Key points

Overview

Bipolar disorder (BD) is a common mental health disorder affecting approximately 2% of people worldwide regardless of sex, ethnicity, or social income. ^, People suffering from BD experience periods of high mood (mania or hypomania) and may also experience episodes of depression. BD is split into 2 main types, bipolar I and bipolar II. Bipolar I disorder (BDI) is defined by manic episodes that last for at least 7 days or are severe enough to require immediate hospital care. Manic episodes are characterized by elevated, expansive, or irritable mood, increased energy levels, racing thoughts, decreased need for sleep, grandiosity, impulsivity, and sometimes psychosis (delusions or hallucinations). Bipolar II disorder (BDII) is characterized by episodes of hypomania. Hypomania is less severe than mania and involves a distinct period of elevated or irritable mood and increased energy, but it does not typically cause severe impairment in functioning or require hospitalization. Depressive episodes in BDI and BDII typically accompany manic or hypomanic episodes or may alternate with them. Some individuals with a BD diagnosis undergo rapid cycling of their mood symptoms defined as having 4 or more distinct episodes within a 12-month period. BD, with a median age of onset at age 33 year old, ranks among the most debilitating illnesses especially affecting young and working-age adults resulting in decreased productivity, social impairments, and high levels of somatic and psychiatric comorbidities. For example, the odds ratio for alcohol use disorders in BD has been estimated at 4.09 (95% confidence interval [CI] 3.37 to 4.96) and the relative risk (RR) of all-cause mortality rates in people with BD is RR = 2.02 (95% CI: 1.89–2.16). The cause of the increased mortality rate is not limited to unnatural causes (eg, suicide) but also involves infectious diseases, as well as cardiovascular and cerebrovascular disorders. The high rate of death by suicide is considerably reduced when BD patients are successfully treated (particularly with lithium).

Individuals with BD typically encounter their initial (hypo) manic or depressive episodes during adolescence or early adulthood. However, diagnosis often occurs with a significant delay, ranging from 5 to 10 years later, particularly for those with an earlier onset or those with an initial depressive episode. Yet, intervention in early illness course of BD is associated with lower recurrence risk and better global functioning (especially with lithium).

Both environmental and genetic factors contribute substantially to the etiology of BD.

Understanding the genetic basis of the disorder and identifying specific genetic risk factors come with great promises for patients and all those involved in their personal and clinical care: (i) genetic risk is linked to biologic factors that may provide opportunity for improved and more accurate diagnosis and therefore earlier intervention (with currently available treatment options) resulting in better outcomes for patients ; (ii) new genetic discoveries might lead to identification of novel disease mechanisms that can be leveraged for the development of novel therapeutic targets and drug development; (iii) genetic risk assessments of BD can also be used for epidemiologic studies to better understand how BD disease risk may impact health, behavior, and disease in the general population from sociologic, as well as a biologic perspective; and (iv) lastly, genetic studies leading to an improved understanding of the disorder within society can help with the destigmatization of BD by highlighting its biologic underpinnings thereby promoting an improved understanding of the disorder within society.

Heritability and molecular genetics of bipolar disorder

Traditional twin study-based estimates for the heritability of BD range from 60% to 85%. ^, These studies also showed a limited role for the shared family environment (for example parenting style) in the risk of BD. A recent population-based family study from Sweden demonstrated a strong familial component to psychosis in BD that was stronger in BD (Hazard ratio [HR] 4.49 [95% CI: 3.21–6.29]) compared to psychosis in psychotic major depression (HR 2.98 [95% CI: 2.38–3.72]) or in schizophrenia (HR 2.33 [95% CI: 1.78–3.06]).

Despite the high levels of heritability of BD, the concordance rates of BD in monozygotic co-twins, who share close to 100% of their genomes, is 40% to 70% indicating the important role of an environmental component to disease risk. The rate of BD in a first degree relative of someone with BD is 5% to 10%. Thus, if one has a first-degree relative with BD the likelihood of NOT developing BD goes from the population average of 98% to between 95% and 90%.

