All this has angered the doctor who pioneered the aggressive therapy, Emanuel P. Rivers, vice chairman of emergency medicine at Henry Ford Hospital in Detroit. The controversy has taken on an ugly tone rare in medicine. Dr. Rivers has called his critics in Pittsburgh "the Pittsburgh pirates." In an email to a Toronto doctor, Dr. Rivers said some see criticism of his therapy as a "thinly veiled academic lynching."
We’ve already seen an attempted academic lynching of the sepsis guidelines here, which I debunked here. But Dr. Rivers, I believe, has more reason to be angry at the Wall Street Journal than the academics. The Journal article is full of distortions and confusing statements, and those readers whose attention span doesn’t extend beyond the sound bite may rush to a faulty conclusion about EGDT.
The article claims, but in a very roundabout and confusing way, that Rivers didn’t use an intention to treat analysis. And, all too typical of the popular press, primary sources are not given. The major point of confusion concerns the 25 patients who were excluded from the study. One statement quotes an uncited interview with Dr. Rivers:
Dr. Rivers, in an April interview describing how he proceeded, said "all 288 are randomized" -- that is, placed into either the standard-therapy group or the other group -- "and then some are excluded because they're not considered severe septic shock."
But once patients are randomly placed in a group, they must be included in a final analysis, according to a fundamental principle of medical research. Twenty-five were not, the medical-journal article makes clear.
That would imply that an intention to treat analysis was not used. This informal, uncited interview (which I was unable to find despite serious Googling) cannot be regarded as a reliable source to determine the methods of the Rivers study.
A quote from Rivers later in the article seems to contradict the interview (italics mine).
Dr. Rivers said that "randomization was valid," that "all patients enrolled in the [early goal-directed therapy] study were appropriately selected, randomized and examined statistically," and that "there are no scientific integrity issues in the study."
Well, when in doubt, go to the methods section of the original paper (italics mine):
We evaluated 288 patients; 8.7 percent were excluded or did not consent to participate. The 263 patients enrolled were randomly assigned to undergo either standard therapy or early goal-directed therapy; 236 patients completed the initial six-hour study period. All 263 were included in the intention-to-treat analyses.
Rivers really put one over on the NEJM editors if the WSJ allegation is true. I hope he responds to the WSJ (not that the quality of his research needs explaining).
There are more distortions. From the second paragraph:
Adding to the concerns: That hospital held patents on a medical device critical to the therapy.
We’re not told what the device was, but it must be in reference to a central line and/or an O2 sat monitor attached to it. The use of a central line is nothing new. It’s been common place in the management of critically ill patients for years. And, unless I misunderstand blood gas technology the use of a continuous monitor to measure SVC saturation is not an essential part of EGDT. You can simply draw a sample from the line and run it through a blood gas machine.
The next paragraph implies that EGDT is expensive---
The new therapy typically costs about $1,100 more per patient, by the estimate of Derek Angus, chief of critical care at the University of Pittsburgh. The U.S. has about 750,000 cases of sepsis each year.
---but ignores evidence that EGDT actually saves money.
The article’s description of the protocol is simplistic:
Sepsis kills in part because tissues don't get enough oxygen from the blood. Dr. Rivers's idea: Continuously monitor the level of oxygen in the blood and take steps to raise it if it gets too low.
For monitoring, he used a catheter, inserted in the jugular vein and slid to near the heart. If it showed that blood oxygen was too low, he elevated it with transfusions and often added a drug to make the heart beat harder.
Not quite. The rationale for the Rivers study was to asses whether the achievement of multiple hemodynamic goals makes a difference when done early. Those questions were driven by disappointing results of sepsis treatment reported in the 1990’s: that “any old time” goal directed therapy doesn’t work.
Arguably the most important function of the central line is the initial measurement of CVP to guide fluid resuscitation. The oxygen measurement, SVC saturation, is not a measurement of oxygen delivery as much as an indicator of tissue extraction.
Lead investigators in the NIH funded trial were interviewed in the accompanying video. Aside from what I believe to be a distortion about how widespread adoption of EGDT would impact the blood supply the video was helpful in explaining the NIH study. The three arms will be 1) standard (non protocol driven) care, 2) the full Rivers protocol and 3) the Rivers protocol minus the central line. I think this will be useful because it will help evaluate individual components of EGDT. It adds to the Rivers study, which looked at the whole package. From the NIH press release about the new study:
To implement this approach across the entire country is a daunting task. As a first step, the consortium aims to find out whether the protocol will have similar success at multiple hospitals across the nation.
Following a year-long period of establishing and training sepsis teams, the consortium plans to treat early severe sepsis in nearly 2,000 patients, enrolling patients over a two- to three-year period. It will randomly assign patients to Rivers’ protocol or to the existing “usual care” approach, which does not include such aggressive resuscitation. It will then follow all the patients for a full year after their recovery to detect any long-term differences in health and mortality rates.
Therein lies an unintended consequence. It’ll be years before we know the results. Meanwhile, aided and abetted by the press, clinical inertia will drag on. Hospitals administrators, already obsessed with perfunctory “core measures”, will have yet another excuse (“let’s just wait for the results of the NIH trial”) to delay institutional commitment.
That, of course, violates fundamental tenets of evidence based medicine: science is tentative; go with the best evidence you have at the time. So how good is the evidence for EGDT in 2008? As far as I can tell, the Rivers study is sound. I’ll need better sources than the WSJ to convince me otherwise. But if that’s not enough, his conclusions have been validated independently in other studies (here, here and here).