Suicidal ideation and suicide plans or attempts in adults with Asperger’s syndrome attending a specialist diagnostic clinic: a clinical cohort study

Summary Background Asperger’s syndrome in adulthood is frequently associated with depression, but few studies have explored the lifetime experience of self-reported suicidal ideation and suicide plans or attempts in this clinical group. We aimed to assess this prevalence in a clinical cohort of patients in the UK. Method In a clinical cohort study, we undertook a retrospective analysis of clinical survey data from adults newly diagnosed with Asperger’s syndrome at a specialist diagnostic clinic between Jan 23, 2004, and July 8, 2013, in England. Patients completed a self-report questionnaire before clinical assessment, recording lifetime experience of depression, suicidal ideation, and suicide plans or attempts, along with self-reported measures of autistic traits and empathy. We compared the rate of suicidal ideation in the sample with published rates of suicidal ideation in the general population and other clinical groups. We also assessed associations between depression, autistic traits, empathy, and likelihood of suicidal ideation and suicide plans or attempts. Findings 374 adults (256 men and 118 women) were diagnosed with Asperger’s syndrome in the study period. 243 (66%) of 367 respondents self-reported suicidal ideation, 127 (35%) of 365 respondents self-reported plans or attempts at suicide, and 116 (31%) of 368 respondents self-reported depression. Adults with Asperger’s syndrome were signiﬁ cantly more likely to report lifetime experience of suicidal ideation than were individuals from a general UK population sample (odds ratio 9·6 [95% CI 7·6–11·9], p<0·0001), people with one, two, or more medical illnesses (p<0·0001), or people with psychotic illness (p=0·019). Compared with people diagnosed with Asperger’s syndrome without depression, people with Asperger’s syndrome and depression were more likely to report suicidal ideation (p<0·0001) and suicide plans or attempts (p<0·0001


Introduction
The autism spectrum is a set of heterogeneous neurodevelopmental conditions, characterised by diffi culties in social communication and unusually restric tive and repetitive behaviours and interests. 1 Although the term autism spectrum disorder is widely adopted, autism spectrum condition avoids the negative overtones of the disorder label, while recognising autism as a medical diagnosis for which individuals need support. Asperger's syndrome is a subgroup on the autism spectrum, 2 showing core symptoms in the absence of language delay or intellectual disability. A clinical diagnosis is often delayed until about 11 years of age, 3 or even into adulthood, 4-6 compared with the average age of 5 years for a diagnosis of classic autism. Transition into adulthood for people with Asperger's syndrome is often accompanied by a lack of support services 7 and poor outcomes in terms of health and social diffi culties, 8,9 quality of life, achievement of occupational potential, 10 social exclusion and isolation, 9,11 and high rates of depression. 12,13 Depression is a risk factor for completed suicide, and more than 90% of people who die by suicide have depression. 14 In view of the evidence for high incidence of poor outcomes and depression in adults with Asperger's syndrome, these individuals might be at particular risk of experiencing suicidal ideation and suicide plans or attempts. However, very few studies have explored suicidality in this group. Suicidal ideation is anecdotally reported to be very common in people with Asperger's syndrome, especially in adolescence and early adult life. According to Gillberg,15 suicidal acts in Asperger's syndrome are anecdotally reported to be quite frequent from the onset of puberty, and can be associated with a preoccupation with death, a reaction to stress associated with a depressive disorder, social isolation, low selfesteem, or feelings of social exclusion.
The scarce empirical evidence available (just four studies 8,[16][17][18] ) suggests that suicidal ideation and attempts are common in individuals with autism spectrum conditions. First, parent-reported rates of suicidal ideation and attempts are signifi cantly higher for children with autism (14%) than for typically developing children (0·5%). 16 Second, the rate of suicidal ideation and attempts in 102 young people with autism spectrum conditions and comorbid anxiety was 11%. 17 Third, in a small community sample of 42 adults with Asperger's syndrome, the rate of suicidal ideation was 40% and the rate of suicide attempts was 15%. 8 Finally, in 26 adult psychiatric patients diagnosed with autism spectrum conditions, 30·8% presented with suicidal ideation and two had died by suicide. 18 However, the lifetime experience of suicidal ideation and suicide plans or attempts in a large sample of adults with Asperger's syndrome, how it compares with other clinical groups, and what risk factors (particularly depression) are associated with suicidality in adults with Asperger's syndrome, is unclear from these studies.
The Cambridge Lifespan Asperger Syndrome Service (CLASS) is a specialist diagnostic clinic established in 1999 with charitable funding because the UK National Health Service (NHS) did not recognise the clinical needs of adults seeking late diagnosis of possible Asperger's syndrome. The clinic provides a diagnostic service to adults in recognition that many adults with possible Asperger's syndrome but born before 1994 (when Asperger's syndrome was fi nally recognised in the Diagnostic and Statistical Manual of Mental Disorders IV 19 and the WHO International Classifi cation of Diseases 10 2 ) would have been missed. In 2011, the clinic became part of the Cambridgeshire and Peterborough NHS Foundation Trust, following the UK Parliament's Autism Act (2010) and the UK Department of Health's Autism Strategy (2011), making the provision of such clinics a statutory requirement. Because existing gold standard diagnostic instruments for autism spectrum conditions (the Autism Diagnostic Interview-Revised 20 and the Autism Diagnostic Observation Schedule 21 ) miss up to 50% of adults with a diagnosis of Asperger's syndrome or high-functioning autism 22 because they were designed to detect autism spectrum conditions in childhood, the CLASS clinic developed a diagnostic tool for adults, called the Adult Asperger Assessment. 23 This test is based on (but is more stringent than) the criteria in the Diagnostic and Statistical Manual of Mental Disorders IV. 19 We analysed survey data from a large clinical cohort of people with Asperger's syndrome to compare rates of suicidal ideation with known rates in the general population and other clinical groups, and explore risk factors for suicidal ideation and suicide plans or attempts (ie, depression and self-reported autistic symptoms) in this population.

