Stopping Antidepressants in Bipolar Disorder

(updated 12/2014)

Bottom line: go slow, much slower than you would have thought necessary; and of course, not without your prescriber’s direct involvement.


This page used to be entitled Breaking News: Two Psychiatrists Agree on 31 Points!  Hard enough to get agreement on anything important, let alone 31 different items. But on this issue, my colleague Tam Kelly (Fort Collins, CO) and I are in complete agreement.

Dr. Kelly evolved these 31 recommendations during 20+ years of seeing patients. Like me he has become a mood and anxiety disorders specialist, not entirely on purpose, but because we both found that some people with mood and anxiety problems were not getting better with standard approaches. (Obviously many people have gotten better with standard approaches, which is great; but for those who don’t, well, that’s where these 31 recommendations come in).

I was going to write my own page, and did publish some evidence for this approach, but Dr. Kelly’s list is more fun [with just a few explanations from me in brackets]. As you’ll see, he has a dry sense of humor and likes to speak plainly.

When to stop antidepressants in people with bipolar disorder

1. If they have been on them a short time I stop them.

2. If less than 2-3 weeks and they are suffering from the antidepressant, taper quickly.

3. Less than a week, stop: two weeks then cut in ½ , a week later stop.

4. Likewise if they just increased their dose I will do the above, decreasing to their previous dose and get rid of the rest later.

5. If the pt is doing well, no mixed state symptoms or cycling, leave it.

6. If manic or severely hypomanic, get rid of them now. Usually can stop abruptly.

7. If cycling get rid of them.

8. If mixed get rid of them.

9. I usually wait until the patient is doing better to much better. Trust is an issue. If the first thing we do is make them suffer more they will be unlikely to stay around long or even go to another psychiatrist. Even though we know the antidepressant is causing harm oft time the patient thinks either that the antidepressant is helping or that every time they try to go off they feel much worse. Waiting until they are better is usually a good thing.

10. Also waiting longer usually means that the patient is going to be more educated about bipolar in general.

11. If they are not getting better after several add on meds then slowly decrease.

12. There are more exceptions to the above rules than there are rules.

Guidelines on how to get off antidepressants

1. Educate/prepare the patient well ahead of time and repeatedly.

2. Chart GAF scores [a psychiatric standard, Global Assessment of Function, a single number summarizing how you’re doing] over time. Sometimes getting off anti-depressants isn’t the right thing to do and can be used to identify “Sweet Spot” for dosing. For example, I recently had a patient who was doing poorly on 300 Effexor XR started when she was still “unipolar”. Took two years to wean off. Retrospectively I was able to see that she was doing best around 75mg. Charting the GAF at appointments and the Lowest in between is best.

3. If the patient stops them AMA [against medical advice] abruptly and they are doing well then leave them off. Watch for manic symptoms. (Sometimes patients get better despite our best efforts.)

4. If the patient stops them AMA abruptly and they are doing worse don’t jump back up to the whole dose. The longer they were at the lower without feeling bad before felling worse, the lower dose you can return to. You can sometimes use half-lives to calculate this. Calculate the dose based on when they started feeling bad. Watch patients very closely during this time, even daily by phone or at the office.

5. Warn patients that they will have mood swings if they do this. Warn patients that they will have mood swings if they don’t do this, probably worse. Warn them of
this over and over again. The point is to try and stop them from major panic when they do have a down.

6. Slowly is best. The slower the better. I usually wait . . . at least 6 – 8 weeks between dosage decreases. Prozac/fluoxetine can be an exception to this.

7. Longer if anxiety is a major feature.

8. Faster if they feel better as they decrease dose.

9. Longer if they have difficulty with dosage decreases.

10.Longer if they are doing relatively well.

11.Never decrease before a major event or holiday.

12. Avoid decreasing during times of major stress.

13. The pt can take longer if they want to take longer for any reason.

14. Reduce in the smallest possible increments. As you approach zero then take the dose changes smaller or longer. Get out that pill cutter. If you can’t get dosage changes in small enough changes do every other day between the smaller dose and the larger dose. You would be surprised how often this works even on very short half-life drugs like Effexor XR.

15. You can go faster if they feel better as they decrease dose, but not too fast. Look for signs and symptoms of mania as well as depression. I have seen both hypomania and even mania in a [patient with Bipolar II] who stopped their antidepressant without taper. This has been reported in the literature as well. Going down slowly also avoids manic reactions.

Special Rules

1. Effexor XR. If the pt can tolerate doing this then this is by far the best way to do this. Open up the capsule and take one more bead out each day. Rules 11 – 13 of how to get off antidepressants apply. Pour the beads out on a creased piece of paper and count out the correct amount of beads. Then using the crease of the paper to get the beads back in the capsule. [In my town I have the advantage of a compounding pharmacist who can make small doses from the patient’s large doses and allow us to decrease]

2. If pts can’t count beads or don’t want to do this then take out about ¼ capsule for 6 – 8 weeks and repeat.

3. For any anti-depressant you can add in 20 mg of Prozac, get them off the anti-depressant, then taper the Prozac.

4. Prozac is a special case because of its long half-life. I generally will drop of one day at a time when reducing dose, e.g. decrease to 6/7 days a week for 6- 8 weeks then decrease to 5/7 days a week. Prozac is also a good candidate for every other day decreases, e.g. from a dose of 40mg a day go to 20 alternating with 40 mgs a day [to make a 30 mg-equivalent dose].

5. Every other day dosing of a medicine often work when reducing doses even when the pharmacology (“half life”) suggests it shouldn’t work. [I asked: Dr. Kelly has indeed used this technique for both duloxetine/Cymbalta and venlafaxine/Effexor, two of the trickier ones to taper. He says it works there too.]