May 24, 2023

Association of Bipolar Disorder Diagnosis With Suicide Mortality Rates in Adolescents in Sweden

Peter Andersson, MSc1,2Jussi Jokinen, MD, PhD3,4Håkan Jarbin, MD, PhD5,6et alJohan Lundberg, MD, PhD3,7Adrian E. Desai Boström, MD, PhD3,4,7
Author Affiliations Article Information
JAMA Psychiatry. Published online May 24, 2023. doi:10.1001/jamapsychiatry.2023.1390
Key Points

Question  Is diagnosing bipolar spectrum disorder in adolescents associated with suicide prevention?

Findings  In this cross-sectional study of 585 confirmed suicide deaths in Sweden, regional bipolar disorder diagnosis rates in adolescent males were associated with a lower suicide death rate at an estimated magnitude of approximately 4.7% of the national average. Independent of annual regional depression and schizophrenia diagnoses, lithium dispensation, and psychiatric care affiliation rates, results were consistent with previous reports suggesting that bipolar disorder is implicated in approximately 4.9% of unselected suicide deaths in young adulthood.

Meaning  The findings indicate that diagnosis of bipolar disorder in male adolescents may be important for suicide prevention.


Importance  The association of early diagnosis and management of bipolar disorder with adolescent suicide mortality (ASM) is unknown.

Objective  To assess regional associations between ASM and bipolar disorder diagnosis frequencies.

Design, Setting, and Participants  This cross-sectional study investigated the association between annual regional ASM and bipolar disorder diagnosis rates in Swedish adolescents aged 15 to 19 years in January 1, 2008, through December 31, 2021. Aggregated data without exclusions reported at the regional level encompassed 585 suicide deaths, constituting 588 unique observations (ie, 21 regions, 14 years, 2 sexes).

Exposures  Bipolar disorder diagnosis frequencies and lithium dispensation rates were designated as fixed-effects variables (interaction term in the case of males). An interaction term between psychiatric care affiliation rates and the proportion of psychiatric visits to inpatient and outpatient clinics constituted independent fixed-effects variables. Region and year comprised random intercept effect modifiers. Variables were population adjusted and corrected for heterogeneity in reporting standards.

Main Outcomes and Measures  The main outcomes were sex-stratified, regional, and annual ASM rates in adolescents aged 15 to 19 years per 100 000 inhabitants as analyzed using generalized linear mixed-effects models.

Results  Female adolescents were diagnosed with bipolar disorder almost 3 times more often than male adolescents (mean [SD], 149.0 [19.6] vs 55.3 [6.1] per 100 000 inhabitants, respectively). Median regional prevalence rates of bipolar disorder varied over the national median by a factor of 0.46 to 2.61 and 0.00 to 1.82 in females and males, respectively. Bipolar disorder diagnosis rates were inversely associated with male ASM (β = −0.00429; SE, 0.002; 95% CI, −0.0081 to −0.0004; P = .03) independent of lithium treatment and psychiatric care affiliation rates. This association was replicated by β-binomial models of a dichotomized quartile 4 ASM variable (odds ratio, 0.630; 95% CI, 0.457-0.869; P = .005), and both models were robust after adjusting for annual regional diagnosis rates of major depressive disorder and schizophrenia. No such association was observed in females.

Conclusions and Relevance  In this cross-sectional study, lower suicide death rates in adolescent males was robustly associated with regional diagnosis rates of bipolar disorder at an estimated magnitude of approximately 4.7% of the mean national suicide death rate. The associations could be due to treatment efficacy, early diagnosis and management, or other factors not accounted for.


