Study finds Echinacea useless for colds

RobertR · 07-28-2005, 01:33 PM

http://www.latimes.com/news/printedi...ck=1&cset=true

I have no doubt this won’t dissuade “true believers”, but as the article says, good science speaks for itself and enough is enough.

D. Scott MacDonald · 07-28-2005, 02:56 PM

I thought that a study several years ago already proved this. On the other hand, another study found that zinc provided a small benefit, but only if taken consistently every four hours or so from the time that the cold first starts.

Mort Corey · 07-28-2005, 03:22 PM

Not registered with the LA Times so couldn't read the article, but that stuff sure does dry up my sinuses (mouth and everything else) for some reason when I've taken it in the past. Haven't had a cold in the last four or so years so I haven't had the occasion to take it in a while.

Mort

Chris Lockwood · 07-28-2005, 03:30 PM

You can "prove" anything with a study. Aren't there studies showing cigarettes are good for you?

JeremyErwin · 07-28-2005, 03:37 PM

Here's the abstract from The New England Journal of Medicine (full text costs $10)

Quote:

An Evaluation of Echinacea angustifolia in Experimental Rhinovirus Infections
Ronald B. Turner, M.D., Rudolf Bauer, Ph.D., Karin Woelkart, Thomas C. Hulsey, D.Sc., and J. David Gangemi, Ph.D.

Background Echinacea has been widely used as an herbal remedy for the common cold, but efficacy studies have produced conflicting results, and there are a variety of echinacea products on the market with different phytochemical compositions. We evaluated the effect of chemically defined extracts from Echinacea angustifolia roots on rhinovirus infection.

Methods Three preparations of echinacea, with distinct phytochemical profiles, were produced by extraction from E. angustifolia roots with supercritical carbon dioxide, 60 percent ethanol, or 20 percent ethanol. A total of 437 volunteers were randomly assigned to receive either prophylaxis (beginning seven days before the virus challenge) or treatment (beginning at the time of the challenge) either with one of these preparations or with placebo. The results for 399 volunteers who were challenged with rhinovirus type 39 and observed in a sequestered setting for five days were included in the data analysis.

Results There were no statistically significant effects of the three echinacea extracts on rates of infection or severity of symptoms. Similarly, there were no significant effects of treatment on the volume of nasal secretions, on polymorphonuclear leukocyte or interleukin-8 concentrations in nasal-lavage specimens, or on quantitative-virus titer.

Conclusions The results of this study indicate that extracts of E. angustifolia root, either alone or in combination, do not have clinically significant effects on infection with a rhinovirus or on the clinical illness that results from it.

"Challenged with rhinovirus type 39..." That sounds fun!

Here's a link to the new york times (no registration required).

Quote:

But in an editorial accompanying Dr. Turner's paper, Dr. Wallace Sampson, editor of The Scientific Review of Alternative Medicine, which analyzes alternative-medicine claims, said there was no reason to have believed that echinacea was effective against colds in the first place. In the early 20th century, echinacea "somehow became popular for the treatment of respiratory illness in Germany," Dr. Sampson wrote, while in the United States it was used for wound healing and other purposes for which today people would take an antibiotic. Its use faded away when real antibiotics were discovered, but it re-emerged in the 1960's as a cold remedy, with no particular reason to think it would work.

**HTF Ads**

RobertR · 07-28-2005, 03:42 PM

Quote:

You can "prove" anything with a study. Aren't there studies showing cigarettes are good for you?

Chris,

Saying some studies are flawed is far from a sufficient critique of this one. What specific criticisms do you have of their methodology? It looks sound to me.

andrew markworthy · 07-28-2005, 04:02 PM

Quote:

You can "prove" anything with a study. Aren't there studies showing cigarettes are good for you?

The short answers - no you can't and yes there are, but I still wouldn't smoke.

The long answer:

You can only 'prove' anything with a study if the study is badly constructed or downright fraudulent. If a study is properly controlled and the sampling is correct, then it should be accurate. By 'properly controlled' we mean that in addition to the experimental question being asked, other possibile explanations have been addressed. The classic case is studying effects of drug treatments. Suppose you have drug X that is supposed to cure disease Y. The most obvious experiment is to give a bunch of people drug X and see if they get better. Obvious, but wrong. Even if the participants all get better, you haven't proved that X cured Y because the participants could have got better even if they hadn't been treated. So then you run a study where you look at people with disease Y, and only treat half of them with X. Then you see who gets better. If you find that the untreated group don't get better but the group receiving X do, you still haven't proved that X cures Y, because there could be a placebo effect. Therefore, you redo the experiment, this time with a no treatment, a placebo treatment and drug X treatment. Suppose you find that only the drug X group get better. You still haven't proved that X cures Y because perhaps the people observing the patients are biased and so classify the participants they know are getting X as better, even though objectively there is no difference. Therefore, you have to redo the experiment yet again, this time running a double blind procedure (where neither the participants nor the observers know which group anyone belongs to).
You will appreciate that I am describing a simple experiment - the more complex ones are of course a lot more convoluted. In many instances you not only have to deal with the experimental variables in front of you but also have to allow for confounding variables. A simple example - suppose that you find that obesity in linked to a disease. Can you say that being overweight is the prime cause of the disease? No - because obesity is associated with many other factors, such as lack of exercise, social class (the lower the socio-economic group, the higher the proportion of overweight people), etc. So obesity may only be associated with the disease, not cause it.
And then you have the issue of sampling. You can't usually test the entire population of people with a particular medical condition, so you therefore take a sample and then conduct statistical analyses to see if what you find in the sample is likely to be true of the population from which it's drawn. However, unless your sample is accurate to begin with and also you are good at choosing the right statistical techniques (and in my experience medics aren't terribly good at this) then the results are likely to be faulty.
Which all leads to the conclusion that if everything is done properly, then the results should be accurate, but there are potentialy pitfalls. However, that is not the same as saying that you can prove anything with a study.

There is also the argument that some people seem to buck the trend. Although a study says one thing for the general population, they are the reverse. This in no way disproves the validity of a study. Statistics usually only tells us what is true for a group. Exceptions to most things can be expected. For example, I have a naturally 'high' temperature - that doesn't disprove statements about what is a normal temperature range.

Chu Gai · 07-28-2005, 06:20 PM

How silly. Everyone knows you starve a cold and feed a fever!

07-28-2005, 07:30 PM

Quote:

Saying some studies are flawed is far from a sufficient critique of this one.

Robert, what are you talking about? Everyone knows that merely referencing a bad example in a set, even without specificity, obviates the supposed virtues of all other members of that set. It is a debating technique that -- like ad hominem attacks, straw man arguments, and mooning -- cannot be countered.

Example:

Proposed: "My wife is very smart."

Counter Argument 1: "Oh, yeah? My sister is stupid, so all girls are stupid, so your wife is stupid."

See? You can't argue with logic!

----------------------------------------

Quote:

Everyone knows you starve a cold and feed a fever!

Actually, my understanding is that this is not a recipe for treatment, but an admonishment: If you starve a cold, it will get worse, and you'll end up feeding a fever -- much like, "A minute on the lips, a lifetime on the hips." In other words, it's not what you should do, it's what you should avoid.

In any case, I doubt whether this admonishment is worth heeding. Everyone knows that all such sayings are just made up stuff that never works.