Kelly, T.P., 2001. Open verdict vs. suicide—importance to research. Br. J. Psychiatry 178, 465–468.
119
Lopez, V.A., Detera-Wadleigh, S., Cardona, I., Kassem, L., McMahon, F.J., 2007. Nested association between genetic variation in tryptophan hydroxylase II, bipolar affective disorder, and suicide attempt. Biol. Psychiatry 61, 181–186. Mahon, K., Burdick, K.E., Wu, J., Ardekani, B.A., Szeszko, P.R., 2012. Relationship between suicidality and impulsivity in bipolar I disorder: a diffusion tensor imaging study. Bipolar Disord. 14, 80–89. NICE, 2004. Self-harm: the short-term physical and psychological management and secondary prevention of self-harm in primary and secondary care, NICE Clinical Guideline 16. Available at: 〈www.nice.org.uk/CG16〉. NICE, 2011. Self-harm: longer-term management, NICE Clinical Guideline 133. Available at 〈www.nice.org.uk/CG133〉. Nordentoft, M., Mortensen, P., Pedersen, C., 2011. Absolute risk of suicide after first hospital contact in mental disorder. Arch. Gen. Psychiatry 68, 1058–1064. Oquendo, M.A., Galfalvy, H., Russo, S., Ellis, S.P., Grunebaum, M.F., Burke, A., et al., 2004. Prospective study of clinical predictors of suicidal acts after a major depressive episode in patients with major depressive disorder or bipolar disorder. Am. J. Psychiatry 161, 1433–1441. Pompili, M., Innamorati, M., Rihmer, Z., Gonda, X., Serafini, G., Akiskal, H., et al., 2012. Cyclothymic-depressive-anxious temperament pattern is related to suicide risk in 346 patients with major mood disorders. J. Affect. Disord. 136, 405–411. Rihmer, Z., 2007. Suicide risk in mood disorders. Curr. Opin. Psychiatry 20, 17–22. Sarısoy, G., Kaçar, O.F., Pazvantoğlu, O., Oztürk, A., Korkmaz, I.Z., Kocamanoğlu, B., et al., 2012. Temperament and character traits in patients with bipolar disorder and associations with attempted suicide. Compr. Psychiatry 53, 1096–1102. Silverman, M.M., Berman, A.L., Sanddal, N.D., O'Carroll, P.W., Joiner, T.E., 2007a. Rebuilding the tower of Babel: a revised nomenclature for the study of suicide and suicidal behaviors. Part 1: Background, rationale, and methodology. Suicide Life Threat Behav. 37, 248–263. Silverman, M.M., Berman, A.L., Sanddal, N.D., O'carroll, P.W., Joiner, T.E., 2007b. Rebuilding the tower of Babel: a revised nomenclature for the study of suicide and suicidal behaviours. Part 2: Suicide-related ideations, communications, and behaviors. Suicide Life Threat Behav. 37, 264–277. Simon, G.E., Hunkler, E., Fireman, B., Lee, J.Y., Savarino, J., 2007. Risk of suicide attempt and suicide death in patients treated for bipolar disorder. Bipolar Disord. 9, 526–530. Skegg, K., 2005. Self-harm. Lancet 366, 1471–1483. Swann, A.C., Dougherty, D.M., Pazzaglia, P.J., Pham, M., Steinberg, J.L., Moeller, F.C., 2005. Increased impulsivity associated with severity of suicide attempt history in patients with bipolar disorder. Am. J. Psychiatry 162, 1680–1687. Tarrier, N., Taylor, K., Gooding, P., 2008. Cognitive-behavioural interventions to reduce suicide behaviour. Behav. Modif. 32, 77–108. Tondo, L., Baldessarini, R., 2005. Suicidal risk in bipolar disorder. Clin. Neuropsychiatry 2, 55–65. Tondo, L., Isacsson, G., Baldessarini, R., 2003. Suicidal behaviour in bipolar disorder. CNS Drugs 17, 491–511. Tondo, L., Lepri, B., Baldessarini, R.J., 2007. Suicidal risk among 2826 Sardinian major affective disorder patients. Acta Psychiatr. Scand. 116, 419–428. Valtonen, H.M., Suominen, K., Haukka, J., Mantere, O., Leppamaki, S., Arvilommi, P., et al., 2008. Differences in incidence of suicide attempts during phases of bipolar I and II disorders. Bipolar Disord. 10, 588–596. Watkins, H.B., Meyer, T.D., 2013. Is there an empirical link between impulsivity and suicidality in bipolar disorders? A review of the current literature and the potential psychological implications of the relationship. Bipolar Disord. 15, 542–558. Zahl, D.L., Hawton, K., 2004. Repetition of deliberate self-harm and subsequent suicide risk: long-term follow-up study of 11,583 patients. Br. J. Psychiatry 185, 70–75.