Participants
Between Jan 23, 2004, and July 8, 2013, the CLASS clinic in Cambridge, UK, routinely administered a patient screening questionnaire before appointments to all individuals newly diagnosed with Asperger's syndrome. Diagnoses were made on the basis of a face-to-face interview by an expert clinician, who used both the Adult Asperger Assessment 23 and clinical judgment, with input from an informant (usually a parent) who knew the patient in their early childhood when possible. We identifi ed participants from the CLASS clinic database.
Ethics approval for the study to migrate anonymised clinic data into the Cambridge University Autism Research Centre for research purposes was granted by the Cambridge University Psychology Research Ethics Committee, with approval by the Cambridgeshire and Peterborough NHS Foundation Trust Research and Development Offi ce.

Assessment
All patients were required to complete a self-report patient screening questionnaire before being invited for further clinical assessment. Relevant questions were "have you ever been diagnosed with depression?", "have you ever felt suicidal?", and "if yes, have you ever planned or attempted suicide?". Patients posted the questionnaire back in a freepost envelope before the appointment. The responses to these questions were entered into a clinic database with the demographic details and diagnostic outcome of each patient. These data were gathered as part of the routine practice of the clinic. If patients reported any lifetime experience of suicidality or depression, it was explored further during a face-to-face clinical assessment interview.
Patients were also screened before their appointment with the Empathy Quotient, 24 a self-report questionnaire that quantifi es individual diff erences in empathy. Scores range from 0 to 80, with highest scores suggesting greatest empathy. Empathy Quotient scores can show sex diff erences (with typically developing women on average scoring higher than typically developing men), and individuals with autism on average scoring lower than typically developing men. 24 80% of patients with Asperger's syndrome score less than 30 on the Empathy Quotient, compared with 12% of the general population. 24 Patients were asked to complete the Autism Spectrum Quotient questionnaire, which measures the degree to which an adult with average or above-average intelligence has cognitive-behavioural traits associated with autism. 25 Scores range from 0 to 50, with highest scores suggesting most autistic traits. The Autism Spectrum Quotient has consistently showed higher mean scores in those with autism spectrum conditions than in typically developing individuals. 25-27 80% of patients with Asperger's syndrome score 32 or higher on the Autism Spectrum Quotient, compared with only 2% of the general population. 25 Empathy Quotient and Autism Spectrum Quotient scores were fed into the Adult Asperger Assessment electronically, before the assessment, providing guidance to the clinician about how to gather evidence for symptoms causing impairment in functioning. Because the Empathy Quotient and Autism Spectrum Quotient are collected before the appointment, they also serve as a confi rmation that the referral is likely to be an appropriate use of clinic time, because each appoint ment takes up to 3 h.
We identifi ed comparison samples from the general population, general medical population, and psychi atric population through searches of Medline, PsycINFO, and Google Scholar for reports published between Jan 1, 1995, and Sept 20, 2013, with the general search term "suicid*" in combination with search terms for each disorder (eg, "ADHD", "autism"). To be eligible for inclusion, studies had to include self-report questions on the lifetime experience of suicidal ideation in adult men and women between 16 and 75 years of age.

