The FDA Has Punted Decisions About Luvox Prescription To The Deepest Recesses Of The Human Soul
Here’s my pitch for fluvoxamine (Luvox) for COVID.
In the midst of all the hype about ivermectin and hydroxychloroquine, scientists put together the giant 4,000-person TOGETHER trial, intended to test all these exciting COVID early treatments. You know what happened next: ivermectin and hydroxychloroquine crashed and burned.
But a different drug, the SSRI antidepressant fluvoxamine, actually did really well! It decreased COVID hospitalizations by about 30% - not the perfect cure rate the rumors attributed to ivermectin, but a substantial decrease. Given the size and professionalism of this study, and another smaller one that also got positive results, I and many others take Luvox pretty seriously. At this point I’d give it 60-40 it works.
Can you prescribe a medication when you’re only 60% confident in it? There’s some thorny philosophical issues around this, but I think in the end you have to compare risks and benefits.
What are the risks? Like every medication, including Tylenol, aspirin, etc, Luvox has some common minor side effects and some rare major ones. But let’s step back a second. Fluvoxamine is a bog-standard SSRI. Its side effects are generic SSRI side effects. We give SSRIs to 30 million people a year, or about 10% of all Americans. As a psychiatrist, I’m not supposed to say flippant things like “we give SSRIs out like candy”. We do careful risk-benefit analysis and when appropriate we screen patients for various risk factors. But after we do all that stuff, we give them to 10% of Americans, compared to 12% of Americans who got candy last Halloween. So you can draw your own conclusion about how severe we think the risks are.
For some reason the same experts who don’t mind prescribing SSRIs when people have mild depression freak out about prescribing them when they’re the only evidence-based oral medication for a deadly global pandemic. “What about SSRI withdrawal?”, they ask. After a ten day course? On 100 mg imipramine-equivalent dose? Minimal. “What about long QT syndrome?” The VA system took 35,000 high-risk older patients off of an unusually-likely-to-cause-QT-syndrome SSRI in 2011, and were unable to find any evidence that this prevented even a single case of the syndrome, let alone any negative outcome!
The objection I take most seriously is actually the worry about post-SSRI sexual dysfunction, a very rare condition where people on an SSRI can have sexual problems for months or years after they come off. I would be shocked if you could get that from a ten-day course. But technically nobody has ever tested this - there’s never been a good reason to put someone on an SSRI for only ten days before - so I can’t rule it out. Still, the risk from adding a few extra Luvox prescriptions for COVID is still much less than the risk we incur all the time from having 10% of Americans on SSRIs for years at a stretch, so this seems like a weird time to get cold feet.
I conclude that the risk-benefit calculation probably favors using Luvox. And I’m not alone here. Johns Hopkins University’s COVID treatment guidelines recommend fluvoxamine for appropriate COVID patients. Some leading psychiatrists, especially the Washington University psychiatrists who helped discover the new indication, support fluvoxamine for appropriate COVID patients. Many of the epidemiologists and statisticians most instrumental in debunking the hype around ivermectin have spoken out in favor of fluvoxamine, saying this one is the real deal (1, 2). The National Institute of Health hasn’t quite come out in support, but they have taken the unusual step of not disrecommending fluvoxamine the same as they disrecommend every other oral early COVID treatment, saying that the evidence "provides the sort of flexibility for the treating clinician to go either way".
Unfortunately, none of these bodies alone or combined are powerful enough to make the average doctor prescribe differently. That’s why all eyes are on the FDA.
The FDA has a weird role here.
They already approved fluvoxamine as an antidepressant. That means it’s legal, pharma companies can make it, pharmacies can stock it, and individual doctors can prescribe it whenever they want, including for COVID.
But they approved it with a label saying “For Depression”. Doctors are kind of . . . well, “hidebound” is a harsh word, but they really hate doing weird new things that no one has explicitly given them permission for. It’s not illegal to prescribe fluvoxamine for COVID. It’s not even going to get you in any trouble. It might not get covered by insurance, but it only costs about $10 anyway. The problem is just that it’s weird.
So in order to make doctors feel completely comfortable prescribing it, the FDA would have to add “…And For COVID” to the label. The scientists involved in the big study have asked them to do this.
I hoped that the FDA would say “Since the COVID pandemic is an emergency, we’ll do this right away”.
I predicted they would say “Please give us a year to figure out our opinion on this.”
I feared they would say “There’s just not enough evidence”.
What I never imagined at all was their actual response, which was “Sorry, we don’t understand our own bureaucracy well enough to figure out how to do this.”
