Finnish Study Links Six Medicine Groups With Higher Odds of Fatal Poisoning in Bipolar Disorder

A Finnish registry study has found that recent use of six medicine groups was associated with higher odds of fatal poisoning among people with bipolar disorder. The groups were tricyclic antidepressants, gabapentinoids, opioids, Z-drugs, benzodiazepines and antipsychotics.

The findings come from an observational study, meaning they show which medicines were dispensed before poisoning deaths—not that the medicines caused those deaths. The researchers said some of the associations may reflect more severe illness or other differences between the people who died and those who did not.

What the study examined

Researchers used Finnish national registry data from 1996 to 2018. The wider cohort included 60,045 people diagnosed with bipolar disorder between ages 15 and 65.

The analysis compared 777 people whose main cause of death was poisoning by medicines or illicit substances with 3,108 living people matched by age and sex. Four controls were matched to each case.

Medicine exposure was based on pharmacy dispensing records from the two weeks before death, or the corresponding date for controls. Those records show that a medicine was dispensed, not necessarily that it was taken.

Medicine groups associated with higher odds

After adjustment, the study reported the following adjusted odds ratios for fatal poisoning:

• Tricyclic antidepressants: 3.07
• Gabapentinoids: 3.05
• Opioids: 2.65
• Z-drugs: 2.19
• Benzodiazepines: 2.13
• Antipsychotics: 1.94

These figures describe the relationship between recent dispensing and fatal poisoning in the study population after statistical adjustment. They do not establish that a particular medicine was involved in an individual death.

Opioids, sedatives and gabapentinoids were described as carrying more risk when used together than alone. The combination of all three had the highest odds reported in the study, with an adjusted odds ratio of 4.82, although the difference from other combinations was not statistically significant.

Findings for other medicines

Lithium, valproate, lamotrigine and carbamazepine did not show a significant association with fatal poisoning in the main analysis. Lithium was recorded for 10.6% of people who died and 10.1% of controls; its adjusted odds ratio was 1.07, with a 95% confidence interval of 0.77 to 1.48.

SSRIs also were not significantly linked with fatal poisoning. Their adjusted odds ratio was 1.19, with a 95% confidence interval of 0.94 to 1.51.

Some results differed by the type of poisoning. Carbamazepine was associated with higher odds of suicidal poisoning, while mood stabilizers overall and valproate were associated with lower odds of accidental poisoning.

Context around the deaths

Among the poisoning deaths, 56.5% were recorded as suicidal and 38.4% as accidental. People who died were more likely than controls to have a substance use disorder—67.1% compared with 27.1%—and a history of suicide attempt—44.3% compared with 12.8%. Employment was less common among those who died, at 9.8% compared with 32.5% of controls.

These differences are important context because they indicate that the people who died had substantially different health and social profiles. The researchers said illness severity may partly explain the medicine associations.

Why the results need careful interpretation

The study did not include information on symptom severity, the reason a medicine was prescribed or bipolar disorder subtype. Pharmacy records could not confirm whether dispensed medicines were taken, or whether medicines from older prescriptions had been kept.

The researchers also lacked complete information about alcohol, over-the-counter medicines and illicit drugs. Toxicology data were not linked to the records, so the study could not identify which substance, if any, was involved in each death.

The results come from one country and cover prescribing practices from 1996 to 2018. Patterns of treatment and medicine use may differ in Finland today and in other countries.

What the findings mean for care discussions

The researchers said clinicians should consider overdose toxicity when selecting medicines, limit dispensed quantities and treatment duration where appropriate, monitor combinations, ask about alcohol and other substances, and help patients recognize poisoning symptoms.

The study is not a basis for changing treatment without clinical guidance. Its central message is about medication safety and risk assessment in a population with a high burden of poisoning—not proof that any listed medicine caused a death. Earlier work on the same Finnish cohort found that 13.1% of people with bipolar disorder had experienced at least one poisoning leading to hospital admission or death.

AI tools were used to assist with the preparation of this article under our editorial standards.