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

Joseph F. Goldberg, MD, sat down with Henry Nasrallah, MD, to discuss the essentials of off-label prescribing and uses of polypharmacy. Nasrallah highlighted his landmark analysis showing that 88% of Diagnostic and Statistical Manual (DSM) psychiatric diagnoses have no US Food and Drug Administration (FDA)-approved medication, leaving clinicians needing to prescribe off-label for many patients.^1,2

Goldberg and Nasrallah both emphasized that off-label prescribing, when rationale-based and neuroscientifically informed, is responsible clinical care. Goldberg observed that "off-label practices are legitimate…good for patients and save a lot of lives," while prioritizing that prescribers must understand what a drug does in the brain, not merely blindly follow or completely ignore labels.

Below is their expert discussion, and a video version of the conversation is available here.

Joseph F. Goldberg, MD: Our guest today is a long-time friend and colleague of mine, Dr Henry Nasrallah. Many of you may know of Nasrallah; he's been in the field for a very long time. He's trained literally generations of psychiatrists over the course of time. Henry currently is Vice Chair for Faculty Development and Mentorship, and Professor of Psychiatry, Neurology, and Neuroscience at the University of Cincinnati College of Medicine. Henry, welcome to the program today.

I'm going to single out one topic, and this paper of yours from a few years back which is entitled, "An Analysis of the High Psychotropic Off-Label Use in Psychiatric Disorders: The Majority of Psychiatric Diagnoses Have No Approved Drug." Our topic today is going to be the off-label use of psychotropic medications and all the variations on that—what does it mean when a medicine is being used off-label? How does that comport with the scientific evidence base for it? How does on-label bear on proprietary branded compounds that a pharmaceutical manufacturer has a vested interest in for a very particular use? And what if you or I want to use it for something other than that? And if that's off-label, what does that mean, and how should patients think about this? So with that very broad backdrop, Henry, how should our audience can be thinking about this concept?

Henry Nasrallah, MD: Absolutely. This is a very important topic because off-label use of medications, proprietary medications that are approved by the US Food and Drug Administration (FDA), has been going on for decades. And the reason I embarked on that paper that you mentioned is because I was annoyed when a paper came out in Journal of Clinical Psychiatry about 15 years ago, implying that psychiatrists are acting inappropriately because 68% of all atypical antipsychotics were used off-label. So this fellow of mine and I, we went through every drug ever approved by the FDA since 1960, and we listed them right next to the DSM-IV-TR at the time, 300 categories, and matched them for what diagnoses have an approved drug.

And we ended up finding, as this paper shows, that 88% of psychiatric disorders do not have any approved drug—only 12% do. So what do you do with 88% of the patients who show up in your clinic and have nothing approved?

Take borderline personality as an example: you have to just to use medications that work on their symptoms, whatever symptoms they have, with depression, mania, brief psychotic episodes, etc. You use it judiciously, compassionately, and you treat them. You do not just send them away, saying "Sorry, I don't have any drug for you." This is true for other branches of medicine as well.

So essentially, off-label practice is essential for every practitioner who receives patients in their clinic, and our clinics are full of patients who are treatment resistant, who did not respond to the approved drug. You have to do something else, right? You cannot just let them suffer and, and eventually potentially harm themselves or harm other people. So this is why off-label practice is an important key ingredient of general psychiatric practice. Why do you think the pharmaceutical industry has not been so interested in that about 90% of disorders for which there are no indications? I think the real reason, in my opinion, is that they are not as prevalent and the pie is not big enough to bring in enough revenue.

Goldberg: So consider this, as far as I'm aware, there is 1 FDA-approved treatment for schizoaffective disorder, right? It is an atypical antipsychotic. We often use others, but technically, that is off-label. If I have diagnosed a patient with schizoaffective disorder, and I'm not using valbenazine (Ingrezza)—if I want to use anything else, even if it were clozapine—it is off-label. How do I reconcile the fact that I am borrowing from my scientific knowledge of how these drugs work, as opposed to saying, "Well, have you failed the one and only FDA-approved drug for this particular ailment?"

Nasrallah: You know, the FDA used to be much more broad: they would develop an antipsychotic for patients with any psychotic symptoms. But about 20 or 25 years ago, they decided, "No, drug companies can only develop a drug for schizophrenia. Schizoaffective is a different disorder, even though they have very similar symptoms.” Now this delays drug development because you have to do spend hundreds of millions and collect different kinds of psychosis, etc.

This is, I think, part of the problem, is that the FDA has several, maybe unintentional, obstructions to drug development that lead to off-label practice. For example, one of the things that I wish the FDA would change their mind about is, they are demanding that a drug company do a clinical trial, a lot of studies, and culminate a phase 3 clinical trial in a specific DSM diagnosis, which is an artificial construct developed by a group of people around a table in a committee. It has very little validity, frankly, in the real world. I mean, if the FDA allows a change in the indication to be about symptoms, then nothing will be off-label anymore because that's what forces us to use medications off-label; there are symptoms existing in a disorder that does not have any approved drug, but that symptom resembles another condition which does have a medication. This is how we end up using off-label, because it's rational.

