More Mini-Trials And Fewer Mega-Trials?

Just as dinosaurs ruled the earth for millions of years, for more than 30 years now cardiology has been ruled by mega-trials. Over the years the cardiology landscape would shake and tremble as ever more gigantic mega-trials emerged to  dominate the field.

Just as the giant dinosaurs paid no attention to the seemingly inconsequential little furry mammals running around beneath their feet, cardiology has tended to ignore or downplay smaller trials— what I will here call mini-trials.

Consider the reflexive dismissal or diminution of the ORBITA trial. Critics, many of them proponents and leaders of mega-trials, immediately opined that ORBITA need not be taken seriously, that it could not possibly raise serious questions about current practices and beliefs, because it “only” enrolled 200 patients and “only” followed them for 6 weeks. But these critics never explained why a trial designed to study whether patients with stable ischemic heart disease experience symptomatic relief from PCI would actually need more patients or longer followup. For most critics there was just the blind assumption that bigger is not only better but necessary. In fact, in this case bigger could well have been worse, since it could well have produced a statistically significant difference that in clinical terms would have been meaningless. (This should not be a new concept to anyone who has been paying attention.)

It’s not yet widely known that the same group, led by principal investigator Darrel Francis (Imperial College London), is now performing an even smaller trial that could have an equal or even greater impact than ORBITA.

Although it will only enroll 50 patients, an astonishingly small number in a universe dominated by mega-trials, SAMSON (Self-Assessment Method for Statin Side-effects Or Nocebo) is designed to help solve one of the most pressing, recalcitrant, and annoying problems in all of cardiovascular and preventive medicine: patient resistance to statins.

Statins of course have their critics but most knowledgeable observers have been persuaded by the clinical trial evidence demonstrating their benefits, at least in a very high risk population of people. The problem is that a large and growing number of people, and even many healthcare professionals, including some cardiologists, believe that statins cause all sorts of side effects. Muscle weakness is the most commonly cited side effect but there is a litany of others, such as cognitive impairment, that can be terrifying to many potential statin takers.

A few years ago a group of leading mega-trialists published an enormous meta analysis that included data from 175,000 randomized patients. They concluded that the evidence supporting both the benefits and safety of statins was overwhelming. The authors were dismayed when their shock and awe strategy did not provoke an immediate retreat or surrender by the anti-statinista forces.

A major limitation of the paper is that the trials included in the meta-analysis did not assess side effects in a uniform or rigorous fashion. Another area of concern is that the patient-level data from the trials has not been made available for outside scrutiny.

But even if we accept the conclusions of the meta-analysis authors and we believe that the true incidence of statin side effects is quite small, we are left with a remaining and quite difficult problem. How does an individual person know whether or not he or she is one of those very few people who have real side effects? This is an inevitable limitation of mega-trials, which study large populations. Patients, however, want to know what it means for them.

SAMSON does something that mega-trials are unable to do. It delivers– even before the trial is completely finished!— specific, valuable and actionable information to the trial participants.

The trial design is remarkably simple and clever. The study population includes adults who have stopped taking statins because of side effects. At the start of the trial participants receive 12 bottles, one for each month, which they take in a a random predefined order. Four bottles contain a daily statin (atorvastatin), 4 contain a daily placebo, and 4 are empty. Participants record their symptoms every day in a smartphone app. In addition to the aggregate data from all the trial participants, individual participants will serve as both their own blinded placebo control and unblinded (because 4 of the bottles are empty) no treatment control.

Although the trial will only enroll 50 patients it should be powerful enough to help answer both the question of the true rate of statin side effects and, even more importantly, at least for the individual participants, whether their own symptoms are caused by statins. Participants will receive their own results at the end of the 12 months— even before the remaining patients have completed the trial. Thus they should have conclusive information about their own individual side effects. (Of course we have no way of knowing at this point how that knowledge will be utilized, but at least there will be no question of how the data applies to the individual patient.)

I am not going to argue that mega-trials have no value. Of course they do. But it is should also be clear to all that mega-trials have become increasingly difficult and expensive to perform, and that they are often no longer as impactful or interesting as they once were. Too often enormous resources have been devoted to answering questions that industry wants answered instead of being used to answer questions of fundamental importance.

By contrast, small studies like ORBITA and SAMSON demonstrate that by thinking outside the box it is still possible to perform original and important research without huge resources. Unless we want the cardiology field to serve as the research arm for industry or Wall Street we need to encourage and welcome original mini-trials trials like ORBITA and SAMSON that challenge the dinosaurs of our time.






  1. I wonder how a trial – large or small – is going to sort out the effect of interactions between the statin and the other medicines the patients might be taking. I suppose that at least the designers of the trial urged the patents not to change their medications during the trial. Or would that be impractical and unethical? Perhaps the most that can be hoped for is that the patients record all changes, down to each dose of aspirin or paracetamol, of laxative or diahhorea-blocker.

    Anyway, it sounds to be a wonderfully cheap way to have at least a preliminary inspection of the issue. I say “preliminary” since I can’t take seriously a study whose perpetrators decline to release their data for other workers to analyse.

  2. In case I’ve been so rude as not to say so before, Happy New Year, Mr Husten.

    I thought this might interest you.

  3. The quality of any meta analysis cannot be any greater than that of the constituent studies.

    In looking at statin studies on record, factors to consider:
    • Incompleteness or obfuscation of data
    • Obsolescence of research technique

    Any sort of fresh look would be a good idea, presuming state-of-the-art research standards.

    • Larry Husten says

      The statin studies were not bad studies. But they were not designed for a rigorous assessment of side effects. The availability of patient level data would be helpful.

  4. A month’s sunny vacation?
    I check (almost) daily.

Speak Your Mind