Results Summary
What was the research about?
Bipolar disorder is a chronic mental illness that causes shifts in mood, energy, and activity. People with bipolar disorder can have high and low moods, known as mania and depression, which differ from usual ups and downs. The disorder may also affect people’s ability to think clearly. They may see or hear things that aren’t real. Severe episodes may require a hospital stay.
Many medicines can treat bipolar disorder. Three types are mood stabilizers, such as lithium and some anticonvulsants; antipsychotics; and antidepressants. Doctors often prescribe two or more medicines together. But these medicines may have negative effects on physical or mental health.
In this set of studies, the research team compared the most used medicines or medicine combinations for bipolar disorder. The team looked at the risk of four health outcomes:
- Hospital stays for mental health concerns, comparing time periods on lithium alone versus other medicines for 852,063 people
- Self-harm, comparing time periods on lithium alone versus other medicines for 529,359 people
- Kidney disease, comparing medicine-free periods versus times on different medicines for 591,052 adults
- Diabetes, comparing medicine-free periods versus times on different medicines for 565,253 adults
What were the results?
For people with bipolar disorder, use of the following medicines and combinations had the lowest and highest health risks.
Hospital stays for mental health concerns. Compared with lithium alone, lamotrigine and bupropion alone or with a mood stabilizing anticonvulsant had a lower risk of hospital stays for mental health concerns. MAOI antidepressants and antipsychotics had a higher risk. Taking three or more medicines together also had a higher risk.
Self-harm. Compared with lithium alone, some mood stabilizing anticonvulsants, SSRI and SNRI antidepressants, new antipsychotics, bupropion, and a combination of SSRIs and bupropion had a lower risk of self-harm. Some combinations of new antipsychotics and other medicines had a higher risk.
Kidney disorders. Compared with medicine-free periods, lithium, MAOIs, and some combinations of psychotropic medicines had a higher risk of kidney disorders.
Diabetes. Compared with medicine-free periods, lithium, bupropion, mood stabilizing anticonvulsants, and SSRI antidepressants had a lower risk of diabetes. Some antipsychotics, SNRI antidepressants, and medicine combinations had a higher risk of diabetes.
What did the research team do?
The research team worked with people with bipolar disorder, their family members, advocates, and doctors to plan and conduct the study. The team held group discussions with people with bipolar disorder and their families to understand their concerns and choose study questions. Then the team reviewed health records from 2003 to 2016 for 1.3 million people with bipolar disorder. They looked at which medicines had better or worse effects on people’s health.
What were the limits of the study?
The study couldn’t assign people by chance to the medicines. As a result, the research team can’t say for sure if differences are due to the medicines or something else. The study didn’t look at medicine dose or whether people actually took their medicine.
Future studies could assign medicines to patients by chance.
How can people use the results?
Patients with bipolar disorder and their doctors can use the results when considering treatments.
Professional Abstract
Objective
To compare the risk of negative mental and physical health outcomes in people with bipolar disorder taking commonly prescribed psychotropic medications
Study Design
| Design Element | Description |
|---|---|
| Design | Observational: cohort study |
| Population | 1.3 million people with bipolar disorder (any type) |
|
Interventions/ Comparators |
|
| Outcomes |
Primary: psychiatric hospitalization, self-harm Secondary: kidney disorders, diabetes mellitus |
| Timeframe | Up to 13-year follow-up for primary outcomes (2003–2016) |
In these retrospective observational cohort studies, researchers used administrative claims data from 2003 to 2016 for 1.3 million people with bipolar disorder. For each outcome, researchers compared time periods when patients with bipolar disorder received one or multiple medications against periods when patients received lithium alone or against periods with no medications. They examined the risk of
- Psychiatric hospitalizations among 852,063 patients, comparing time periods on different medications and medication combinations versus lithium alone
- Self-harm among 529,359 patients, comparing time periods on different medications and medication combinations versus lithium alone
- Kidney disorders among 591,052 patients, comparing time periods on different medications and medication combinations versus medication-free periods
- Type 2 diabetes among 565,253 patients, comparing time periods on different medications and medication combinations versus medication-free periods
Researchers did not study people with other serious mental health conditions or, for secondary outcomes, people who had kidney disease or diabetes prior to starting medication. Researchers examined outcomes for up to 13 years after treatment initiation.
