Suicidality in Body Dysmorphic Disorder: A Prospective Study

Two hundred subjects were enrolled in this prospective observational study of BDD’s course. The study did not provide treatment. Inclusion criteria were meeting criteria for DSM-IV BDD or its delusional variant, being age 12 or older, and being able to be interviewed in person. The only exclusion criterion was an organic mental disorder. Fifty-two percent of subjects were self-referred; 48% were referred by professionals. An institutional review board approved the study, and the subjects signed statements of informed consent (assent plus parental consent for adolescents). The subjects were interviewed at intake and then annually. Because the subjects were enrolled over 2.4 years, they had varying follow-up durations. Of the 200 enrolled subjects, 185 (92.5%) had a 1-year follow-up interview, 163 (81.5%) had a 2-year interview, 95 (47.5%) had a 3-year interview, and 25 (12.5%) had a 4-year interview.

The Longitudinal Interval Follow-Up Evaluation, a semistructured rating system for assessing the course of mental disorders (7), assessed weekly BDD severity during follow-up with the 7-point Psychiatric Status Rating for BDD; scores of 1 and 2 indicate minimal or no BDD symptoms, 3 and 4 indicate subthreshold symptoms, and 5–7 reflect full BDD criteria. The Longitudinal Interval Follow-Up Evaluation also assessed treatment received and suicide attempts during follow-up, including medical threat and intent. We determined the importance of BDD in precipitating the attempt (in the subject’s view) using a 5-point scale; attempts were considered attributable to BDD if BDD was a moderate, major, or primary reason. We used the semistructured BDD Form (3, 5) to assess suicidal ideation during the year since the previous interview. Suicidal ideation was attributed to BDD if BDD was considered primarily responsible for the ideation. This measure also obtained data on demographics and BDD’s clinical features, including past suicidal ideation and attempts (cross-sectional/retrospective suicidality data from the intake assessment are reported elsewhere [8] ). Means, standard deviations, frequencies, annual weighted means, and 95% confidence intervals (CIs) were computed. Rates are based upon the number of subjects interviewed at follow-up; the rates of completed suicide are based on the number of subjects interviewed or due for an interview. Suicides were determined by searches of the Social Security Death Index, other search procedures, and examination of death certificates.

A total of 126 of 185 subjects (68.1%) were women, 138 (74.6%) were unmarried, 157 (85.8%) were white, and 13 (7.2%) were Hispanic. At intake, the mean age was 33.0 years (SD=12.2), and the mean duration of BDD was 16.0 years (SD=12.5). At intake, 164 (88.6%) of the subjects met full DSM-IV BDD criteria (11.4% were in partial or full remission but had met full BDD criteria in the past).

At intake, 147 of 185 subjects (79.5%) reported a history of suicidal ideation; 51 (27.6%) had a history of a suicide attempt. Table 1 shows prospective suicidality data. The mean score on the Psychiatric Status Rating for BDD over 1–4 years of follow-up was 4.6 (SD=1.4) for the entire group and 4.9 (1.2) for subjects who met full criteria for BDD at intake. A total of 167 (90.3%) of the subjects received mental health treatment at some point during the follow-up period. The nine subjects (four men and five women; age: mean=37.1 years, SD=11.6) who attempted suicide during follow-up made a total of 30 suicide attempts; when we considered all attempts, the maximum level of medical threat was minimal for one subject, moderate for four subjects, severe for one, extreme for one, and resulted in death for two. The maximum level of intent (excluding the suicides) was definite but ambivalent in one, serious in three, very serious in two, and extreme in one.

Two subjects committed suicide (both were receiving psychiatric treatment at the time of their deaths). One was a 31-year-old single white man who was receiving disability pay and had no past hospitalizations or suicide attempts who hanged himself. The other was a 54-year-old single white unemployed man who stabbed himself. He had four past hospitalizations, all primarily for BDD, and one prior suicide attempt, attributed primarily to BDD. At their last interview, both subjects had extremely severe BDD (score of 7 on the Psychiatric Status Rating for BDD), and both considered BDD their primary (most problematic) diagnosis.

Consistent with previous studies of past suicidality (2 – 5), this prospective study found that individuals with BDD had high rates of suicidal ideation and attempts. The mean annual suicidal ideation rate of 57.8% is approximately 10–25 times higher than in the U.S. population, and the mean annual suicide attempt rate of 2.6% is an estimated 3–12 times higher (9, 10) . With adjustment for age, gender, and geographic region, our completed suicide rate was approximately 45 times higher than in the general population (11), that is, the standardized mortality ratio for BDD was approximately 45. Although rates of completed suicide for other disorders vary, depending on the study, and comparisons should be made with caution, examples of estimated standardized mortality ratios for suicide for other mental disorders (based on meta-analyses) are 23 for eating disorders, 20 for major depression, and 15 for bipolar disorder (12) . Thus, our standardized mortality ratio for BDD appears markedly high. This result is preliminary. However, individuals with BDD have many suicide risk factors, including having high rates of psychiatric hospitalization, being single or divorced, and having high comorbidity, poor social supports, poor self-esteem, and high levels of anxiety, depression, and hostility (8) .

Our study’s limitations include the lack of a control group or data directly comparing subjects with BDD to individuals with other disorders, use of different instruments than in studies of other disorders, and conduction of the study in a BDD specialty setting, which may attract more severe cases. On the other hand, our rate of suicide could be an underestimate because our methods may not have detected all suicides. Our group size and limited duration of follow-up limited the stability and precision of our suicidality estimates. Studies are needed over longer follow-up periods and in other BDD samples and settings.

Received March 18, 2005; revision received May 13, 2005; accepted Sept. 19, 2005. From Butler Hospital, Brown Medical School. Address correspondence and reprint requests to Dr. Phillips, Butler Hospital, 345 Blackstone Blvd., Providence, RI 02906; [email protected] (e-mail).Supported by NIMH grant R01-MH-60241 to Dr. Phillips.