When depression gets worse on an antidepressant, it doesn’t necessarily mean the patient has bipolar disorder or that they are sensitive to medications. Today, we detail 7 explanations for this paradoxical phenomena.
This month we featured a new study on an old controversy: Can antidepressants cause violence? The study is the best designed one we’ve seen to tackle this question, which dates back to the early 1990’s when case reports were described on fluoxetine. The idea goes against the grain of practice, where SSRIs are used to treat anger and aggression due to conditions like personality disorders and intermittent explosive disorder. But it reminded us of another issue – one that we see a lot more often in clinical practice – can antidepressants make depression worse?
This is the kind of problem that – even if it’s rare – you’re likely to encounter in psychiatric practice because the majority – 80% – of antidepressants are prescribed in primary care. The ones that get better are not likely to come to our doors, but the ones who don’t respond or get worse will.
1. Side Effects
When a patient says they feel worse on an antidepressant, the first thing to do is to rule out physical side effects like nausea, fatigue, and insomnia. One to pay attention to is akathisia, as patients may have difficulty describing this inner sense of restlessness which is more associated with antipsychotics but can occur on antidepressants as well, particularly serotonergic ones. And akathisia can cause anxiety, insomnia, and even suicidality.
How can serotonergic antidepressants cause akathisia? I thought it was due to dopamine blockade?
It’s thought that the inhibitory effects of serotonin have indirect effects on the dopamine system, and can lead to dopamine antagonism there in the striatum. Another, related side effect you may see on serotonergic antidepressants is restless leg syndrome, which is almost like akathisia at night. If that happens, you could add gabapentin or pramipexole – both of which treat restless legs and akathisia – and both of which have psychiatric benefits – gabapentin helps sleep and anxiety and pramipexole helps depression. Or you could switch to bupropion, which treated restless leg syndrome in a randomized controlled trial, even when dosed in the morning.
And while we’re on the subject, sleep disruption is another reason that antidepressants can make people feel worse. Bupropion often gets a bad rep here because it’s thought to be an activating antidepressant, but it actually causes initiation insomnia at about the same rate as SSRIs. On the other hand, the SSRIs tend to lighten sleep, making sleep more restless, while bupropion tends to deepen sleep quality. And if your patient has poor sleep quality, they may not complain of it directly – they may just say that they are tired, can’t concentrate, and are easily frustrated and can’t manage stress. The antidepressants that are pretty favorable when it comes to sleep quality are three that have sedating effects:
And three that are relatively non-sedating:
So, Dr. Aiken, once you’ve ruled out side effects and sleep problems, what else goes through your mind when a patient tells you they feel more depressed on an antidepressant?
2. Bipolar Disorder
The most common reason after side effects is bipolar disorder. For a long time psychiatrists debated whether antidepressants caused mania – there were studies showing it did, and some showing it didn’t – but in my mind the issue was settled in 2014 when Alexander Viktorin published a large study of over 3,000 patients in the American Journal of Psychiatry. His team focused on the first 3 months after starting an antidepressant, and only looked at bipolar patients. They found no increased risk of manic switching when antidepressants were taken with a mood stabilizer, but when taken on their own as monotherapy in bipolar the risk of manic switching was 3-fold higher.
So does that mean it’s safe to give an antidepressant as long as they are on a mood stabilizer?
No, this study only looked at mania, but other studies have found problems like mixed symptoms and rapid cycling after starting an antidepressant, even when a mood stabilizer is on board. However it is clearly less risky when mood stabilizers are on board, but then again staying up all night is also less risk with a mood stabilizer – we still wouldn’t advise a person with bipolar disorder to do that.
But aren’t there studies where antidepressants treat bipolar depression?
There are, but most of them have design flaws, like small sample size, enriched design, or lack of a placebo arm. In the well-designed studies, like one from that Gary Sach’s group published in the New England Journal of Medicine, antidepressants did not work. My take on all this is that they help a minority of patients with bipolar disorder – Terrence Ketter estimates it at about 1 in 30 to 1 in 40, and they possibly work better in bipolar II than bipolar I, but they also have a high chance of causing harm and an even higher chance of doing nothing. And the harm they cause is rarely mania – antidepressants are usually not powerful enough to flip someone out of depression and into mania. Instead they fail to treat the depression, and they sprinkle a few manic symptoms on top of it – which is called a depression with mixed features. And when this happens to patients, they are not going to call you up and say “That antidepressant you gave me caused mixed features.” Instead they will say “It’s making my depression worse – I feel anxious, wired, like I’m crawling out of my skin.” Antidepressants will turn a depression – which is a passive, inactive state – into dysphoria – which is a desperate state that craves urgent relief.
