Understanding why antidepressants don’t work for everyone

Why antidepressants don’t work for everyone with depression

Medicines don’t help everyone with depression, and we don’t yet know why. A recent King’s College London study looked at people with depression who often switch medication to find those who don’t respond to treatments. This could help doctors give more tailored care sooner.

Major depression is common, affecting around 300 million people worldwide. More than feeling sad, it involves constantly feeling low and losing interest in life. Common treatments are medications that boost ‘feel-good’ chemicals in the brain called ‘Selective-Serotonin-Reuptake Inhibitors’ or SSRIs. An example is Prozac. Together with similar medication, these are called antidepressants.

Finding the right antidepressant may take trial and error. Like choosing fertiliser for a garden, sometimes the first one works, sometimes you need to try another before plants start to grow. About one third of people with depression recover with their first antidepressant. Others do not improve and may need to try different medications. Some eventually develop what is called treatment-resistant depression, where at least two antidepressants fail to work, adequately.

For this group, it may not be about finding the right fertiliser. They may need a different approach, like reworking the soil itself or planting in a new way.

Why don’t antidepressants work for everyone?

We don’t know why some people respond well to antidepressants while others don’t. Research has found common features in those who don’t respond, including experience of childhood trauma. Studies also suggest that recovering from depression with antidepressants may run in families. Many questions remain unanswered. For example, do people showing nonresponse share a particular genetic profile? A genetic profile is your unique code or instruction manual. As well as guiding features like eye colour, it provides directions for how your body reacts to medicines.

So how to find those who don’t respond?

Treatment response isn’t recorded in UK standard electronic health records. Smaller studies on treatment response exist, but findings are difficult to generalise to the millions who show nonresponse.

The UK Biobank (UKB) holds genetic data volunteered by 500,000 people, much of it linked to medical record diagnoses and prescriptions. Lo and colleagues at King’s College London aimed to use this data to spot those who don’t respond to treatment and find their shared characteristics. Their work was supported by the Wellcome Mental Health Award and the Maudsley Biomedical Research Centre.

Because treatment response isn’t collected in UKB, the team looked for a different way to find those who don’t respond. A new approach. Doctors often suggest switching medicines when one medication doesn’t work. So, the team tested whether people who switched antidepressants (i.e., tried a new fertiliser) were more likely to represent a group with poor response to treatment.

The researchers first used a formula to find those in the data with depression, then defined “switchers” and “non-switchers”. Switchers were people prescribed an SSRI followed by another antidepressant within 90-days, while non-switchers received three or more SSRI prescriptions in a row. The team found 5,133 switchers and 33,680 non-switchers in the data, then used these groupings to test several questions.

Is switching a sign of nonresponse?

The researchers examined whether people who switched antidepressants shared genetic patterns previously linked to poor antidepressant response.

Using statistical analysis, the team also looked for common features in the switchers, such as income, or education. These traits were compared with findings from previous treatment response studies as another way to check whether switching could suggest nonresponse.

Finally, advanced statistical tools were used on the genetic data to look at whether inherited genetic differences contributed to antidepressant switching.

Where possible, the same analyses were repeated on ‘Generation Scotland’ data to check whether the UKB results were correct. Generation Scotland is genetic, mental and physical health information for 24,000 volunteers, linked to prescription records.

If switchers are the nonresponse group, Lo and colleagues’ approach opens the door to studying thousands who don’t respond to treatment. In doing so, researchers can learn why some people end up in this group. If we know why, doctors can assess who is at greater risk of nonresponse and look at alternative, personalised treatments to help them with depression.

The results

As suspected, the team found that switchers and nonresponse share genetic similarities and other traits. Lower education and income level were linked to more switching, just like previous research on treatment resistance. The findings suggest that antidepressant switching can act as a useful indicator of nonresponse in large health datasets. The findings also suggested that antidepressant switching may partly reflect genetic influences that are different from the genetics linked to developing depression itself.

This study provides a way to spot nonresponse to antidepressants in big datasets and highlights traits linked to nonresponse. It has the benefit of using large, real-world data. While previous studies have looked at switching and nonresponse, Lo and colleagues used a longer timeframe in their switching definition. This made it possible to collect more data and deliver more believable results. They also based switching on prescription dates, avoiding guesswork about treatment length. The results now need to be checked in larger groups to be more certain.

Moving forward cautiously

While the switching definition is useful, it’s not the same as nonresponse. UKB has 13% switchers, much less than the one third estimated to go on to be treatment resistant. Prescriptions don’t tell the whole story. People may switch for other reasons, such as bad side effects. Some who don’t respond, may stop all medication, meaning they wouldn’t be counted as switchers. Also, prescriptions don’t guarantee that patients take the medication, not to mind their response.

Most UKB participants were white and around two thirds were female, so it’s difficult to generalise findings to the whole population. Moreover, Generation Scotland’s data’s too small to draw conclusions from and perhaps too different to be of use checking the UKB findings, with 28% switchers in the Generation Scotland data compared to just 13% in the UKB data.

How will this help those with depression?

Despite drawbacks, this research brings us closer to understanding why some people don’t respond to antidepressants. Switching is a helpful definition for future large genetic studies and for building a clearer picture of nonresponse and treatment-resistance. The findings may guide new drug research to support personalised, alternative solutions, matching the medicine to the person. In the future, this kind of research may help doctors predict which treatments are more likely to work for different people. It’s a step towards making sure everyone with depression can find the right help.

Image references for video: All background images created using ChatGPT