Bipolar spectrum disorder is a chronic severe mental illness characterized by recurring episodes of elevated mood and depression with a general peak age of onset at 12 to 25 years.1,2 Long delays have been observed from disease onset to diagnosis and initiation of treatment. A recently published meta-analysis found median times from onset to help seeking, diagnosis, and initiation of mood-stabilizing pharmacotherapy of 3.5, 6.7, and 5.9 years, respectively.3 Factors influencing latencies were sex, treatment setting, aspects of disease presentation, and geographic location, with shorter delays to diagnosis observed in Europe and Scandinavia vs North America.3The diagnosis of bipolar disorder in pediatric populations is a widely discussed topic in the literature, particularly due to conflicting perceptions of prevalence and treatment, as well as the potential role of irritability.4 The National Comorbidity Survey Replication has estimated that bipolar disorder affects 2.8% of US adults and 2.9% of US adolescents, with higher prevalence among adolescent females (3.3%) vs adolescent males (2.6%), but largely equal prevalence rates across the sexes in adults.5,6 Van Meter et al7 reported a weighted average prevalence of youth bipolar disorder to be 3.9% (95% CI, 2.6%-5.8%). The authors suggested that heterogenous prevalence rates could be attributed to nonstandard diagnostic criteria and narrow definitions. Their study also indicated that prevalence rates in the US are not higher than in other Western countries and that there is no increasing trend over time. Moreover, in a review of the evidence base, Findling et al4 found consensus favoring diagnosis according to DSM-5 criteria, consistent with the report by Perlis et al,8 that a majority of adult patients with bipolar disorder presented with onset of mood symptoms in youth (aged <18 years) and that early onset was associated with more recurrences, greater rates of substance abuse, and a greater likelihood of suicide attempts and violence. Emerging evidence supports psychological and pharmacologic treatments in early disease stages to improve outcomes.9,10 Nevertheless, practices of diagnosing bipolar disorder in youth have been controversial, and issues related to a perceived lack of awareness, diagnostic confusion, stigma, and other factors may have negatively affected diagnosis rates.11Accordingly, international clinical guidelines communicate incongruent recommendations for the diagnosis of youth bipolar disorder. For example, National Institute for Health and Care Excellence guidelines posit substantially stricter criteria for diagnosing bipolar disorder in adolescents vs adults (eg, requiring the presence of mania),12 whereas the American Academy of Child and Adolescent Psychiatry recommends adhering to traditional DSM criteria.13

Epidemiologic reviews have suggested that 44% to 76% of adolescents who die by suicide meet criteria for an affective disorder.14 The exact contribution of bipolar disorder to suicide deaths in adolescence is not fully elucidated. Clements et al,15 however, reviewed registered suicides in a national English sample during 1996 to 2009, and found that 4.9% (57 of 1163) of suicides in young adulthood (aged ≤24 years) had a primary diagnosis of bipolar disorder. Lithium treatment has shown greater efficacy in reducing suicide attempts in youth with bipolar disorder compared with other mood-stabilizing treatments.16 However, the exact ratio of attempted suicides to suicide deaths in adolescents with bipolar disorder remains unclear and is estimated to be between 50:1 and 100:1 in the general adolescent population.17This uncertainty complicates the interpretation of the potential suicide-protective effects of lithium in this specific population.

The aim of our study was 2-fold. First, we hypothesized that the historical lack of solid evidence in support of practices for diagnosing bipolar disorder in youth may have contributed to health inequities regarding diagnosis and treatment rates at the regional level. Our goal was to comprehensively examine and compare the national prevalence of bipolar disorder in its various forms (bipolar 1, 2, and not otherwise specified) across the 21 regions of Sweden during the years 2008 to 2021 in individuals aged 15 to 19 years. Additionally, we aimed to investigate the potential association between bipolar disorder diagnosis frequencies and the total regional dispensation of lithium treatment within this age group. Second, we hypothesized that regional diagnosis rates of bipolar disorder and treatment with lithium are associated with reduced suicide death rates in adolescents, ie, that early diagnosis (and management) of bipolar disorder and/or lithium treatment would bestow reduced adolescent suicide mortality (ASM) at the regional level.


Study Design and Patients

In this cross-sectional study of sex-stratified suicide death rates, bipolar disorder diagnosis frequencies, and lithium dispensation rates in youth aged 15 to 19 years, data were retrieved for the 21 Swedish regions from January 1, 2008, through December 31, 2021, from the Swedish National Board of Health and Welfare,18 a freely available Swedish data set.19-21 Extracted data included registered bipolar disorder diagnosis frequencies (ICD-10code F31) in both specialized outpatient and inpatient care and confirmed suicide death rates (codes X60-X84) per 100 000 inhabitants, as well as the number of dispensations to adolescents recorded for lithium (Anatomical Therapeutic Chemical code N05AN01) per 1000 inhabitants in each region and age group, respectively. This work was conducted in accordance with the ethical standards of the Declaration of Helsinki and in accordance with Swedish laws on research ethics. As the study pertained to openly available data, no ethical permission or informed consent for publication were required by Swedish jurisdiction. Karolinska Institutet regulatory standards were, however, followed. The study followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline.22Sources for additional control variables and initial processing steps are detailed in the eMethods in Supplement 1.


FYI: Joey Eckenfels