Some of the first genome-wide association studies (GWAS) studies were pivotal in expanding our understanding of the shared genetics between BD and schizophrenia. Purcell and colleagues, in one of the first applications of the utility of polygenic risk score (PRS) approaches in human disease, demonstrated that there is a substantial overlap in genetic risk between the two disorders. Whilst the sample sizes available at that time were small by modern standards, the PRS for schizophrenia explained approximately 2.3% to 3.2% of the variance in independent schizophrenia samples of European ancestry and 1.5% to 2% of the variance in BD samples of European ancestry. The variance explained in a schizophrenia sample of African American ancestry was only 0.4% (see section on the use of PRSs in BD). It is important to note that the initial GWAS studies of BD primarily involved cases with severe bipolar disorder (BDI) that were diagnosed using clinically rated and validated instruments meeting the standard classification of mental disorders. The relatively high proportion of genetic risk shared between BD and schizophrenia led to further discussion in the field about the validity of the more than a century-old Kraepelinian dichotomy, which assumes that BD and schizophrenia are distinct entities with separate etiologies.

At the same time, the earliest significant findings of the genome-wide studies of psychiatric disorders also highlighted shared genetic risk loci across diagnostic boundaries; (a locus is a specific physical location in the human genome). For example, the locus containing the L-type voltage-gated calcium channel gene CACNA1C was first reported to be associated with risk of BD but was then also found to be associated with schizophrenia and major depression. Subsequent to the initial GWAS studies of a few thousand subjects, and through the establishing of the Psychiatric Genomics Consortium ( initially named Psychiatric GWAS Consortium ) (PGC), much larger genome-wide studies of BD have been completed. With each substantial increase in sample size, an increasingly large number of loci have been identified to be significantly associated with BD. First, Stahl and colleagues 2019 identified 30 independent genome-wide significant (GWS) findings with BD in a discovery sample of 20,352 cases and 31,358 controls of European descent. More recently, Mullins and colleagues (2021) extended this work in a meta-analysis of data from 41,917 BD cases and 371,549 controls and identified 64 independent susceptibility loci that included 33 novel loci. These large-scale studies have identified many risk loci and provided significant evidence of enrichment of voltage-gated calcium channel genes, with brain-expressed genes, and with genes involved in synaptic signaling, in the regulation of insulin secretion and endocannabinoid signaling, all to be involved in BD susceptibility. The GWAS studies clearly showed the important contribution of genetic risk to BD and revealed the extensive polygenic nature of the genetic architecture of severe mental illness. The genome-wide assessment of genetic risk for BD was also used to calculate PRS, which in turn is used to examine the overlap of genetic risk between psychiatric disorders and neurobehavioral traits. It was shown that the PRS for schizophrenia was higher in BDI subjects and that the PRS for major depression was higher in BDII subjects. Furthermore, the genetic analysis between BDI and BDII revealed a strong but incomplete genetic correlation structure (rG = 0.85) indicating the presence of distinct contributing biologic factors between the 2 subtypes of the same disorder. The larger GWAS study with more power for discovery has enabled the researchers to point beyond the general assessment of brain expressed genes involved in BD susceptibility, to more specifically highlighting significant involvement of excitatory and inhibitory neurons in the cortical and subcortical regions, in hippocampal pyramidal neurons and in interneurons of the prefrontal cortex.

Large-scale genetic study of intermediate phenotypes of BD is in its early days but comes with the promise to better understand the biologic factors contributing to the heterogeneity of the disorder and examining the genetic basis of shared and disorder-specific symptoms between severe mental illnesses such as schizophrenia and BD. For example, PRS analyses in a combined sample of BD and schizophrenia identified several significant correlations within case-only phenotypes, including schizophrenia PRS with psychotic features and age of onset in BD. These and other results point to the utility of genetics to inform symptomology and potential treatment.

As part of the BD working group of the PGC, a recent effort was made to examine the age at onset and polarity of onset of BD. The latter measure considers whether BD is manifested first through depressive or manic episodes. The study showed that individuals with an earlier onset are characterized by an increased polygenic liability for a broad spectrum of psychiatric traits but also revealed that there are systematic differences in age at onset across cohorts, continents, and phenotype definitions, introducing significant heterogeneity in heritability estimates and thereby negatively affecting statistical power for discovery.

