Rational Off-Label Prescribing: A Review With Joseph F. Goldberg, MD, and Henry Nasrallah, MD, Part 2

Joseph F. Goldberg, MD, sat down with Henry Nasrallah, MD, to discuss the essentials of off-label prescribing and uses of polypharmacy. Goldberg and Nasrallah both emphasized that off-label prescribing, when rationale-based and neuroscientifically informed, is responsible clinical care.

See part 1 of this discussion here, and more from the series “Brain Trust: Conversations in Psychopharmacology” here.

Henry Nasrallah, MD: You know, speak of underdosing, here's another example. I like to use the low dose of lumateperone (Caplyta)—it comes in 10.5 and 21 mg. The company developed them for patients who have an inhibitor, so you are going to start with a lower dose. But I find that they are very patient-friendly to start, Caplyta at 10.5 for a couple of days and 21 for a couple of days, and then go to the standard dose of 42 mg, which is what the FDA approved.¹ And guess what? It's considered off-label. The company will not give me samples of10.5 and 21 mg to use because it's off label to them, but it is off label in a very innocuous way. You are using a low dose initially, which I think should be used in most psychiatric disorders. So titrate slowly to enter the brain of the patient, so that you do not shock them with a full dose on day 1.

Joseph F. Goldberg, MD: I think one can certainly be cognizant that the idea of tolerability issues can quickly clobber a trial. So the patient takes a medicine, and on day 1 or day 2, they have a headache or they are dizzy, and they do not want to continue. Even if the target dose is here, the language of the label does not really say things like, "You are welcome to use your clinical judgment and graduate up to that target dose." Now, if along the way, someone is benefiting at a lower dose, there is no law that says you must go higher; if they are not benefiting, but they are tolerating it, you are now free to continue to the dose that was shown in the clinical trials, which may have been the dose that was studied. But we do not necessarily know if other doses might work better.

If my patient is on lithium with a level of 1.1 and they are still symptomatic, I do not want to push that one too high. If someone has a very high margin for concern about safety, the good prescriber understands safety first, but also recognizes what is the reasonable balance of safety and efficacy.

Nasrallah: You reminded me of another very common off-label practice by many clinicians around the country, where they combine 2 or 3 antipsychotics, which is generally considered bad polypharmacy, right? Because we do not have any studies to prove what you are doing is right. But those clinicians are trying to get their patient to respond, so they start with, let's say risperidone, and they go to a pretty good dose, and the patient's psychosis continues.

So they say, "Why don't I add some ziprasidone," for instance. Well, after 2 weeks, the patient looks better. I think what they are doing is they are just increasing the amount of dopamine blockade in a difficult patient, where the normal doses do not work, and they may need a higher dose. But there are no studies. Not a single drug company has bothered to do a combination therapy, let's say antipsychotic A and B, and for or antipsychotic A, B, and C together. Thirty-eight percent of US patients with schizophrenia are receiving 2 or more antipsychotics, and that is considered polypharmacy. But there are some studies showing that maybe polypharmacy can actually be helpful.

I wish the companies would do a large, double-blind, placebo-controlled study comparing drug A alone by itself, with another group getting A and B, and with another group getting B only; this may show that the A plus B was actually better than both of A and B. No need for placebo. But nobody wants to put money to show that this might be helpful to a subset of patients that we encounter every day.

Goldberg: Experts like yourself, when you think about combinations and sort of using the materials that you have available, you will think about what makes sense.

You know, it would make sense probably to you to augment clozapine with, say, aripiprazole as opposed to quetiapine, right? Because you are just getting more antihistamine, anticholinergic effects and you are not really doing anything unique, whereas a partial agonist makes sense. You can build a rationale for that. Now, there are reasons why we are not going to see that study of polypharmacy comparison being adequately powered. One is, I very much doubt the aripiprazole people, if they were still around, would want to show the limitations of the drug alone. And the second, what happens once the drug goes generic? We are not going to see any more studies with aripiprazole or with lithium or with clozapine, or anything that is adequately powered in the modern era.

I have this gripe when it comes to antidepressants, because the only antidepressants that are getting studied at all nowadays are not tricyclics, are not selective serotonin reuptake inhibitors (SSRIs), are not even generic selective norepinephrine reuptake inhibitors, they are newer things. And so the current generation, is no going to have a way to know if a drug like venlafaxine could work in OCD at high doses, because no one is going to study that now.

Nasrallah: That's right. Here is another issue that came to mind: 30% of patients with schizophrenia will not respond to any dopamine antagonist. So we have about 30 or 35% who need clozapine, right? But many clinicians do not use it, unfortunately—there are all kinds of barriers. It is labor-intensive, or the patient does not want it, too many blood draws and numerous side effects, even though it can be a lifesaver for many patients with treatment-resistant schizophrenia.

It's like my patient, Bethany, with whom I established the CURESZ Foundation. She was psychotic for around 5 years, dropped out of college, became homeless, was eventually hospitalized and treated with 5 different atypicals, nothing worked. And I said, "What are you guys waiting for?" and I started her on clozapine—within 8 or 9 weeks she was back to normal.