But according to Kelsey Piper at Vox, that’s where they are right now:
[Professor Ed] Mills, who thinks that fluvoxamine and budesonide are both appropriate to prescribe to patients sick with Covid-19, compares public messaging on fluvoxamine to communications about Merck’s drug molnupiravir. The evidence for molnupiravir is in many ways weaker than the evidence for fluvoxamine, but molnupiravir was produced by a major pharmaceutical company that can shepherd it through the process of becoming a recommended drug. On a call last week, Mills said, the FDA told him “they don’t know how to deal with submissions where there isn’t someone to be responsible for it.”
That is, FDA procedures usually assume there is a pharma company sponsoring a drug. But fluvoxamine is cheap and off-patent and no pharma company is involved in repurposing it for COVID. Nobody has a procedure for a drug without a sponsor, so they won’t do anything.
Kelsey’s article focuses on the systemic failure: the FDA, guideline-making agencies, and public health communicators have dropped the ball on this. I think that’s a perfectly fine thing to focus on. I’m not usually one to defend the FDA, and their actions here hardly seem defensible. I’m with Kelsey in hoping they find a way to solve their institutional dysfunction.
But I can’t help wondering if this is entirely on the FDA. Fluvoxamine is legal. The only reason we need the FDA to get involved here at all is because if it’s not on the label, doctors will feel uncomfortable prescribing it.
What if, in order to save thousands of lives and help beat back a global pandemic, doctors just did the uncomfortable thing?
Am I being harsh in saying that the problem is doctors who don’t want to do something uncomfortable?
There are many reasons not to prescribe a medication for a new indication. Maybe you genuinely think the risks outweigh the benefits. If it were me, I would trust the Johns Hopkins guidelines team on this, but honest opinions can differ. I have no problem with doctors who are holding off for this reason, and look forward to arguing with them in the appropriate venues.
Or maybe you’re afraid of lawsuits. If you get sued for malpractice, it’s nice to be able to tell the jury “it says this drug is okay for this condition right on the label”. But this doesn’t usually stop doctors from doing off-label prescriptions. Gabapentin is the 18th most-prescribed drug in the US, almost always for nerve pain or anxiety, but its label only officially endorses use for seizures or shingles. Beta-blockers for social anxiety? Off-label and dirt common. Prazosin for PTSD nightmares? Off-label and dirt-common. How do doctors sleep at night, knowing they’re constantly at risk of getting sued for off-label prescriptions? Probably using trazodone, the #2 most popular sleeping pill in the US, whose label says it should only be used for depression. No, seriously, it’s because most doctors don’t even know these indications are off-label, plus their medical school professors all did it too so it doesn’t feel transgressive.
Or maybe you suspect the benefits outweigh the risks, but you have a principled heuristic of not trusting your own suspicions. Doctors are constantly meddling with systems we don’t fully understand, people die when we make mistakes, and hordes of scammers and profiteers are trying to exploit us at any given moment. “Never do anything that five government bodies haven’t enthusiastically recommended” is a great meta-level heuristic for staying sane in that environment, and one which I follow 95% of the time. If someone else follows it 99.9% or 100% of the time, and even a Johns Hopkins endorsement isn’t enough of a recommendation for them, I can understand that.
Or maybe you’re a coward.
I’m not saying doctors are generically cowards. My father is a doctor and he’s one of the bravest people I know. Every time there’s a typhoon or an earthquake in some terrorist-infested country on the other side of the world, he hops on a plane to go there and treat victims, sometimes before the rubble is even cold. If this was something simple, like treating river-blindness in war-torn parts of the Congo or containing an Ebola epidemic in Nigeria, I’m sure doctors would be all over it.
But the Devil knows the weaknesses that lurk in the hearts of men. When he wants to scare off doctors, he doesn’t threaten us with insanely hard acts of self-sacrifice. He knows we love that stuff! He threatens us with the prospect of looking slightly weird in front of our colleagues.
Here is a doctor who, if nominative determinism is any guide, knows a thing or two about diabolic temptation. Yet he talks about how he “almost felt dirty” prescribing fluvoxamine, even though he knew there was strong evidence supporting it. He worried the nurses were making fun of him (protip: if you are a doctor, the nurses are always making fun of you). He made the right decision in the end, but I wonder how many doctors in similar situations don’t.