The need of increasing sample sizes for GWAS to uncover the genetic architecture of BD more fully, has also led to the inclusion of self-reported and biobank-identified cases in the most recent efforts of the BD working group of the PGC. The advancement of much larger and more ancestry-diverse inclusion numbers comes with increased power for discovery, but it also results in major challenges regarding a much broader (spectrum) case definition compared to previous GWAS studies of BD. The BD working group of the PGC is reporting (in a not yet peer-reviewed paper) results from a much larger; multi-ancestry GWAS meta-analysis of 156 thousand BD spectrum cases; and almost 2.8 million controls, combining clinical, biobank, and self-report samples. The number of BD susceptibility loci increased to 298 with 337 independent genome-wide significant variants. Perhaps not altogether surprisingly, the genetic correlations between the self-report and biobank derived samples are stronger with BDII compared to BDI. The single nucleotide polymorphism (SNP)-heritability on the liability scale assuming a BD lifetime prevalence of 2% for BDI was 0.25, 0.22 for clinically ascertained BD, 0.05 for biobank ascertained BD, and 0.08 for self-report BD. From the perspective of BD GWAS studies for the identification of common risk alleles, it may be that this massive increase of sample size (leading to the major increase of known disease loci) is counterbalanced by decreased phenotypic precision and, as a consequence, discovery of disease biology outside the immediate bounds of BD. Furthermore, despite the increased ancestry-diverse sample in this study, the vast majority of cases is still of European background. There remains an impetus to significantly increase non-European ancestry cohorts for a better understanding of the genetic risk of BD across the globe and discovery and refinement of disease mechanisms. Major efforts are underway to collect and analyze large case/control BP cohorts in Latin-America, Africa, and Asia. It is expected that these efforts will serve as the next frontier in BD research for the discovery of the underlying genetic susceptibility residing in common alleles in the populations of the world. There currently remains, however, a gap, between the estimated broad-sense heritability of BD ( h ² ∼65%) and the heritability explained by common alleles (with minor allele frequency of >1%) (h ² _snp ∼ 20%). This difference suggests that other sources of genetic variation are likely to contribute to disease risk, including copy number variants (CNVs), and rare (oftentimes deleterious) genetic variants.

Involvement of copy number variants

Rare sub-chromosomal deletions and duplications, known as CNVs are important risk factors for schizophrenia and autism. This is not so much the case for BD where the role of CNVs is less pronounced. However, some of the same CNVs that confer risk for schizophrenia, namely 1q21.1 duplications, 3q29 deletions, and 16p11.2 duplications do confer risk for BD with odds ratios of 2.64 (95% CI: 1.19–5.88), 17.31 (95% CI: 1.57–190.97), and 4.37 (95% CI: 2.12–9.00), respectively. A later study reported that CNV burden was restricted to schizoaffective BD cases and that this increased burden was not seen in BDI cases regardless of the presence of psychosis.

Rare single nucleotide variants in bipolar disorder

The advent of whole exome and whole genome sequencing has allowed the reliable detection of rare variants that are likely to confer a high level of risk for a particular trait. If negative selection on BD (or any other trait) exists, it is expected that underlying high-risk alleles are present at very low frequency in the population and that disease-causing, ultra-rare variants are mostly affecting the protein-encoding regions of the genome. It has been shown that BD risk is associated with a burden of rare and ultra-rare protein truncating variants (PTVs) in genes that are generally intolerant of deleterious variants (OR = 1.2, P =8 × 10 ⁻⁴ ). The largest exome sequencing study of BD (thus far) was performed by Palmer and colleagues (2022) and included almost 14,000 cases and a similar number of control subjects. The study showed an enrichment of ultra-rare PTVs within genes previously implicated in schizophrenia. However, no single BD disease gene was identified through this effort, except when schizophrenia results (based on a schizophrenia cohort of 24K patients and 97K controls) and BD were combined, resulting in the identification of AKAP11 as a definitive risk gene (OR = 7.06, P = 2.83 × 10 ⁻⁹ ). ^, It is hypothesized that AKAP-11 interacts with GSK3B , the hypothesized target of lithium, the most prescribed drug worldwide for the treatment of BD. Unpublished results from the BD PGC working group showed that genes implicated in the most recent GWAS of BD are significantly enriched for ultra-rare PTVs (OR = 1.16, 95% CI = 1.05–1.28, P =.002). These results show that rare coding variants do contribute to BD disease risk with growing evidence of convergence of common and rare variant signals in the human genome. We expect that the increased sequencing efforts of large BD case/control cohorts in diverse ancestry populations will provide one of the biggest advances in the near future, which will shed more light on the contribution of rare coding variants to the etiology of BD.

