Intended for healthcare professionals

Editorials

Gabapentinoids linked to new risks, including suicidal behaviour

BMJ 2019; 365 doi: https://doi.org/10.1136/bmj.l4021 (Published 12 June 2019) Cite this as: BMJ 2019;365:l4021

Linked Research

Associations between gabapentinoids and suicidal behaviour, unintentional overdoses, injuries, road traffic incidents, and violent crime

  1. Derek K Tracy, consultant psychiatrist
  1. Queen Mary’s Hospital, London DA14 6LT, UK
  1. derek.tracy{at}nhs.net

Young people taking pregabalin look most vulnerable

Medical training might start with “first, do no harm,” but in truth the prescribing of drugs is a balance of risks. All drugs can cause harm, although the extent of this varies among different people, drugs, dosages, and durations of use.

Gabapentinoids (primarily gabapentin and pregabalin) are indicated for epilepsy and neuropathic pain. Pregabalin is also used for anxiety disorders. Off-label use has included restless leg syndrome, migraine, menopausal flushes, and alcohol dependency.1 Acute, well documented side effects of gabapentinoids include motor incoordination, dizziness, cognitive impairment, and suicidal ideation.2 Few data exist on how these might translate into longer term harm over months and years, and existing evidence is patchy and inconsistent.

Molero and colleagues (doi:10.1136/bmj.l2147) help to fill this evidence gap3 by exploring a range of harms among adolescents and adults (>15 years) prescribed gabapentinoids, including suicidal behaviour, unintentional overdose, injuries, road traffic incidents, and violent crime. Their study is timely. Prescribing rates have risen sharply in recent years4; evidence has emerged that these agents are being used clinically (but ineffectively) as an opioid substitute5 and diverted to recreational use6; and detection at post mortem has increased, along with direct attribution as cause of death.7 These concerns have led to legislative changes to restrict or control prescribing in several countries, including the United Kingdom.8

Acute harms associated with drug treatments are relatively easy to study, although reporting is often inadequate. Unpicking complex downstream effects is much more challenging. Most trials are too short term and underpowered to capture these events reliably, and the number of potential confounders is formidable.

Molero and colleagues overcome these issues by analysing data from a large cohort of almost 200 000 people prescribed gabapentinoids in Sweden between 2006 and 2013. They used a within-individual design, which compares the time someone takes a drug with the time he or she is not taking the drug. This enabled determination of adjusted hazard ratios indicating the risk of specific harms during treatment periods, compared with baseline risk during periods without treatment. During treatment periods, participants were at increased risk of suicidal behaviour or death from suicide (age adjusted hazard ratio 1.26, 95% confidence interval 1.20 to 1.32), unintentional overdose (1.24, 1.19 to 1.28), head or body injuries (1.22, 1.19 to 1.25), and road traffic incidents or offences (1.13, 1.06 to 1.20).

The authors report no statistically significant associations between treatment with gabapentinoids and violent crime. Crucially, subanalyses showed that only pregabalin, not gabapentin, was associated with increased risks of harm, and that this was largely in a dose-response manner. Stratification showed age as the only factor associated with differential risk, and participants aged 15-24 were the most vulnerable group.

These findings are important and clinically helpful. Against a backdrop of growing concern and legislative restrictions, these authors report solid data to inform patients on the risks associated with treatment. The authors’ large dataset and study design cleverly overcome the limitations of case reports and small trials that fail to capture off-label prescribing or longer term problems.

This work could not determine causality—namely, that gabapentinoids definitively produced these adverse outcomes. The important issues of drug adherence and any acute interplay of alcohol and illicit drug use also could not be determined. However, within-individual analyses allow participants to act as their own controls, removing important confounders, particularly confounding by indication, where bias is introduced when people prescribed gabapentinoids already have conditions such as depression or dyspraxias that are associated with the outcomes of interest.

This study also reports invaluable information differentiating risks by age group, and between the two most commonly prescribed agents in this class. The findings challenge the clinical utility of the broad categorisation “gabapentinoid” and suggest it might be time to uncouple pregabalin and gabapentin for the purposes of legislation and guidelines. Both might need to be more specific about age group.

Further outstanding challenges include improving our understanding of gabapentinoid dependency: how it happens and its prevalence, incidence, and determinants.910 We also need to understand what is driving the age related differences in risks and how recent legal restrictions will affect the illicit market in diverted drugs.

Since gabapentinoids became scheduled (restricted) drugs in April 2019, clinicians in the UK have faced new prescribing challenges: prescriptions must be signed by hand, repeat prescriptions are banned, and pharmacies must dispense the prescribed drug within 28 days of issue.11

Although subsequent lobbying has meant that gabapentioids do not need to be locked in a controlled drug cupboard,12 the other inconveniences associated with rescheduling, combined with the new findings by Molero and colleagues, could have a chilling effect on prescribing. How this will reduce these specific harms remains unclear.

Despite reasonable concerns, gabapentinoids remain a valued therapeutic option for many people. Medicines can harm as well as heal, and the best treatment decisions are made in full partnership with patients, after consideration of all available evidence on both.

Acknowledgments

I thank the patients and relatives who shared their personal experiences of benefits and harms from gabapentinoids; these have helped shape this editorial.

Footnotes

References

View Abstract