Goldberg: You are getting at 2 interesting points here. One is dimensions of psychopathology, like you mentioned suicidality or maybe hallucinations or impulsivity, or some target symptom. Then the second is brain circuits. The FDA has put a toe in the water with at least some drugs, for instance in the autism space, when they will talk about a drug like risperidone to treat agitation or aggression in autism. And why don't they do more? Because you could talk about treating suicidal behavior in borderline personality disorder or you could treat micropsychosis or impulsive aggression, and these are transdiagnostic concepts. I wonder if the FDA is sort of hesitant to go farther in that direction, even though that is exactly what I am sure most of our clinician audience today are doing in their own practices, whether they realize it or not.

Nasrallah: Yes, and the FDA allows companies to have a secondary outcome. The primary outcome is, for say schizophrenia, is improvement in psychosis, right? But they allow them to have a secondary outcome, like negative symptom or cognitive deficits. Which is not a single symptom, but it is a category, a cluster of symptoms. So if the FDA would loosen the rules a little bit, the whole notion of off-label becomes moot because we are allowed to use a medication for a symptom.

Goldberg: And to help us think less concretely about these things, if I may, let's take clozapine and suicide, for example: it is the only drug that has an FDA labeling that indicates it can reduce the chance of suicidal behavior in schizophrenia. Now, if I am thinking concretely, I cannot take that information and say, "Oh, I should give clozapine to my suicidal depressed patient because the FDA identified an antisuicidal effect specifically in that category." One would have to be mindful of the fact that probably the suicidality in schizophrenia comes from psychosis, whereas the suicidality in depression comes probably from other things. It really requires the clinician to think about what is going on in the brain, as opposed to just saying, "what does the label say?"

Nasrallah: Well, thankfully the FDA allowed the manufacturer of clozapine to do a study called the InterSePT study. It was started in 1999 and culminated in 2003 with an approval. They focused on suicidality because there were observations that clozapine seemed to reduce the suicidality in patients with schizophrenia. So they did a 1-year study against olanzapine in terms of effectiveness, but focused only on suicidality, suicide attempts, suicidal thoughts, etc. And lo and behold, clozapine beat olanzapine at the end of the study, showing that it does have an effect on this symptom of suicidality.

It paid off, and now all of us realize we can use clozapine if our patient with schizophrenia is responding to one of the antipsychotics but is still suicidal. We should not only use it for treatment-resistant schizophrenia—we can add clozapine to patients who are responding to risperidone, olanzapine, aripiprazole, whatever, but they are still suicidal, we can add clozapine strictly for that symptom. So now we solve the problem of off-label treatment for suicide at least.

Goldberg: There are so many things we do in clinical psychopharmacology, I would venture to guess, without necessarily calling it out overtly or realizing it along these lines. If I am giving a patient valproic acid or lithium for their impulsive aggression, regardless of their diagnosis, and someone says, "Oh, but they don't have bipolar disorder," or if I'm looking at affective instability and I have the observation that a particular medicine is helpful for that, I'm going in the off-label territory because I'm treating a target symptom that may or may not necessarily fall under the rubric of the overarching diagnosis.

Nasrallah: You mentioned valproate (Depakote), bipolar, etc. We know that valproate works in bipolar disorder and reduces the manic symptoms, including especially the irritability, the anger, which is one of the hallmarks of bipolar. You have no idea how many patients I have in my neuropsychiatric clinic with traumatic brain injury, who are angry, impulsive, and agitated, and I use Depakote for them right and left. They don't have any seizures. They don't have mania but I use it specifically for that symptom, and with great results, I must say.

The patients themselves recognize it and thank me for it, because they become less hostile, less angry, and are getting along with people better. But we are doing it on our own, so it's called off-label.

Goldberg: And Henry, you are also doing it based on your knowledge of neuroscience. And I'm wondering if our audience may be unwittingly doing something similar, but thinking about it wrong. They might say, "I gave this angry, impulsive patient valproate, and they seem less angry. Does that mean they have bipolar disorder?" As opposed to Henry comes in and says, "Perhaps they have GABAergic dysfunction. Perhaps there's something in their limbic, bottom-up regulation that's overriding top-down,” and you're using something that will have a kind of a pro-inhibitory effect in the brain.

Nasrallah: You know, a few colleagues and I have been focusing on reinventing the DSM. The DSM has a lot of silos, with individual disorders arbitrarily and artificially defined, but it becomes the rule. But over the last 8 to 10 years, we now have ample evidence of the transdiagnostic model, which you mentioned earlier. The transdiagnostic model shows that 14 different psychiatric disorders share many genes. So do not be surprised if your bipolar patient comes with attention-deficit hyperactivity disorder (ADHD), or with panic disorder, or with obsessive-compulsive disorder, or with alcohol use. We really need to reinvent DSM to include all of those common comorbidities. We may eventually develop a drug that works across the board—a transdiagnostic drug.

