Sepsis Cure Needs An RCT

Epistemic Status: Confident

Every now and then the news comes out with a totally clear-cut, dramatic example of an opportunity to do a lot of good. This is one of those times.

The story began in January, 2016, when Dr. Paul Marik was running the intensive care unit at Sentara Norfolk General Hospital. A 48-year-old woman came in with a severe case of sepsis — inflammation frequently triggered by an overwhelming infection.

“Her kidneys weren’t working. Her lungs weren’t working. She was going to die,” Marik said. “In a situation like this, you start thinking out of the box.”

Marik had recently read a study by researchers at Virginia Commonwealth University in Richmond. Dr. Berry Fowler and his colleagues had shown some moderate success in treating people who had sepsis with intravenous vitamin C.

Marik decided to give it a try. He added in a low dose of corticosteroids, which are sometimes used to treat sepsis, along with a bit of another vitamin, thiamine. His desperately ill patient got an infusion of this mixture.

“I was expecting the next morning when I came to work she would be dead,” Marik said.”But when I walked in the next morning, I got the shock of my life.”

The patient was well on the road to recovery.

Marik tried this treatment with the next two sepsis patients he encountered, and was similarly surprised. So he started treating his sepsis patients regularly with the vitamin and steroid infusion.

After he’d treated 50 patients, he decided to write up his results. As he described it in Chest, only four of those 47 patients died in the hospital — and all the deaths were from their underlying diseases, not from sepsis. For comparison, he looked back at 47 patients the hospital had treated before he tried the vitamin C infusion and found that 19 had died in the hospital.

This is not the standard way to evaluate a potential new treatment. Ordinarily, the potential treatment would be tested head to head with a placebo or standard treatment, and neither the doctors nor the patients would know who in the study was getting the new therapy.

But the results were so stunning, Marik decided that from that point on he would treat all his sepsis patients with the vitamin C infusion. So far, he’s treated about 150 patients, and only one has died of sepsis, he said.

That’s a phenomenal claim, considering that of the million Americans a year who get sepsis, about 300,000 die.

Sepsis is a really big deal. More people die from sepsis every year than from diabetes and COPD combined. Ten thousand people die of sepsis every day.  A lot of these cases are from pneumonia in elderly people, or hospital-acquired infections.  Curing sepsis would put a meaningful dent in the kind of hell that hospital-bound old people experience, that Scott described in Who By Very Slow Decay.

Sepsis is the destructive form of an immune response to infection. Normally the infection is managed with antibiotics, but the immune response still kills 30% of patients.  Corticosteroids, which reduce the immune response, and vitamin C, which reduces blood vessel permeability so that organs are less susceptible to pro-inflammatory signals, can treat the immune response itself.

Low-dose corticosteroids have been found to significantly reduce mortality in sepsis elsewhere in controlled studies (see e.g. here, here, here) and there’s some animal evidence that vitamin C can reduce mortality in sepsis (see here).

This treatment seems to work extraordinarily well in Marik’s retrospective study; it is made of simple, cheap, well-studied drugs with a fairly straightforward mechanism of action; the individual components seem to work somewhat on sepsis too.  In other words, it’s about as good evidence as you can get, before doing a randomized controlled trial.

But, of course, before you can start treating patients with it, you need an RCT.

I wrote Dr. Marik and asked him what the current status of the trials is; he’s got leads at several hospitals: “two in CA, one at Harvard, and one in RI. In addition the Veterinary University of Georgia is proposing a neat study in horses — horses are at increased risk of sepsis.”

But he needs funding.

Medical research does not progress by default. The world is full of treatments that one doctor has tried to great success, which never went through clinical trials, and so we’ll never know how many lives could have been saved.  Some of the best scientists in the world are chronically underfunded. The world has not solved this coordination problem.

By default, things fall apart and never get fixed. They only get better if we act.