There are lots of reasons to feel nervous and awkward when you prescribe a medication your colleagues won’t. Maybe they think you’re a loose cannon who doesn’t care about evidence. Maybe they they think you’re defecting against the team and going to get everyone in trouble. Maybe they’re remembering the ivermectin debacle and wondering if you secretly prescribe horse dewormer and vote Trump. Maybe they think you hold them in contempt for not being as up on the literature as you are, and only prescribing normal stuff.
You should always consider your colleagues’ opinions insofar as they are good smart people and you want to use their expertise as a check on your own fallible mind. But it’s hard to keep that separate from considering your colleagues’ opinions in the sense where it would be socially awkward to disagree. And that’s how the Devil gets us.
I faced the Devil last year and lost. In March 2020, when everyone was freaking out about ventilator supply, a team of very smart engineers asked me to prescribe them a medical-grade oxygen concentrator. I can’t remember the details, but something about trying to tinker around with a bunch of cheaper machines and jury-rig a budget ventilator, which they could pitch to people as a solution to the ventilator shortage. I punted. I said that this wasn’t really what the prescription system was for, you can’t prescribe things to healthy people just so they can tinker with them, and I might get in trouble with my clinic or the government or somebody. I told them to try to go through the proper channels for obtaining medical equipment, even though I was unsure whether those channels existed, and doubtful they would move with appropriate urgency.
Later I thought about this, and realized the choice before me was “You can contribute to a desperately important project that might save thousands of lives, but only if you do something kind of weird that might get you in a tiny amount of trouble”, and I had said no. Devil 1, Scott 0.
I faced the Devil the year before that, and I . . . well, let’s say it was a tie. I had a bunch of patients with treatment-resistant depression. Everyone knew ketamine was great for treatment resistant depression. But the only people using it were anaesthesiologists giving it IV, which was inconvenient and unaffordable for most patients. The FDA was trialing a new version of ketamine that could be given by psychiatrists via inhaler, and there was no reason whatsoever to think this wouldn’t work with normal ketamine, but nobody I knew was doing it and they all thought it seemed kind of weird. My severely depressed patients kept asking me for ketamine, and I kept saying “Sorry, I can’t prescribe that to you”, secretly ending the sentence with “…unless I use this one weird loophole I’ve never heard of anyone else using”. Finally I called up a compounding pharmacy near me and asked if anybody knew about this, and they said they knew a doctor who did, and did I want his phone number? I talked to him, and he said he’d been doing this for years and it had always gone well. For some reason, knowing that someone else was doing it was the permission I needed, I prescribed it to my patients, and it went well (I’ve since written up a guide for others). But I still didn’t have the courage to do the weird thing without knowing other people were doing it first.
(When I finally got around to prescribing ketamine, one of my patients told me I’d given her her life back. Usually I love hearing that kind of thing. This time it was bittersweet, because I knew I could have given more patients their lives back if I’d done it earlier. There are a couple of people who had six months of terrible depression that I maybe could have prevented if I had more courage. That’s partly on the FDA for making poor decisions such that optimal treatment required virtue on the part of individual doctors. But mostly it’s on me, for not having it.)
I will face the Devil in the future and I’ll fail again. Medicine is too big and complicated and scary to stray from the herd most of the time, and the sort of person who never fails at this problem is probably crazy, and constantly gives his patients snake oil or ivermectin or whatever. Doctors should generally stay within their area of expertise and doubt any argument leading them away from consensus. Certainly if you’re my patient and you somehow find this essay and read it back to me and tell me I need to prescribe you the latest whatever, I’m going to nope out of whatever you’re offering (especially if opioids are involved).
Still, I do want to stress the “facing the Devil” aspect, where this is a difficult moral battle. I know that’s a weird way to frame a prescription decision. But CS Lewis is a leading expert on devils and he was very clear that moral battles generally don’t happen in war-torn parts of the Congo. They happen in ordinary decisions about whether to do slightly unusual things that we worry might affect our social status among people we respect.
(by the way, when the other psychiatrists in my clinic learned I was prescribing intranasal racemic ketamine, they all said that was cool, and a few asked me to walk them through the process).
So if you haven’t been giving fluvoxamine to patients, please take a second, sit down, and decide whether it’s because:
You honestly think the risks outweigh the benefits.
You’re trying to follow some complicated meta-level heuristic that you need in order to practice good medicine or at least stay sane.
If it’s 1 or 2, you’re valid and I support you. If it’s 3, man up and write the prescription.
If you’d feel happier doing this after you talked to a psychiatrist who has experience with this medication, feel free to email me at scott[at]slatestarcodex[dot]com. I have no personal experience using it against COVID, but I can direct you to the studies and protocols that explain how.