'Tantalizing' Results For A Test Of Vitamin C For Sepsis
Could the leading cause of death in hospitals be curtailed with a safe and inexpensive treatment that includes vitamin C?
That question has been on the minds of intensive care doctors since the spring of 2017, when a well-known physician asserted that he was saving the lives of most of his patients with sepsis by using a treatment based on intravenous vitamin C.
Scientists from around the world have jumped on this claim to support or refute it because, if it worked, it would save millions of lives. The biggest study published to date now comes to a vexing conclusion. If you read the study summary, vitamin C didn't help the patients. But if you dig deep into the paper, you will find that the people who got the treatment were much more likely to survive.
Confusing? Absolutely. The rub comes from the way the study, published Tuesday in the American Medical Association journal, JAMA, was designed.
Sepsis is the body's overreaction to a severe infection, leading to multiple organ failure and, frequently, death. Though many people haven't heard of sepsis, the Centers for Disease Control and Prevention says it afflicts more than 1.7 million Americans a year and kills more than 250,000, making it one of the top 10 causes of death.
Over the years, scientists have stumbled across clues that vitamin C might be effective against sepsis. For one thing, people with sepsis tend to have surprisingly low levels of vitamin C in their blood. In 2014, Dr. Alpha "Berry" Fowler published a paper involving just 24 patients, hinting that vitamin C was a benefit. In particular, Fowler noted that a measure of organ failure, known as a mSOFA score, improved far more in the patients who had received vitamin C.
That study was far too small to draw any real conclusions. But Dr. Paul Marik at the Eastern Virginia Medical School in Norfolk read the study and was intrigued. When Marik encountered a sepsis patient who seemed on the verge of death, he decided to give her vitamin C. He added thiamine (a B vitamin) and a dose of steroids to her treatment. And, much to his surprise, the woman quickly recovered.
Marik then started using it regularly in his intensive care unit. He reported in 2017 that the mortality rate for sepsis in his ICU had plummeted after he switched to this treatment. That paper led some ICU doctors to start using the so-called Marik protocol, figuring that it was quite safe and possibly effective. In the meantime, researchers geared up to do careful studies to see whether the treatment really worked.
Meanwhile, Fowler, at Virginia Commonwealth University, pressed on with his own research. He rounded up a group of colleagues to further explore the role of vitamin C in one set of sepsis patients — those who became extremely ill after they developed an associated disease called acute respiratory distress syndrome.
Fowler's team studied 167 patients, half of whom received the vitamin C treatment and half of whom received a placebo. They chose as the mark of success the same endpoints he had used in his study of 24 patients.
The result: no difference in these scores at all. The bottom line was that vitamin C "did not significantly improve organ dysfunction scores or alter markers of inflammation and vascular injury," the study says.
Here's the twist. The researchers also looked at 46 other "secondary" endpoints. And one in particular stands out: 28 days following treatment, 25 of the 84 patients treated with vitamin C died, compared with 38 of the 83 patients given the placebo. That's a big difference, with the mortality rate dropping from 46% to 30%.
If death had been the primary endpoint of the study, this result would have been highly significant. The conclusion would strongly support the hypothesis that vitamin C is an effective treatment of sepsis.
But there's a catch. Since Fowler and his colleagues looked at 46 secondary endpoints, it's likely that something would randomly pop up as statistically significant. It's as though they had 46 bites at the apple to find something meaningful. To guard against that, statisticians have strict rules for analyzing multiple comparisons. And when those rules are followed, the death numbers in the JAMA paper don't pass that test.
What patients really care about, of course, is not their mSOFA scores and blood test results, but whether they live or die. Fowler tells NPR that he now rues his decision to select an endpoint that seemed more likely to show a benefit (as it did in his study of 24 patients). "So many sepsis studies have failed that used mortality as a primary endpoint," he says. He and his colleagues worried that if the result was not dramatic, "vitamin C probably would have fallen on its face."
Though he's now bound by the rules of experimental design to downplay the mortality results, he personally feels a sense of success. "We're all whooping and hollering because of what we found," he says.
There is less celebration in other quarters. "Despite this rigorous clinical trial, the question remains: does administration of vitamin C in patients with organ dysfunction confer benefit?" write Drs. Emily Brant and Derek Angus at the University of Pittsburgh in an editorial that is published alongside the paper. While the death results are "tantalizing," they note that none of the many clinical measurements normally associated with better outcomes can explain the higher survival rate in vitamin C patients.
"They have to explain how vitamin C would have an impact on mortality through a mechanism that did not affect the organ failure scores," Dr. Steve Goodman at Stanford University writes in an email to NPR. Goodman, who is an expert in experimental design, notes that the Food and Drug Administration has at least once approved a drug (a heart medication called carvedilol) based on improved mortality, even though the drug didn't appear to work based on its primary endpoint. But in his view the mortality result in this new study isn't so convincing.
Fowler has some ideas to explain the results — at least after the fact. One issue is that these patients were far sicker than the patients in his original study of 24 patients. Some were septic for four days before treatment began, unlike the other study in which patients often got the treatment promptly.
One requirement of this study was that the patients must have developed the serious complication of ARDS and had been put on mechanical ventilators.
The findings also don't have a direct bearing on the Marik protocol. That's not simply vitamin C (at a much lower dose than the Fowler study) but also includes corticosteroids and thiamine (vitamin B1). Marik has been treating patients at the first signs of sepsis, when treatment is more likely to be successful.
"One could argue this [new result] doesn't say very much about the Marik protocol," Brant tells NPR.
Marik agrees. He says the vitamin C and corticosteroids have a synergistic effect that is not evident in the new study. While Marik agrees that the study is not definitive, he says it importantly finds no side-effects and it does show a mortality benefit. "You can argue about all the statistical nuances, but that's what the study showed," he said.
Given how polarized physicians' opinions are about the value of vitamin C for sepsis, "I suspect when all is said and done this is not going to change a single person's point of view," says Dr. Craig Coopersmith, interim director of the Emory Critical Care Center at Emory University.
The ultimate lesson here is that a single study is rarely definitive. In this case, everyone agrees with the cliché that more research is needed. Fortunately, those studies are in the works. Two of the biggest studies looking directly at the Marik protocol are in their end stages and could provide answers in the coming months.
You can contact NPR Science Correspondent Richard Harris at email@example.com.