Clozapine is a good example, by the way. It works in psychosis, but it also works in mania, and it works in aggression, and it works in suicidality, and it works in depression some people say. So this may be why a drug like clozapine, which has a lot of receptor binding, may actually be a transdiagnostic drug. We need more of that in the future.

Goldberg: You know, I'm so glad you said that because I want to encourage our audience to think about this very carefully. I think this is one of the most important things we do as practitioners. I would like to ask our audience to think back to your more difficult patients. How many have just a single diagnosis? They may have a primary diagnosis, but patients with serious mental illnesses usually have more than one thing going on.

So this is where you could say, is inattention by virtue of comorbid ADHD, or are these the cognitive symptoms of depression or mania? Or is anxiety a subcomponent of depression, like anxious distress, or is it a comorbid anxiety disorder? In the bipolar world that I study, about 75% of people with bipolar illness have a second psychiatric diagnosis, 40% have a third, 20% have 3 or more psychiatric diagnoses. So how can you talk about on-label, off-label, when you are treating 4 things?

Nasrallah: Exactly. That is exactly the point. I wish all clinicians would absorb that and realize that they are going to encounter a lot of so-called comorbidities. I call them coexisting—it is a really coexisting disorders. It is not a comorbidity, because they are being generated from the same genes, probably, that are shared by those disorders.

As you said, we have a lot of patients who are called treatment-resistant, and they have multiple disorders. Not to forget the Axis II, by the way. So when you look at the FDA studies, and you and I have done a lot of them, we basically collect patients that meet the diagnosis of X disorder, and they cannot have substance use, they cannot have diabetes, they cannot have thyroid dyskinesis, they cannot have this and that.

All of those inclusion and exclusion criteria, you end up doing a study on 300 or 400 patients who are really stable, without any medical or psychiatric comorbidities. And you, you do the study, and guess what? It beats placebo. The drug goes to market, approved for this condition. But then in the real world, the patients are messy. They are affected with multiple symptom, multiple disorders, and this drug may or may not work for them. So we have to end up using some off-label method to make those real-world patients respond to a drug that might work for the average patient.

Goldberg: Not to beat up on the FDA or pharma, however, you are right. The model that now exists is to capture the rarefied monodiagnosis patient, when, in fact, our colleagues in the clinical trenches are not seeing those people. So when the results that occur in their hands are so different from what happens in the rarefied monodiagnosis patient, they throw their arms up, they say, "This drug doesn't work," or, "There's something wrong with the trials." And yet, as you point out, the real-world patients are not the ones that get studied.

When we were doing the old studies of Depakote, Bob Hirschfeld, our late colleague, published an interesting paper in Psychopharmacology Bulletin. He looked at all the screen failures for the Depakote clinical trials for the FDA, and he found that over 90% of the patients who were screened with bipolar to be in the study were not eligible to be in the study. Now, just think about this: 90% of the people in the world who have this ailment are not representative, and so they did not get studied. Then when they come to you and to me and our colleagues, we don't know what to do with them. And, and we either have to go off-label, or we have to say, "The on-label data falls short."

Nasrallah: You know what also bothers me, Joe, is the different types of off-label. There's the baddest off-label in the eyes of the FDA, when you use a drug approved for condition X, and you use it for condition Y. That's off-label.

But what bothers me is another really innocuous off-label, which is using a medication above the approved limit in the FDA studies. Let's say, olanzapine as an example. I mean, when we did the studies with it, it came to the market with a dosage from 10 to 20 mg. Twenty was the highest dose used in the clinical trials, so FDA automatically makes it the upper limit. But in the real world, like we talked earlier, we have very difficult, complex, and severe patients who may need 25 and 30 mg, and we go up that way, and we end up getting some benefit. Even 35 or 40 mg have been published, actually, showing olanzapine at 40 mg may act almost like clozapine. But the FDA says, "Nope, and 20 is the upper limit." We are basically discovering the benefit of other uses for these drugs at higher doses that were not included in the FDA study.

Goldberg: And you make me think, neither the FDA nor any other entity points out things like, "If your patient is tolerating a medication without any meaningful adverse effects, and they're not getting better: A, consider whether a supratherapeutic dose might be better and B, check a blood level and see if they may be in the lower range. And if somebody happens to be, for example, an ultra-rapid metabolizer of a drug that's a substrate for an enzyme that's under pharmacogenetic variation, you, you may, you may be underdosing somebody without realizing it because they are at the lower end.

See the rest of this discussion on off-label prescribing here.

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. Devulapalli KK, Nasrallah HA. An analysis of the high psychotropic off-label use in psychiatric disorders. The majority of psychiatric diagnoses have no approved drug. Asian J Psychiatr. 2009;2(1):29-36.

2. Van Norman GA. Off-label use vs off-label marketing of drugs: part 1: off-label use-patient harms and prescriber responsibilities. JACC Basic Transl Sci. 2023;8(2):224-233.