Polygenic risk scores in predicting bipolar disorder risk: challenges and limitations in clinical applications

PRS approaches have been used to summarize the common genetic risk for a trait into a single score. They work best at predicting risk at the population level compared to at the level of the individual. As the training data (GWAS results from large-scale studies) that is used to generate a PRS increases in size it is expected that the PRS will prove better at predicting disease outcome. Important considerations for the use of PRS include the ancestry matching between the training and test data, or at least the incorporation of data from multiple ancestries. Most GWAS are biased toward individuals of European ancestry, and this therefore has the likelihood of exacerbating existing health inequalities. Assuming a lifetime disease prevalence of 2% for BD the PRS from the latest PGC BD working group GWAS analysis explains 8.4% of the variance on the liability scale trained using clinically or biobank ascertained BD subjects of European ancestry PRS on European ancestry samples. The inclusion of the community-reported BD subjects reduced this to 7.0%. Training the PRS on BD subjects of all ancestries without the community-reported BD subjects improved the variance explained to 9.0%, however. The variance explained by the multi-ancestry PRS with no self-report data in the African American cohort was 1.0% and 2.3% with self-report. The results from East Asian cohorts were somewhat intermediate to the results from the African American cohort and together the results are indicative of the importance of including people of multiple ancestries in future genomic studies of BD.

Genetic counselling for bipolar disorder

Genetic counseling in BD can have an important role to play in reproductive and other life event planning. The evidence suggests that genetic counseling in a psychiatric context given in the absence of a genetic test has a measurable impact on well-being of people with a diagnosis of BD and their family. As discussed earlier, the genetic risk factors and the variance explained by current PRS do not allow clinically relevant prediction of BD. Common genetic risk factors and PRS are not deterministic for BD and have therefore limited diagnostic relevance. It needs to be seen how ultra-rare PTVs leading to BD (identified via large-scale sequencing efforts) will impact genetic counseling and clinical care of BD.

Environmental risk factors

A range of environmental risk factors, such as adverse childhood events, stress, and migration have been shown to influence the risk of developing BD. Adverse events in childhood (ACE) are an established risk factor for BD and are associated with a 2.63 times greater risk of developing BD. ACEs are also associated with an increased risk of relapse in BD and with increased severity of BD symptoms. ^, Chronic stress-induced exhaustion disorder and depression have also been shown to be risk factors for a constellation of related affective disorder diagnoses that included manic episodes, bipolar affective disorder, and persistent mood disorder with an OR of 6.80 (95% CI: 5.32–8.69). Refugee and migration status are risk factors for psychotic affective disorder that includes psychotic depression, as well as BD with psychosis (HR 2.07; 95% CI 1.55–2.78 and HR 1.40; 95% CI: 1.16–1.68, respectively). Interestingly, the rate of non-psychotic BD was found to be lower in refugees (HR 0.24; 95% CI: 0.16–0.38) and in non-refugee migrants (HR 0.34; 95% CI: 0.28–0.41).

Pharmacotherapy

There are substantial differences in the patterns of prescribing treatments for BD across the world. A recent study by the Global Bipolar Cohort collaborative network found that lithium is prescribed at a greater frequency in Europe and Australia compared to North America, with an overall treatment rate of approximately 34% in BDI and 23% in BDII subjects. The same study also found evidence for heterogeneity in the prescribing practices for benzodiazepines in BDI; for first-generation antipsychotics in BDI and BDII (higher in Europe); for second-generation antipsychotics in BDI and BDII (lower in North America); and for mood stabilizing anticonvulsants in BDII (higher in Europe).

Pharmacogenetic-guided pharmacotherapy for bipolar disorder

Serious dermatologic reactions, that include Toxic Epidermal Necrolysis (TEN) And Stevens-Johnson Syndrome (SJS), have been reported in people treated with the mood stabilizing anticonvulsant carbamazepine. The risk is associated with being a carrier of the HLA-B∗1502 allele. The prevalence of this allele differs across global populations being largely absent in people of European, African, and Native American ancestry but ranges from 2% in people of South Asian ancestry to approximately 15% in people from Hong Kong. Another HLA allele HLA-A∗3101 is also associated with TEN and SJS. The frequency of HLA-A∗3101 also differs across populations ranging from 2% to 5% in Northern European populations to 15% in people of Japanese, Native American, and Southern Indian ancestry.