Right now, we do a lot of off-label. I do a lot of off-label treatment for patients who fail clozapine, which is almost 50%, but I've salvaged almost 47%. What we do right now is we send them to electroconvulsive therapy (ECT). And it does not work on everybody, but it really helps another 50%. So what I do off-label, based on the literature, there are case reports of lamotrigine. Tiihonen in Finland wrote a meta-analysis where 5 studies showed improvement in patients who failed clozapine.²

Allopurinol is used for a completely different medical condition, and there are several reports showing that it can actually help patients with clozapine resistance. But nobody is bothering to do FDA-type studies in order to get them to be approved for this indication. And clinicians may not be completely unaware that those options available, but they are off-label, and many clinicians may want to use them. We should never, ever give up on our patients with schizophrenia, as an example, or severe mania or severe depression, same thing with a TRD and so on.

Goldberg: And then we do get limited by coverage. So here is a bugaboo of mine, for example with intravenous ketamine. It was first studied in bipolar depression, where it worked very nicely. Nobody got manic, but then when the manufacturer of intranasal esketamine did their registration trials, they chose to study unipolar major depression, not bipolar patients.

So technically, if I wish to give that molecule to a patient and have it covered by insurance, they need to have unipolar depression, not bipolar depression. And it is not as if to say that it is a bad idea to use that molecule for bipolar depression, but it is off-label, which one implication is it will not be covered by insurance. What should we tell our colleagues who have to sort of wrestle with that coverage issue? Or worse, to have the insurance company say, "Oh, this is not approved for that ailment," as if to make them think that it is a bad idea to try it?

Nasrallah: That is one of the biggest stumbling blocks for the off-label issue, that insurance companies deny the usage of a medication that may help the patient off-label simply because it is off-label.

I had a patient who failed every antidepressant on the market, failed ketamine, failed ECT, failed TMS, everything, and the only thing that worked for her, by chance was modafinil. Not 250, not 500 mg. Once she got to 1000 milligrams a day, she woke up. Her depression was gone. Now, no company would approve it every time I prescribed it. It is completely off-label. But, you know, I saved this woman's life by calling the drug company and explaining to them that her mother committed suicide, and she had a high risk for suicide. This is the only thing that worked. They do not want to approve it because they want to save money, basically. So do not forget the insurance company can be a monkey wrench in the off-label practice.

Goldberg: It's funny you mention that example. I had a somewhat similar patient once with bipolar depression, and, Mark Frye's study of modafinil in bipolar depression just come out, so I wanted to try it because I, too, had a resistant patient. and back then, I think you could get samples, and so I gave this fellow samples. He was doing terrific, but then the samples came to an end, the drug went generic, and then when I tried to get it covered, not only did they not approve it (he'd been stable for years on this drug now), but they sent him a letter saying something to the effect of, "Your doctor is giving you a drug that has not been shown to work." And the language sort of aroused concern on the patient's part, that there was something not just off-label but sinister going on.

And the patient looked at me and said, "I don't understand. I'm better. I've been better, and they sent me a letter saying this has been shown to not work." And you have to explain to everybody involved the difference between something that is evidence-based, that has been shown to work, and then how the nuances of how coverage works.

Nasrallah: What we are ending up recognizing is that off-label practices are legitimate. They are actually good for patients and it saves a lot of lives, preventing deterioration and chronicity and suicide. You discover that it works, and the patient stabilizes, and they feel good, and then the insurance company refuses to give it to the patient.

And we are coming full circle here, that there are a whole bunch of patients in psychiatric clinics with no approved drug, and yet we discover by chance. And this is where I love clinicians who end up writing papers. There is so much wealth of knowledge by ordinary clinical psychiatrists who have a lot of challenging patients and they try trial and error. They find that something works, but they never tell anybody. Unless you write it up as a case report or letter to the editor and share it with others, who will then say, “I had something like that, too." And pretty soon, you have enough signals from a lot of clinicians that drug companies bring in their big capital and invest in a clinical trial. That is the only way we can start reducing the 88% who have nothing approved, because somewhere in the clinical arena, clinicians are using something for those patients that works, but it never sees the light of day because they do not disseminate it.

Goldberg: Now, in fairness to the FDA, one of their missions is to protect from harm, and too let it not become the wild West, where you can do anything you want without cognizance of what the potential damages might be. For instance, imagine that I decided I am going to treat Alzheimer dementia with very high dose amphetamine, because in my opinion, I think that is a good idea. And it turns out, not only is that off-label, it is not evidence-based, and you could run into other problems and get cardiovascular effects. So there we get into the dilemma that it is not just an off-label use—is it a wise use? Is it a sensible use?

And I humbly would say, here is where I think we cannot necessarily count on the FDA to protect everyone from harm, but that the prescriber really has to understand what the drug is doing. And that is a big ask, because it does not just mean I follow the label. It does not just mean I know how the drug is prescribed. It means I understand what it is doing in the brain, both good and bad.

