Depression, Diagnosis, Treatment, Medication, and Biomarkers

How many times has it been said that “There’s no lab test for depression.” That was the case for thousands of years. Even since the beginning of the modern mental health treatment era with Freud and Kraepelin in the late 1800s, the causes of mental illness were guessed at by observations of behavior.

There are many reasons a person might be depressed, and there are many tools available to treat depression. How can a person know how to move forward with medication or talk therapy treatment? When is talk therapy most effective? What kind of medications are available to treat depression? Why would you choose one treatment over another? What is a biomarker, and why might it be useful in treatment?

One generally accepted treatment method is to use the STAR*D treatment steps. Start with Citalopram and see how that goes for 14 weeks. Why Citalopram? That medication is a selective serotonin reuptake inhibitor. That means that Citalopram helps keep serotonin in the part of a person’s brain where it can be most effective. Insurance companies tend to reimburse providers for this kind of treatment without complaint, and they’ll pay for those medications.

If you have a serotonin problem then this can be a nice solution. Women get better results from Citalopram than men. This is because women have more problems with serotonin than men do. The female reproductive system and hormone system affect serotonin more than men’s systems.

A biomarker is an objective medical status that can be accurately measured and reproduced. Here is a link to a study that uses an electroencephalograph (EEG) to locate a biomarker for depression and response to Prozac (fluoxetine), a similar SSRI medication. An EEG looks something like a swimming cap with a bunch of spark plugs stuck in it that are connected to a computer. First, the research team screens the patient for depression using the standard questionnaire, the Hamilton Rating Scale for Depression (HRSD). Here is a copy of that questionnaire. Insurance companies want PMHNPs and psychiatrists to use this form to decide whether a patient has depression and will accept it as documentation supporting that diagnosis. Then, researchers measured electrical brain activity in several areas of the brain and recorded their results.


I would like to call your attention to the difference between those two methods. The HRSD takes the patient’s verbal response to questions, and the EEG reads electrical brain activity.

Within that study, researchers found that using the EEG they could measure a patient’s response to Prozac at the end of the first week of treatment and predict how they would do at 8 weeks of treatment. They repeated the EEG at one week and calculated the difference between their first result on day 1 and their second result a week later. Using this calculation (called the Antidepressant Treatment Response Index), researchers could predict the patient’s response to treatment seven weeks in the future. By contrast, the HRSD did not offer a prediction of success. The EEG does not record the placebo effect, the change in the patient’s biomarker reflects the performance of the medication only. The HRSD assessment includes the placebo effect.

Wouldn’t it be nice to know whether you are experiencing the benefit of the medication vs the placebo effect? As a provider, this is vital information. I would not want to continue a medication that has the side effects that many antidepressants have (nausea and sexual dysfunction) when talk therapy would probably be a better option. A strong placebo effect indicates that a patient probably has hope. Hope is a wonderful thing to find in a patient seeking treatment for depression! Their depression might be event-driven, temporary, due to something that is not connected to serotonin, or otherwise resolvable.

Using the EEG in addition to the HRSD shortens the time from 14 weeks to 1 week. Isn’t that an incredible outcome? 13 weeks of medication costs money, and imagine spending 14 weeks waiting to see whether you are going to feel less depressed only to find out that it’s not ultimately going to work for you. Even if the medication seems to be working, there is still a significant chance that you feel better because of the placebo effect. Finding out that Citalopram or Fluoxetine probably isn’t going to work for you might not sound like good news, but it is. That tells us that your answer lies elsewhere. Instead of trying to find a version of serotonin that will be helpful, maybe we should look to dopamine or norepinephrine for a more likely solution.

The fact that a person can distinguish between these outcomes changes everything.

While this study is interesting and I believe very helpful, it was the follow-up to this one that converted me to a biomarker-hunter. We can look at that one next.

What’s new in mental health care?

I’m glad you asked! The good news is that there are an increasing number of tests that can help a clinician chase down what is causing some mental health problems.

Genetic testing can help find out how your body processes medication. Why does that matter? Some medications have to be metabolized before they can be used by the body for their intended purposes. For example, codeine is metabolized in the liver by cytochrome P450 CYP2D6. CYP2D6 is an enzyme that is about 2% of the liver’s CYP system, but it processes about 25% of clinically used drugs (Wang et al, 2006). 2D6 is polymorphic, which means that it comes in a number of different forms as dictated by the DNA that codes for its construction. Some people have a low speed, some a medium speed, and some have a high speed version of 2D6. If you have a low speed 2D6, codeine may not be transformed into morphine at a rate that would provide you with pain relief. A medium speed 2D6 would metabolize codeine as the manufacturer expects. A high-speed 2D6 might convert codeine into morphine too quickly, and oversedate before it wears off much too early. Maybe codeine is not a good fit for a person with a high-speed 2D6!

That information alone is very useful, but it is even more useful in the context of a patient’s personal situation. There are substances that affect how 2D6 processes medication. For example, Prozac (fluoxetine) slows down 2D6. The importance of this is that changes in one medication can affect how another one works. “After I started taking Prozac my depression got better, but now my knees hurt just like before I started taking codeine. What happened?” Now we can find out.

There is a small group of people who have the right enzyme to make the best use of Zyprexa (olanzapine). They can use a low dose and get relief of their symptoms without running the risk of winding up with metabolic syndrome. The weight gain alone can cause people to stop taking that medication, but the diabetes that comes with metabolic syndrome can take years off a person’s life. As a prescriber, I want to avoid causing a person to go through that series of events. Genetic testing can help me avoid causing that particular problem.

Genetic testing is a great tool to have available. Ten years ago, it was simply not available. Psychiatrists had to work with trial and error to find medications that would work for patients. It’s a good thing that era has finally come to an end.

Wang, B., Yang, L., Zhang, X., Huang, S., Bartlam, M., & Zhou, S. (2009). New insights into the structural characteristics and functional relevance of the human cytochrome P450 2D6 enzyme. Drug Metabolism Reviews, 41(4), 573-643. doi:10.1080/03602530903118729