An advisory council of people with bipolar disorder, patient advocates, family members, and physicians was part of the study team. To identify outcomes of interest and interpret the findings, researchers conducted focus groups with people with bipolar disorder and their family members.
Results
For each outcome, use of the following treatments had the lowest and highest risk estimates relative to their comparators. These differences were statistically significant and clinically meaningful.
Psychiatric hospitalization. Compared with use of lithium alone, lamotrigine and bupropion—either alone or with a mood stabilizing anticonvulsant (MSA)—had a lower risk of hospitalization. MAOI antidepressants and antipsychotics had a higher risk. Taking three or more medications generally had a higher risk of hospitalization.
Self-harm. Compared with use of lithium alone, MSAs, second-generation antipsychotics, bupropion, and SSRI and SNRI antidepressants, plus a combination of SSRIs and bupropion, had a lower risk of self-harm, while first-generation antipsychotics as well as combinations of second-generation antipsychotics and other medications had a higher risk.
Kidney disorders. Compared with medication-free periods, use of MAOIs and medication combinations that included lithium, MSAs, or antipsychotics had a higher risk of kidney disorders.
Diabetes. Compared with medication-free periods, use of lithium, bupropion, MSAs, and SSRI antidepressants had a lower risk of diabetes. Some antipsychotics and many medication combinations had a higher risk.
Limitations
Patients were not randomly assigned to medications; unmeasured participant characteristics, such as mental illness severity or medical history, may have affected the findings. Researchers did not control for medication dosage, administration route, or adherence to prescribed regimen.
Conclusions and Relevance
Some medications and medication combinations for bipolar disorder may have fewer negative side effects than seen with lithium or medication-free periods. Taking multiple medications for bipolar disorder may bring additional risks for physical health compared with taking only one medication.
Future Research Needs
Future research could randomly assign participants to medications and combinations and track medication adherence.
Final Research Report
View this project's final research report.
Journal Citations
Results of This Project
Related Journal Citations
Peer-Review Summary
Peer review of PCORI-funded research helps make sure the report presents complete, balanced, and useful information about the research. It also assesses how the project addressed PCORI’s Methodology Standards. During peer review, experts read a draft report of the research and provide comments about the report. These experts may include a scientist focused on the research topic, a specialist in research methods, a patient or caregiver, and a healthcare professional. These reviewers cannot have conflicts of interest with the study.
The peer reviewers point out where the draft report may need revision. For example, they may suggest ways to improve descriptions of the conduct of the study or to clarify the connection between results and conclusions. Sometimes, awardees revise their draft reports twice or more to address all of the reviewers’ comments.
Peer reviewers commented and the researchers made changes or provided responses. Those comments and responses included the following:
- The reviewers asked whether the researchers used validated algorithms to measure the study outcomes: psychiatric hospitalization, kidney disorders, or diabetes. The researchers indicated that they did not use algorithms to measure outcomes, instead using billing codes to determine the presence or absence of a particular outcome. The researchers went on to say that in the report they described the multiple diagnostic and procedure billing codes that they included for each outcome. The reviewers responded by asking for comparisons between the researchers’ methods of identifying the clinical outcomes and existing algorithms created by other authors. The researchers explained further that they created their outcomes definitions from scratch and with input from their clinician-collaborators rather than relying on existing definitions. The reviewers noted that readers would be interested in the comparative definitions of the clinical outcomes so that readers could understand the breadth and accuracy of the definitions that the researchers used. The researchers responded by adding the sensitivity and specificity of previously developed algorithms or definitions of the clinical outcomes, particularly diabetes.
- The reviewers noted that the study data set was restricted to patients with private health insurance so the results of the study might not generalize to patients on public insurance. The researchers explained that although they had originally planned to include Medicaid patient data, the foundation supplying those data discontinued their supply due to funding problems. The researchers added a limitation to the discussion to point out that the lack of Medicaid data meant that most of the most severely affected patients with bipolar disorder were not represented in the study.
- The reviewers expressed concern about the requirement that patients maintain continuous drug exposure and maintain monotherapy, to stay in the study. While acknowledging that these elements were analytic strengths, the reviewers said that these expectations both led to reduced follow-up due to patients changing or adding medications, or not refilling their prescriptions one month. The researchers noted that these were the results of the initial study, which they explored further in focus groups, and that they adjusted their subsequent analyses based on these findings. The researchers concluded that low adherence to monotherapy is a major public health concern requiring system-wide interventions.