Statistical analyses
We anonymised data in Microsoft Excel 2010 and analysed the results in PASW Statistics version 18.0. We used Pearson's χ² tests to explore associations between sex, depression, and suicidal ideation and suicide plans or attempts, and to compare lifetime experience of suicidal ideation in the present sample with that of other    clinical groups. We calculated odds ratios (ORs) and 95% CIs as a measure of eff ect size. We used Bonferronicorrected independent samples t tests to compare Autism Spectrum Quotient scores, Empathy Quotient scores, and age between participants who reported depression, suicidal ideation, or suicide plans or attempts.

Role of the funding source
The funders had no role in the study design, data collection, analysis, data interpretation, or writing of the report. The corresponding author had full access to all the data in the study and fi nal responsibility for the decision to submit for publication.

Results
374 adults attending the CLASS clinic were diagnosed with Asperger's syndrome between Jan 23, 2004, and July 8, 2013. Informants were available for 355 (95%) participants. The mean age at diagnosis was 31·5 years (range 17-67, SD 10·9) and 87 (23%) participants were in full-time education at the time of assessment. 256 (68%) respondents were men. Seven (2%) participants did not complete the question about suicidal ideation, six (2%) did not complete the question about depression, and nine (2%) did not complete the question about planned or attempted suicide. About a third of respondents had a lifetime experience of self-reported history of depression; about two-thirds had lifetime experience of suicidal ideation; and about a third had planned or attempted suicide (table 1). Lifetime experience of depression (p=0·787), suicidal ideation (p=0·164), and suicide plans or attempts (p=0·06) did not diff er signifi cantly between men and women (table 1). Individuals with a history of depression (table 2) were more likely to report suicidal ideation (OR 4·3, 95% CI 2·4-7·7; p<0·0001) and more likely to report suicide plans or attempts (OR 2·4, 95% CI 1·5-3·8; p<0·0001) than were those without any history of depression. A small number (all <2%) of patients had any comorbid diagnoses (table 3). Individuals who reported suicide plans or attempts had signifi cantly higher Autism Spectrum Quotient scores than those who did not (table 4). Empathy Quotient scores and ages did not diff er between individuals who did or did not report suicide plans or attempts (table 4). Patients with self-reported depression or suicidal ideation did not have signifi cantly higher Autism Spectrum Quotient scores, Empathy Quotient scores, or age than did those without depression or suicidal ideation (table 4).