Valproate (valproic acid, sodium valproate, and semi-sodium valproate) is an anti-epileptic that is used as a mood stabiliser in BD. The use of valproate is contraindicated in people who carry POLG mutations. The POLG gene encodes the catalytic subunit of mitochondrial DNA polymerase. Mutations in POLG cause a range of heterogeneous mitochondrial related neurologic diseases including mitochondrial epilepsy, polyneuropathy, ataxia, and progressive external ophthalmoplegia. Carriers of POLG mutations who are treated with valproate are at increased risk of liver failure.

Lithium is a highly effective mood stabilizer not only controlling mania and hypomania symptoms and reducing the severity of depression. Lithium has also been shown to prevent suicides and suicide attempts in BD. ^, Despite this high level of efficacy approximately 30% of BD subjects are excellent responders and 40% do not benefit from treatment and suffer from side-effects. Several studies have attempted to identify genetic factors and risk profiles that may influence the treatment efficacy of lithium. There have been inconsistent findings with several lithium GWAS studies, some with controversial results, suggesting that imprecise and heterogeneous phenotype measurements, as well as the polygenic architecture of lithium response heterogeneity, require a more systematic approach and much larger studies. More recently, a meta-analysis of GWAS data from the largest lithium studies with an additional new cohort found evidence to implicate the ADCY1 (Adenylate Cyclase 1) gene in lithium response (Mcquillin A., Yao K., Van Der Veen T., et al., Implication of the ADCY1 gene in lithium response in bipolar disorder by genome-wide association meta-analysis. 2024. Available at: https://doi.org/10.21203/RS.3.RS-4000581/V1, under review). Adenylate cyclase is an important component in calcium signaling in the brain; it has been implicated in learning and development. Despite early efforts, it remains to be seen if genetic risk of BD (or any other severe mental illness) is associated with the genetic basis of lithium response.

Future perspective

Genetic and genomic studies of BD have improved our understanding of disease risk and etiology. It also has provided new insights into the shared risk between BD and other psychiatric disorders. However, the genetic findings to date are not clinically actionable.

Future studies of BD with increased ancestral diversity will allow the development of polygenic risk scores that are transferrable between individuals of different ancestries and have the potential to uncover additional disease risk mechanisms. The shift toward large-scale genomic sequencing for data collection in BD studies will especially improve knowledge of high-impact, rare variant burden of BD, and has the potential to highlight specific disease genes that may reveal novel drug treatment targets.

Future work aimed at increasing the availability of systematically assessed data on intermediate and clinical sub phenotypes of BD, as well as a renewed effort to systematically collect and analyze (quantitative) pharmacologic data (such as response to specific treatments), will help to improve our understanding of the genetic basis for BD symptoms and bring advancements to currently available treatment options and outcomes.

Clinics care points

• •
  

  Genetics plays a major role in the risk BD.

  
  
• •
  

  Genetic risk factors are not deterministic for BD.

  
  
• •
  

  Genomic studies gave identified hundreds of common disease risk loci.

  
  
• •
  

  Fewer rare disease risk loci have been identified.

  
  
• •
  

  Pharmacogenomic testing in individuals from high-risk populations is recommended prior to prescription of carbamazepine and valproate.

  
  
• •
  

  Polygenic risk scores for BD are not currently clinically useful for diagnosis or treatment of BD.

Only gold members can continue reading. Log In or Register to continue

Share this:

• Click to share on Twitter (Opens in new window)
• Click to share on Facebook (Opens in new window)

Related

Related posts:

Psychiatric Genomics 2025 Tourette Syndrome and Chronic Tic Disorders Genetics of Anorexia Nervosa Recent Advances in Schizophrenia Genomics and Emerging Clinical Implications Mainstreaming Diagnostic Genetic Testing and Precision Medicine for Autism Spectrum Disorder Genomics of Suicidal Behaviors

Stay updated, free articles. Join our Telegram channel

Join