You can click on this Google Form to give me estimates of how much you’d be willing to donate and your contact information; once I get a sense of what’s possible, my next step will be coordinating with Dr. Marik and finding a good vehicle for where to send donations.

(I don’t have any personal connection to Dr. Marik or to the treatment; I literally just think it’s a good thing to do.)


20 thoughts on “Sepsis Cure Needs An RCT

  1. As a total outsider to medical research, I feel nervous about believing this since I do get the impression there’s also a lot of promising sounding, but not actually promising things out there. Obviously this particular thing gets a more cred because you seem to know your shit, but I’m not sure what stance to take to these things in general. So I guess I have some questions:

    Why hasn’t the world fixed the coordination problem for simple, promising treatments like this is claiming to be? I know the usual story is that there’s no money in it, so pharma companies won’t run studies. That makes sense as far as it goes, but doesn’t anyone want the *credit*? Universities? NIH? Big hospitals? Seems like there’s a lot of money for medical research in general. I can understand that lots of money gets diverted to not very promising things, but if this is really *so* promising, won’t people want to be associated with it? Is it that RCTs are ridiculously expensive enough that people can’t afford to do that?

    • The NIH would take more than a year to give a decision on the study. It’s not that the study definitely won’t happen, it’s that it takes time — too much time.

      • The reasons for why this is an unsolved problem boil down to public choice theory. Medical research is (directly and indirectly) government subsidized and regulated. That means established players have a monopoly, new entrants in the field have a hard time breaking in, bureaucracy expands to fill (and exceed) the funding available, and performance-related incentives are weak. It’s not that every promising new treatment will fail to be fully explored and funded, of course — we *do* have recent improvements in medical care, like statins and much better trauma care. But things move slowly and things fall through the cracks.

        It’s sort of like the way you don’t really expect your local transportation infrastructure to improve, you expect train service to be rolled back and interruptions to get gradually more frequent, and the 2nd avenue line will never be built. Maybe medicine is actually a little better than trains, I’m not sure. But it’s flawed for similar reasons.

        In the mid-20th century, it seemed like the government was actually pretty good at providing public goods. But now, not so much. It costs more to do less good stuff. My hypothesis is that institutions have a life cycle, and they naturally decline at some point, and we’re seeing that now.

  2. I have nothing I can give. But you could maybe get in touch with SENS or groups like that? In fact, I can share this with a facebook group on that subject right now.

  3. I can’t fill the form out now, but will later today. I just wanted to say that although I generally like and value most of the content on this blog, I view posts like this as being of exceptionally high value. The chance to leverage your expertise to a real advantage in doing good is a great opportunity. Thanks.

  4. Have you reached out to GiveWell or their sibling organization Open Philanthropy Project? I’d expect them to take longer than your ad-hoc Kickstarter too, but maybe they could surprise me.

    • I don’t have the best relationship with them, but if anybody is interested in pitching to them, I’d be happy about that.

  5. I did a Google News search pre-2016 for some context on Dr Marik.

    He’s sort of a big deal – not Atul Gawande famous, but well-known enough that if he had a history of fraud, probably someone would have noticed already.

    He’s done other work on reducing hospital infections, e.g. by promoting copper-infused linens. Overall it seems like he’s tried a lot of reasonable-sounding things, so this isn’t just some random thing he tried. But presumably there were lots of attempts that didn’t produce such dramatic results, so we should expect some substantial regression to the mean due to publication bias.

    On the other hand, I don’t think that anywhere near overcomes the huge observed effect size for this sepsis cure. Even though there wasn’t proper random assignment yet, the implied significance level is pretty darn good.

    Other things Dr Marik is known for include probably industry-funded research assessing the benefits of medical devices, and public skepticism of ICU treatment for the elderly.

  6. I think this would benefit from clarifying how much potential benefit Marik’s protocol (Vitamin C + corticosteroids) has over Fowler’s (Vitamin C only), which according to that NPR article is currently in a Phase II trial, funded by a $3.2M NIH grant.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s