Moderator: ofonorow
ofonorow wrote:Thank you for this update/report. I understand medical inertia to some extent, but what I find hard to understand is why medicine doesn't immediately begin giving ICU patients these antioxidants? What is the downside? The upside seems to be the chance of preserving organs in half the patients who would otherwise experience organ failure. I can understand the hesitancy w/r to prescription drugs, which are usually highly toxic, but antioxidants are at the other end of the spectrum. If it was a member of your family, wouldn't you want them to be getting the antioxidants?
It's a fair question, but do we really know for certain that the combination of antioxidants is safe in all ICU patients? Based on what we know from healthy patients, it probably is - but it doesn't hurt to have a larger trial to corroborate both the efficacy and safety findings. Sometimes an intervention that appears perfectly safe in a few hundred patients ends up having some rare side effects in a much larger trial.
ofonorow wrote:It's a fair question, but do we really know for certain that the combination of antioxidants is safe in all ICU patients? Based on what we know from healthy patients, it probably is - but it doesn't hurt to have a larger trial to corroborate both the efficacy and safety findings. Sometimes an intervention that appears perfectly safe in a few hundred patients ends up having some rare side effects in a much larger trial.
This is always the excuse given, that "larger" trials are necessary, when in fact larger trials are less scientific, primarily because they are harder (impossible) to replicate by other objective scientists.
In this case, the conditions are controlled better than any human study we might invent. ICU patients diets (lives for that matter) are under 24 hour observation and tightly controlled. (In my case, when I was having abdominal surgery, my wife sneaked me vitamin C which I added to my water, and I subsequently became my surgeon's poster boy for fast recovery. I tried to tell him later, but he was uninterested, to say the least. It is unlikely for the same thing to happen in the ICU.)
A poorly designed, large trial may be less valid than a very well run small trial; nonetheless, larger trials that are well designed help to mitigate sampling error, effects due to chance, poor randomization, etc. They are also essential for finding small but clinically important differences that would not reach statistical significance in a smaller trial.
Large-scale double-blind mega-studies are the current fashion in medicine. The assumption is made that by increasing the number of subjects, the work somehow becomes more valid. Regrettably, these studies are borderline science, as a large scale study is difficult to repeat and replication is at the core of the scientific method. Of particular concern are those studies that are so expensive that only pharmaceutical companies are ever likely to perform them. The way studies are financed and the potential source of any experimental bias or error become more important with larger studies, as they are more difficult to replicate. Replication reduces the possibility of bias. A large-scale study giving positive results is less convincing than, say, three smaller, equally positive studies by independent researchers, in different institutions, separately funded and using dissimilar methods.
Sure, you might be able to control what an ICU patient receives in the form of food or treatment. But then you're assuming that everyone in the trial will receive the same treatments; that's not true, and if you look at the study in the OP, patients with a wide variety of conditions were recruited for the study. The two groups in the study are quite homogenous in terms of baseline characteristics that were recorded, but you always run the risk of unaccounted for differences for the treatment groups when the study is small. For instance, I do not see things like smoking status/history being accounted for in the baseline characteristics of the two groups - chances are that the incidence should be the same in both groups due to randomization, but that is not always the case. Such differences could potential confound the results.
ofonorow wrote:A poorly designed, large trial may be less valid than a very well run small trial; nonetheless, larger trials that are well designed help to mitigate sampling error, effects due to chance, poor randomization, etc. They are also essential for finding small but clinically important differences that would not reach statistical significance in a smaller trial.
This is the general argument used to justify larger trials, and in the case of prescription pharmaceuticals (i.e., highly toxic (toximolecular) substances) there may be some justification, but in general the argument is fallacious. At some value of N the statistical confidence reaches a high enough level that adding more subjects does not provide that much more useful information. But increasing the study size/cost does makes it harder to replicate and increases the chances for error and bias.
The following is, I believe, a profound quote on this very issue that I happen to agree with from Professor Hickey and Robert's book ASCORBATE: The Science of Vitamin C (lulu.com/ascorbate) After reviewing the basics of the scientific method, they write on page 34:Large-scale double-blind mega-studies are the current fashion in medicine. The assumption is made that by increasing the number of subjects, the work somehow becomes more valid. Regrettably, these studies are borderline science, as a large scale study is difficult to repeat and replication is at the core of the scientific method. Of particular concern are those studies that are so expensive that only pharmaceutical companies are ever likely to perform them. The way studies are financed and the potential source of any experimental bias or error become more important with larger studies, as they are more difficult to replicate. Replication reduces the possibility of bias. A large-scale study giving positive results is less convincing than, say, three smaller, equally positive studies by independent researchers, in different institutions, separately funded and using dissimilar methods.
Sure, you might be able to control what an ICU patient receives in the form of food or treatment. But then you're assuming that everyone in the trial will receive the same treatments; that's not true, and if you look at the study in the OP, patients with a wide variety of conditions were recruited for the study. The two groups in the study are quite homogenous in terms of baseline characteristics that were recorded, but you always run the risk of unaccounted for differences for the treatment groups when the study is small. For instance, I do not see things like smoking status/history being accounted for in the baseline characteristics of the two groups - chances are that the incidence should be the same in both groups due to randomization, but that is not always the case. Such differences could potential confound the results.
The logical extension of this argument is that you should only study male smokers, age 50 who weight 150 to 190 lbs, etc. However, then you only know what affects 50-year-old males of medium weight, which means you don't really know anything. We are all biologically unique, and of course people in intensive care will vary across the board in almost all factors. Yes, they may be getting different drugs, so I suppose that creating subgroups on the basis of medications might be useful, but the test is whether innocuous antioxidants added to the normal ICU regimen helped, or hurt patients. The answer in at least 2 studies was yes - vitamin C and E reduced the average length of stay and reduced organ failure by almost half.
You sound like you don't trust these researchers, that you fear they would have somehow put all the smokers in the group that experienced organ failure.
I know the feeling about not trusting researchers. When science is conducted in accordance with the scientific method - is replicated as Hickey/Roberts discuss - the requirement to trust researchers is greatly diminished.
JAMA STUDY: [color=#400040] JAMA Jan 23/30 2002 RANDOMIZED, DOUBLE-BLIND CONTROLLED TRIAL IN HUMANS FOUND STATISTICALLY SIGNIFICANT 60-SECOND TREADMILL EXERCISE IMPROVEMENTS IN 5000 MG VITAMIN C GROUPS [Chelation Therapy for Ischemic Heart Disease: A Randomized Controlled Trial, Knudtson, et. al. JAMA, Jan 23/30, 2002 - Vol 287, No 4. Pp 481-486]
Users browsing this forum: No registered users and 0 guests