That makes sense because mixed states are more severe than depression – they have higher rates of suicide, substance abuse, and psychiatric hospitalization.
3a. Bipolar Features: Depression with Mixed Features
And that’s the second type of patient who can get worse on an antidepressant: Depression with mixed features. This diagnosis is new in DSM 5 and you don’t have to have bipolar disorder to have it. It’s actually quite common in unipolar depression – one meta-analysis found that 25% of patients with depression had at least 3 manic features – which is the DSM-5 cut off for mixed features. We don’t know much about how to treat these unipolar depressions with mixed features. On the one hand, they are more likely to get worse on antidepressants, but on the other hand some will respond to them so it’s tricky.
Most of the patients I see in practice who have mixed features are already taking an antidepressant.
Yes, and the first goal is to figure out if that antidepressant is making things better or worse, but you can’t just yank it away as they are likely to have withdrawal symptoms that can resemble a mixed state, particularly if it’s a serotonergic antidepressant. If it’s clear that the antidepressant has made things worse, I would taper it off, but if there’s any doubt I’d leave it there until I’m able to get the mixed features better – usually starting with lurasidone which has the best evidence for this and almost earned FDA approval for mixed features – and then I’d use any medication that’s appropriate for bipolar disorder. Ideally you’d want to use something that can treat both mania and depression, like lithium or quetiapine, but if the manic symptoms are more prominent you may use something with more anti-manic effects, and if the depressive symptoms are more prominent you might use lamotrigine. Then, once they improve, I might slowly try to taper off the antidepressant over at least a few months.
We covered this in our January 2018 issue, which has a link to the Bipolarity Index. This is a validated rating scale that uses non-symptomatic markers of bipolar – like age of onset, treatment response, and family history – to assess the likelihood that a patient has bipolar disorder on a 100-point scale. It’s particularly helpful when faced with mixed features, because the symptoms are so hard to identify – they can look like anxiety, PTSD, or even ADHD.
3b. Bipolar Features: Depression with Short Duration Hypomania
Another area where we see this is short duration hypomania. About 1 in 20 patients with recurrent depression have brief hypomanias that are classic and recurrent but never last long enough to meet DSM criteria for bipolar – they last less than 4 days. Jules Angst published a remarkable study where he showed the chance mood worsening on an antidepressant rises steadily with the duration of hypomania – from 1% in patients with no history of hypomania, to 9% if they’ve had it for less than a day, 14% for 2-3 days, and then we get into the real bipolar levels – 4 days – the cut off for bipolar II – the risk is 27% – and if they’ve ever had hypomania for more than a week the risk of getting worse on an antidepressant is 38%. But, still, that’s only 38% – it’s not 100% – none of this is absolute.
4. Borderline Personality Disorder
Another reason that patients can get worse on antidepressants is borderline personality disorder. And this isn’t just because there’s a lot of symptom overlap with bipolar. Borderline personality itself has been identified as a predictor of antidepressant worsening, and patients with borderline personality tend to get worse – more depressed or more aggressive – on tricyclics – for example there are two studies showing this effect with amitriptyline. On the other hand, there are studies where some borderline symptoms improved on antidepressants – particularly SSRIs – so this is not an absolute, but keep in mind those studies all had flaws – enough that the evidence for SSRIs in borderline was considered too inadequate to recommend them in several recent reviews.
5. Age Under 25
The fifth area where you may see worsening on antidepressants is in children and young adults. The FDA has placed a warning about suicidal ideation on antidepressants in people under age 25. Now, this is complicated because there is also some epidemiologic evidence that suicide rates have gone down in youth when antidepressants were used, but I don’t think it’s too much of a stretch to acknowledge that they may cause harm in some and help others.
So how common is this problem?