Discussion
Our fi ndings of a 66% lifetime experience of suicidal ideation and a 35% lifetime experience of planned or attempted suicide supports the assertion that these occurrences are common in people with Asperger's syndrome. In our sample, the lifetime experience of suicidal ideation was more than nine times higher than in the general population in England, 28 and signifi cantly higher than rates previously reported in other clinical groups with medical 29 and psychotic 30 illnesses (panel). People with Asperger's syndrome were signifi cantly more likely to report suicidal ideation or plans or attempts at suicide if they also had depression. Individuals who planned or attempted suicide also had a signifi cantly higher level of self-reported autistic traits than those who did not. Although determination of the direction of causation for this increase in risk is not possible, the fact that more people in this sample reported lifetime   experience of suicidal ideation (66%) than were depressed (31%) is puzzling, and could suggest a diff erent process for suicidal ideation in Asperger's syndrome than for other clinical groups. Alternatively, this fi nding might result from under-reporting of depression, perhaps because of alexithymia (diffi culties verbally describing subjective emotional experience). 32,33 The lifetime experience of suicidal ideation reported in our clinic sample was higher than that noted in a previous small sample of adults with Asperger's syndrome living in the community (40%), 8 and in a small sample of 26 adult psychiatric patients diagnosed with autism spectrum conditions (30·8%). 18 This diff erence might be because individuals in our cohort had not been diagnosed with Asperger's syndrome until late in adulthood, with a mean age at diagnosis of 31 years compared with the usual average age of 11 years. 3 Thus, many of these individuals had diffi culties such as social exclusion, unfulfi lled educational potential, diffi culties with getting or keeping a job or being promoted, and diffi culties with developing close relationships, which could have been exacerbated through lack of appropriate support throughout their lives. Delayed diagnosis in adulthood could possibly be another risk factor for suicidal ideation and plans or attempts in people with Asperger's syndrome.
Our study has several limitations. First, the sample only represents the population of people who reach adulthood without a diagnosis of Asperger's syndrome and who subsequently seek assessment. We do not know if these results can be generalised to the population who have Asperger's syndrome identifi ed in childhood. We also cannot say anything about those adults who do not seek a diagnosis of Asperger's syndrome but who might be struggling, undiagnosed, in the community. Second, comparisons with other studies could not be closely matched for age, sex, or other risk factors. Nonetheless, the scale of the diff erences suggests that a signifi cant diff erence would remain even these factors could be controlled for. The patient questionnaire used in the current study had not been previously validated for use in the current sample, and included both plans and attempts at suicide in the same question. Because lifetime experience of planning and attempting suicide were collapsed into one item of the patient screening questionnaire, disentanglement of these two related but distinct outcomes, or com parison with previously established rates in previous studies (in which these outcomes are recorded separately), was not possible. Future research could explore whether people with Asperger's syndrome are more likely to engage in suicide attempts without planning than are people in the general population or other clinical groups. Lifetime experience of depression, suicidal ideation, and suicide plans or attempts were all measured through a self-report questionnaire completed before the clinic assessment; whether this method led to underestimation or overestimation of these outcomes is not clear. However, all comparisons of suicidal ideation in other clinical groups also included similar self-report measures of the lifetime experience of suicidal ideation, to ensure that group diff erences were not confounded by diff erences in reporting methods.
More detailed studies are needed into the triggers and experience of suicidal ideation, the risk-promoting and protective factors for suicide plans and attempts in adults with Asperger's syndrome (such as age at diagnosis), and family history of suicide and aggression. In addition to the social factors known to predispose to depression, the cognitive profi le of people with Asperger's syndrome might further increase the rate and risk of suicidality. For example, cognitive fl exibility can be impaired, and might mediate some of the increased suicidality. On the basis of our fi ndings, services should be alerted to the high lifetime risk of suicidal ideation and suicide plans or attempts, especially in individuals receiving a late diagnosis of Asperger's syndrome, in view of the substantial risk in this group.

Contributors
SC and PB did the literature search. SB-C conceived of the study and SC, PB, JR, and SB-C designed the study. SC and PB prepared the tables. PB, JR, MM, and SB-C collected the data. SC did the data analysis and interpretation. SC, PB, and SB-C wrote the report, and CA, JR, and MM critically revised the report. All authors approved the fi nal submitted version.

Systematic review
We identifi ed articles by searching Medline, PsycINFO, and Google Scholar in September, 2013, with the general search term "suicid*" in combination with search terms for autism spectrum conditions (eg, "autism", "Asperger", etc). No age or date restrictions were applied to the search.

Interpretation
We identifi ed four studies exploring suicidality in autism spectrum conditions. The few previous studies of suicidality in autism spectrum conditions have shown that these individuals are at high risk of suicidal ideation and behaviours. Only one study was done in adults with Asperger's syndrome, with no comparison sample, so the potential role of risk factors (eg, depression) or how the lifetime experience of suicidal ideation, plans, and attempts in adults with Asperger's syndrome compares with that of other clinical groups is not clear. We studied a unique cohort of individuals diagnosed with Asperger's syndrome as adults to explore their lifetime experience of suicidal ideation and plans or attempts by comparison with other clinical groups. In this fi rst, large-scale clinical study, we noted that adults with confi rmed Asperger's syndrome were signifi cantly more likely to report lifetime experience of suicidal ideation than were individuals from a general UK population sample, or a sample with one, two, or more medical illnesses, or a sample with psychotic illness. Individuals with Asperger's syndrome were also signifi cantly more likely to report suicidal ideation or plans or attempts at suicide if they also had depression. Those who planned or attempted suicide also had a signifi cantly higher level of self-reported autistic traits than those who did not. Our fi ndings support anecdotal reports of increased rates of suicidal ideation in adults with Asperger's syndrome, and identify depression as a signifi cant risk factor for suicidality in these individuals. This study emphasises the need for more detailed studies into the triggers and experience of suicidal ideation, and the risk and protective factors for suicide plans and attempts in adults with Asperger's syndrome, to inform appropriate service planning and support to reduce risk in this clinical group.

Declaration of interests
We declare no competing interests.