Six lithium dose predictors for patients with bipolar disorder

Summary: Researchers have identified six predictors that can help determine the appropriate amount of lithium to treat a patient with bipolar disorder.

Source: Karolinska Institute

Six predictors may help determine the amount of lithium needed to treat patients with bipolar disorder, according to a large study led by researchers at the Karolinska Institutet in Sweden.

The study, published in the journal The Lancet Psychiatryalso identifies genetic markers that seem to influence how quickly the body eliminates lithium from its system.

“Our model could already be used to predict how much lithium a patient with bipolar disorder will need. This could save precious time spent finding the right dose for each patient, potentially with life-saving consequences,” said Martin Schalling, professor in the Department of Molecular Medicine and Surgery, Karolinska Institutet and senior author of the study.

Lithium is one of the main treatments for patients with bipolar disorder, a condition associated with an increased risk of suicide.

The chemical acts as a mood stabilizer and reduces episodes of depression and mania. How much is needed varies widely from person to person and finding the right dose for each patient is key, as too much can be toxic, while too little is ineffective.

To minimize the risk of side effects, clinicians tend to start treatment with low doses that increase over time, meaning it can take months for the treatment to take effect.

To overcome this, researchers have long sought a model that could predict dose response in individual patients.

Previous studies have identified markers such as age, gender and kidney function as possible predictors of how quickly the body clears lithium from its system (lithium clearance), which can be used to determine the daily amount needed. However, most studies are limited by small sample sizes.

In the current study, the researchers examined electronic health records and registry data from a total of 2,357 patients with bipolar disorder, which may be the largest sample size for this type of study to date. Both men and women ranging in age from 17 to 89 were included, mainly of European descent.

Lithium is one of the main treatments for patients with bipolar disorder, a condition associated with an increased risk of suicide. Image is in the public domain

The study found associations between the rate of lithium clearance and age, gender, renal function (measured as eGFR), serum lithium concentrations, and medications containing diuretics and agents that target the renin-angiotensin-aldosterone system (RAAS), which can be used to treat hypertension and other conditions.

“Our findings suggest that elderly patients, women, patients with impaired renal function, and those taking certain medications require lower doses of lithium. Interestingly, we also found that the amount of lithium ingested and the lithium concentrations in the blood are not completely proportional which is somewhat contrary to current views.

“Our model based on these predictors explained about 50-60% of the variance in lithium clearance, which is better than previous models and could be used to inform treatment decision,” said first author Vincent Millischer, a postdoctoral researcher in the department. Molecular Medicine and Surgery, Karolinska Institutet, and resident in psychiatry at the Medical University of Vienna.

The study also found associations between lower lithium clearance and one genetic locus on chromosome 11 and was also able to show that genetic variants affecting BMI and kidney function were associated with lithium clearance.

While adding the genetic markers only marginally improved the model’s predictive ability, the researchers say it opens up the possibility for personalized medicine in lithium treatment in the future.

“We will then test our model in a clinical trial to see if it can shorten the time it takes to find the right amount of lithium for each patient,” says Martin Schalling.

“If the outcome is positive, we will develop a digital app that can be used by psychiatrists in the future to help assess lithium dosing for patients with bipolar disorder.”

About this research news on psychopharmacology

Writer: press office
Source: Karolinska Institute
Contact: Press Information – Karolinska Institute
Image: The image is in the public domain

Original research: Closed access.
“Improving lithium dose prediction using population pharmacokinetics and pharmacogenomics: a cohort genome-wide association study in Sweden” by Martin Schalling et al. The Lancet Psychiatry


Improving lithium dose prediction using population pharmacokinetics and pharmacogenomics: a cohort genome-wide association study in Sweden


Lithium is the most effective treatment for bipolar disorder, resulting in strong suicide prevention effects. However, the therapeutic range of lithium is narrow and initiation of treatment requires individual titration to address inter-individual variability. We wanted to improve lithium dose prediction using clinical and genomic data.


We performed a population pharmacokinetic study followed by a genome-wide association study (GWAS), including two clinical Swedish cohorts. The participants in cohort 1 were from specialist outpatient clinics at Huddinge Hospital in Stockholm, Sweden, and participants in cohort 2 were identified using the Swedish National Quality Register for Bipolar Disorder (BipoläR). Patients who received a lithium dose equivalent to at least one tablet of lithium sulfate (6 mmol) per day and had clinically relevant plasma concentrations of lithium were enrolled.

Also see

This shows astrocytes

Data on age, sex, body weight, height, creatinine concentration, estimated glomerular filtration rate (eGFR), lithium preparation, number of lithium tablets per day, serum lithium concentration and drugs affecting renal function (C09 antihypertensives, C03 [except C03D] sodium-retaining diuretics and non-steroidal anti-inflammatory drugs) were obtained retrospectively for different time points, if possible, from electronic health records, BipoläR and the Swedish Prescription Registry.

The median time between time points was 1.07 years for cohort 1 and 1.09 years for cohort 2. The primary outcome of interest was the natural logarithm of total body clearance for lithium (CLLi) associated with the clinical variables.

The residual effects after taking into account age and gender, which represent the effects at the individual level (CLLi,age/gender), were used as the dependent variable in a GWAS.


2357 patients who received lithium (1423 women [60·4%] and 934 men [39·6%]† average age 53.6 years [range 17–89], mainly of European descent) were included and 5627 data points were obtained. Age (variance explained) [R2]† R2cohort1=0·41 and R2cohort2=0·31; both p<0.0001), gender (R2cohort1=0.0063 [p=0·045] and R2cohort2=0.026 [p<0·0001]), eGFR (R2cohort1=0 38 and R2cohort2=0 20; both p<0.0001), co-medication with diuretics (R2cohort1=0.0058 [p=0·014] and R2cohort2=0.0026 [p<0·0001]), and agents that act on the renin-aldosterone-angiotensin system (R2cohort1=0.028 and R2cohort2=0.015; both p<0.0001) were clinical predictors of CLLi

In particular, a link between CLLi and serum lithium was observed, with a lower CLLi associated with higher serum lithium (R2cohort1=0 13 and R2cohort2=0 15; both p<0.0001). In a GWAS from CLLi,age/genderone locus was associated with a change in CLLi (rs583503; β = –0 053 [95% CI –0·071 to –0·034]† p<0 00000005).

We also found an enrichment of the associations with genes expressed in the medulla (p=0.0014, corrected FDR=0.04) and renal cortex (p=0.0015, corrected FDR=0.04 ), as well as associations with polygenic risk scores for eGFR (p-value threshold: 0.05, p=0.01), body-mass index (p-value threshold: 0.05, p=0.00025), and blood urea nitrogen (p-value threshold: 0.001, p=0.00043). The model based on six clinical predictors explained 61.4% of the variance in CLLi in cohort 1 and 49.8% in cohort 2.

Adding genetic markers did not significantly improve the models: within the subsample of genotyped individuals, the explained variance only increased from 59.32% to 59.36% in cohort 1 and from 49.21% to 50.03 % in cohort 2 when including rs583503 and the first four principal components.


Our model predictors can be used clinically to better direct lithium dosing, reducing the time to reach therapeutic concentrations, improving care. Identification of the first genomic locus and PRS associated with CLLi introduces the possibility of individualized medicine in lithium treatment.


Stanley Medical Research Institute, Swedish Research Council, Swedish Foundation for Strategic Research, Swedish Brain Foundation, Swedish Research Council, Söderström-Königska Foundation, Bror Gadelius Minnesfond, Swedish Mental Health Fund, Karolinska Institutet and Hospital.

Leave a Comment