Well, the risk seems to taper off between ages 18-25, which is interesting because 25 is also the age where the brain is thought to reach full maturity. Then, in adults from age 25-65 the risk of suicide stays flat – it doesn’t go up or down with antidepressants – and after age 65 we can detect a clear preventative effect against suicide with antidepressants.
In terms of how common it is, about 1 in 100 children will develop suicidality due to an antidepressant – actually the rate is 3% on antidepressant and 2% on placebo. Sometimes you’ll see this in practice as intrusive suicidal thoughts, but more often it’s part of a bigger picture of mood worsening that may include anxiety, insomnia, or aggressive behaviors. In 2016 the Cochrane group published an analysis of 70 trials of SSRIs and SNRIs, and they concluded that these medications doubled the risk of suicidality and aggression in children but not adults.
And do we know anything else about which children are at risk?
The risk is greater under age 12 and if the antidepressant is raised too quickly or the dose goes too high. The other risks bring us back to bipolarity – children are much more likely to experience mood worsening – which may or may not include suicidality – if a close relative has bipolar disorder or the child has mixed features.
So far we’ve covered 5 reasons why mood can get worse on an antidepressant:
- Side effects like akathisia and insomnia
- Bipolar disorder
- Bipolar features, like depression with mixed features or short-duration hypomania
- Borderline personality disorder
- Age less than 25
And now for #6: Genetics. Some research suggests that patients with the short arm of the serotonin transporter gene – the S/S genotype on SERT – are more likely to develop mania, and possibly suicidality or just more side effects on a serotonergic antidepressant. This is by no means definitive, but then again, neither are any of the other risks we mentioned in this podcast. This is about risk factors, not absolutes.
7. Medication Sensitivity
And #7: Some patients are just more sensitive to medications overall. Maybe they have an anxiety disorder, or hypochondriasis, or maybe they are a slow metabolizer and get really high levels of the med.
Yes, med sensitivity is always that possibility, but I would leave that as a diagnosis of exclusion. If you take away anything from this podcast, know that there are specific diagnoses out there that can cause patients to feel worse on antidepressants. And when you get a phone call from that patient, they are unlikely to recite the DSM-5 criteria for mania or quote the FDA warning on suicidality – those are the 2 risks we are all warned about – but they are unlikely to be what the patient reports, even when they are happening. Mania more likely to present as a dysphoric, anxious mixed state, and patients with suicidal thoughts may not share them with us for fear of being involuntarily committed, but they will share any number of symptoms – anxiety, restlessness, dysphoria – that may have worsened since starting the antidepressant.
So think carefully about these other possibilities before concluding that the patient is med-sensitive. That term doesn’t tell us very much, other than to keep trying with antidepressants with lower and slower doses.
And that is what I was taught to do in the 1990’s – a time that David Healy called the Antidepressant Era. Back then pharmaceutical companies weren’t required to release all their data, and the new antidepressants seemed so much safer than the older ones that few psychiatrists questioned them. One of those psychiatrists was Fred Goodwin, the former Director of the National Institutes of Mental Health, who passed away last September. 20 years ago Fred wrote a punchy paper called “The bipolar spectrum and the antidepressant view of the world” that woke me up to the idea that antidepressants could make mood worse. The coauthor on that paper was a young Nassir Ghaemi, who has since become one of the leading critical thinkers in psychopharmacology. Nassir is launching a new video course on psychopharmacology with CME credits – you can sign up at www.psychiatryletter.net. The first lecture starts February 20th, but you can watch it anytime.
And now for the word of the day…. Dysphoric
Dysphoric comes from the Greek for “Distress that is hard to bear.” It is a profound state of unease or dissatisfaction. Dysphoria is a symptom that is not tied to any specific disorder in psychiatry – it’s been used in premenstrual dysphoria, gender dysphoria, and in mood disorders it is often used to describe mixed states or depressive “dysphoric” mania. This seems fitting, as mania is an unsatisfied state of mind – that is why people with mania do things to excess – nothing satisfies them. So while a depressed patient may delay seeking help for months or weeks because of hopelessness and low motivation, a dysphoric patient is likely to call for urgent relief. The late Athanasios Koukopoulos, who studied mixed states throughout his 50 year career, described it this way: “The patient complains of anxiety, inner tension, irritability, anger, despair, suicidal impulses, crowded or racing thoughts, rumination, and insomnia.”