CREDIBILITY OF CLINICAL TRIALS AND PRACTICE GUIDELINES

September 2009 is a very upsetting month for clinical research and practice guidelines. This is because in September, two important clinical papers were publish, further casting doubts to the value of clinical trials in guiding clinical practice and care of patients.

We have stated before in previous blogs, the issue of " ghost writers " . We have also highlighted the severe biaseness in the FDA for approval of some drugs, applying clinical trials in a rather bias manner. Remember the " prasugrel " issue with Dr Sanjay?.

In the month of September 9th, Dr Ross and colleagues published a paper entitled " Trial publication after registration in ClinicalTrials.Gov: A cross-sectional analysis. PLoS Med 2009; Similarly, Dr Sylvain Mathieu and colleagues across the big pond published "Comparison of registered and published primary outcomes in randomized controlled trials". JAMA 2009; 302: 977-984.

Both these papers dealt with the issue of clinical trials publication after registration. In 2005, the US government mandated that all clinical trials, must be registered with the "International Committee of Medical Journal Editors (ICMJE)", before they can be given permission to begin enrolment. This was a result of problems with the failure to complete clinical trials, especially those sponsored by pharmas, when the trial results seem unfavourable. The US government then decided that that will give the readers a very bias opinion about the drug as only positive views are reported. Sometimes, negative trials also teach us alot.

Dr Ross and colleagues found that from Dec 1999 - June 2007, 7,515 clinical trials were registered. 53% of them failed to report the closing date of the trials when it was ? finished, 66% failed to report the Primary Outcome of the trial, and 87% failed to report when the trials were commenced. Then they took a 10% samples of the 7515 trials, which were completed. Of these 10% completed trials, less then half, only 46% were published, supposedly, the results were favourable. When he did a breakdown, only 40% of trials sponsored by pharmas, had their results finally published, 47% of trials sponsored by the government, were finally published and 56% of trials sponsored by academic institution were finally published. One obviously wish to know, what happened to those trials that were never published. Is it because the authors had inadequate funding, were the authors in some way physically could not complete, or worse of all, is it because the trial results were negative and they wanted to hide that fact.

Dr S. Mathieu's paper had to do with the fact that, not only are trials selectively published, very often, their reported, registered outcomes, are not the same ( significantly different ) from their published outcomes. Somehow, between reporting to the authorities and sending it in for publications, other facts seemed to have come up and be included. This again does not give confidence to the system of scientific research.

Having seen all these facts, should we still trust clinical trials and by extension, clinical guidelines. I suppose, to some degree, many still do because they have nothing better. I am much more sceptical. Those pharmas who come to promote their products to me using clinical trial results, will be told that the trial results are interesting, but should not be taken at face value. The clinical trial results must be compartible with accepted pathophysiological mechanisms. In other words, it must make medical sense. A clinical trial results that is too good to be true, like the old cliche goes " is probably, too good to be true ". A case in point is, the association of COX2 inhibitors and coronary thrombosis. Many studies have shown that COX2 and COX1 inhibitors bothe have the same problem of coronary thrombosis. A study of the patho-physiological mechanism will show that COX2 agents are more prone to coronary thrombosis then the COX1 agents, because COX2 agents cause more " endothelial cell-platelet aggregation" imbalance then COX1. In other words COX1 is safer, from the coronary thrombosis stand point.

Anyway, I must not digress too much. What I was trying to highlight is that clinical trials are not as truthful as they are made out to be. Negative trials are often not reported and may be just as important for clinical practice guidelines than positive trials.

Can we then trust clinical trial results, and by extension, can we then trust clinical practice guidelines that are form on the basis of the clinical trial results?

Like everything else human, the system is not perfect. A sound scientific mind and good medical sense is still very important for patient care.