Now, one other example, I had a patient once who was floridly manic, and the prescriber decided they were going to orally load them with lamotrigine because it is a mood stabilizer, sort of oblivious to the fact that lamotrigine has no mania properties. And the patient ended up getting Stevens-Johnson syndrome, and their mania didn't get better. And the poor clinician was saying, "Well, but the patient was so desperately ill, I did something off-label." Well, you did something off-label, but it also was not very logical, because that is not the right tool for that job. So it was not about being off-label, it was about rationale.

As we are winding down here, maybe you could share a few thoughts of wisdom with our audience about how to think of what is logical in off-label prescribing?

Nasrallah:

Off-label practice, in my opinion, is a creative process, and it is done by clinicians, not by researchers. This is, to me, the cutting edge of scientific advances.

If only they share it, though, if they only publish it, write it up, present it at a meeting. It is the moving edge of psychopharmacology, when we discover off-label practices that eventually become on-label once a clinical trial is done. But first you have to show a signal and share it. Many clinicians say they are too busy to write things up. I wish they would, because some of the best ideas come from clinicians, and then they can partner with the researchers. We need both the clinicians and the researchers to address this issue of off-label prescribing.

Goldberg: You raise a really good point, and I'm not sure how many of our practitioners are thinking about this, but if you discover (like you described with the modafinil patient who did well at a high dose), write it up, send a letter to the editor, and you can present it as a hypothesis, and that will inspire other people, not necessarily to go and copy you, but to say "Hmm, maybe something's going on here." Because so much of psychopharmacology by history is serendipitous discovery, right?

Nasrallah: You know, I want to wrap up by giving you a nonpsychopharmacology tidbit here. Psychiatric disorders show the same deficit in their, on their brain, MRI scans. Three regions, the same 3 regions are abnormal in all those disorders: anterior cingulate, right insula, and left insula. Those 3 are abnormal across psychiatric disorders, which tells you something about the brain, in that they all share the same neurobiology, so no wonder they might share the same psychopharmacology. This is a clue that there must be a way to treat, all patients with 1 drug or 2 drugs that are approved for all of them. It is going to take a lot of off-label research to do that.

Goldberg: I think as long as the prescriber is asking the right questions, being cognizant, paying attention to efficacy and tolerability, and making some determination, "Is this helpful or not?" That is really evidence-based practice.

Nasrallah: Yes. Well, speaking of tolerability, though, a meta-analysis was done where they looked at the combination therapy in patients who did not respond to one antidepressant. They looked at all the studies where they were augmented with Wellbutrin or aripiprazole or another antidepressant, and they found out that the best outcome for patients who don't respond to one antidepressant is not to add, let's say, Wellbutrin to SSRI.

They found that adding mirtazapine, Remeron to a reuptake inhibitor, whether SSRI or SNRI, adding mirtazapine (which is an alpha-2 antagonist, a different mechanism completely) not only enhances the efficacy, but it also offsets the side effects. So you have double benefit: more efficacy, which is what you're looking for but patient tolerability also improves. This is the kind of studies that, that help guide clinicians. It's off-label to, to use mirtazapine with Prozac or with Zoloft. But when they did the the meta-analysis of studies that did off-label, these clinicians are trying things here and there, and they write it up and publish it, then we find that there is a signal that's worth disseminating and adopting.

Goldberg: It's rationale-based, and it is asking the prescriber to think about what they are doing, which is what we hopefully all should be doing. Well, I think that's a good note to end on. Henry, I wanna thank you so much for all your insights and, and your wisdom.

Nasrallah: My pleasure. It's fun to talk about things like this. Off-label is a controversial issue, but it really makes a lot of sense.

Goldberg: It really brings the prescriber to have to think about what they are doing, which is hopefully what we are doing anyways. I think you have given us a lot to think about in terms of being mindful about one's practice, which is a goal. I want to thank all of you for joining us today. I hope you found this program helpful, and we'll be back soon with our series, “Brain Trust: Conversations in Psychopharmacology.”

Dr Goldberg is a clinical professor of psychiatry at The Icahn School of Medicine at Mount Sinai in New York, NY and the immediate-past president of the American Society of Clinical Psychopharmacology.

Dr Nasrallah is a professor of psychiatry, neurology, and neuroscience at the University of Cincinnati College of Medicine. He cofounded the Schizophrenia International Research Society and the organization Comprehensive Understanding via Research and Education into Schizophrenia.

References

1. Caplyta label. US Food and Drug Administration. Accessed April 28, 2026. https://www.accessdata.fda.gov/scripts/cder/daf/index.cfm?event=overview.process&ApplNo=209500

2. Tiihonen J, Mittendorfer-Rutz E, Majak M, et al. Real-world effectiveness of antipsychotic treatments in a nationwide cohort of 29,823 patients with schizophrenia. JAMA Psychiatry. 2017;74(7):686-693.