Gabapentinoids unlikely to be directly linked to self-harm risk: BMJ

Treatment with gabapentinoids - drugs such as gabapentin and pregabalin - is not directly associated with an increased risk of self-harm, finds a UK study published by The BMJ today.

However, rates of self-harm were higher before and shortly after treatment, highlighting the need for close monitoring of patients throughout their treatment journey, say the researchers.

Gabapentinoids are prescribed for conditions such as epilepsy, nerve pain, and anxiety disorders.

Previous studies have raised concerns about potential side effects, including an increased risk of self-harm, but did not examine the risks immediately before starting or after stopping gabapentinoids, despite the fact that these drugs may be prescribed for conditions associated with self-harm. As such, the nature of this relationship is still not fully understood.

To obtain a clearer picture, researchers analysed electronic health records for 10,002 individuals in the UK aged 18 and over (average age 39; 67% female) who were prescribed gabapentinoids between 2000 and 2020 and had at least one hospital record of self-harm.

They examined rates of self-harm in four distinct time periods for each individual - 90 days before starting treatment, the treatment period, the 14 days after stopping treatment, and any other period, which acted as the reference category.

By comparing each individual with themselves across time, the researchers were able to account for all potential confounding factors, including genetic risks, childhood environment and early-onset conditions. They also accounted for age, seasonality, and other prescribed opioid and psychotropic medications.

Of the 10,002 individuals, 4,767 (48%) took gabapentin only, 3,164 (32%) took pregabalin only and 2,071 (21%) took both over an average follow-up period of 13 years.

Compared with the reference period, the rate of self-harm increased by 69% (16.8 per 100 person years) in the 90 days leading up to the start of treatment, suggesting that individuals prescribed these drugs may already be at an increased risk.

This risk declined during the treatment period, but markedly increased to 29.6 per 100 person years in the two-week period after treatment had ended, before returning to reference levels, suggesting that gabapentinoids are unlikely to be linked to self-harm risk.

The researchers acknowledge that these are observational findings and results may be limited by the small sample sizes within some groups. What’s more, they were only able to capture prescriptions issued by a general practitioner and cases of self-harm severe enough to be admitted to hospital.

Nevertheless, their use of a large population based database provided sufficient statistical power to evaluate the association between gabapentinoid use and self-harm, and results were similar after further analyses, suggesting they are robust.

This study provides valuable insights into the association between gabapentinoid treatment and self-harm risks, they write. And while further studies are needed in different populations, they say these findings “underscore the necessity for close patient monitoring of self-harm throughout the gabapentinoid treatment journey.”

This investigation shows the importance of testing associations in primary and secondary care, and the authors' novel approach of considering periods before and after treatment is an important contribution, note researchers in a linked editorial.

They point to some important study limitations, but acknowledge that, “clinically, their results suggest that routine and periodic follow-up of people prescribed gabapentinoids should be considered, particularly in the weeks after medication has been discontinued.”

They add: “Whether young adults and people with no psychiatric diagnoses need more supervision while taking gabapentinoids requires further research to clarify.”

Reference:

Sariaslan A, Fazel S. Use of real world data to understand self-harm risk in people prescribed gabapentinoids BMJ 2025; 389 :r634 doi:10.1136/bmj.r634