Sham versus Placebo: Implications for Interpreting Randomized Controlled Trials

[Epidemiologic Inquiry 2006, 1: 14]

Clinical investigators have often naively touted any trial to be the absolute truth, giving the randomized controlled design precedence over observational studies. However, in a recent issue of the BMJ, it was discovered that the type of control method used can significantly impact the interpretation of a randomized trial result.

By comparing a sham acupuncture procedure versus an inert placebo on physical function, the investigators discovered that the two "control" methods yielded different functional results and different rates of adverse effects. Obviously the psychosocial placebo effect is much different depending on the method of control. This has implications for interpreting other randomized trials which use different intervention for controls (such as: placebo, general advice, normal behavior, alternative diet, etc), which may elicit a different confluence of adaptive factors (also known as "canalization").

Thus, are randomized trials truly interpretable as the absolute gold standard? What other limitations are there? (to be continued).

Sham device v inert pill: randomised controlled trial of two placebo treatments.
Kaptchuk TJ, Stason WB, Davis RB, Legedza AR, Schnyer RN, Kerr CE, Stone DA, Nam BH, Kirsch I, Goldman RH.
BMJ. 2006 Feb 1; EPUB

Resolving Differences of Studies of Estrogen and Cardiovascular Disease

One of the greatest controversies in the world of observational research and randomized trials has been the discrepant findings between epidemiologic results that indicate oral estrogen is beneficial for lowering risk of coronary heart disease (CHD) and the Women's Health Initiative trial results that indicate no effect on risk of CHD.



However, upon closer inspection of the original results, there indeed may be a strong biologic explanation for the heterogeneity not due to residual confounding. This month in the Journal of Women's Health, the original estrogen investigators Grodstein et al. examined the estrogen-CHD association in a reanalysis of the Nurses' Health Study and reviewed the contrasting results between NHS and the WHI.



Notably, they first highlighted that women in the WHI were significantly older (majority a full decade since menopause) than women in the NHS at initation of estrogen therapy (who were mostly perimenopausal and <5 year since menopause at baseline). Previous experimental evidence from primates indicated that oral estrogen has strongly divergent effects of being protective against CVD risk factors among younger primates, while not protectives in older primates.



More importantly, the original NHS study found that the benefits of estrogen therapy diminished over time with longer duration of use (longer time since menopause). Indeed, the latest reanalysis confirm that among a subset of nurses of comparable age and years since menopause to WHI women, results were consistent in which there was no effect of estrogen on CHD risk; meanwhile among younger nurses of relatively few years since menopause, NHS women using oral estrogen did show decreased risk of CHD.



To top it off, Grodstein et al.'s reanalysis of the WHI age-stratified results indeed also confirms such heterogeneity in which younger WHI women age 50-59 showed protective benefits of estrogen RR=0.56, while moderately older women age 60-69 and older women age 70-79 showed no benefit of estrogen therapy on CHD risk, RR=0.92 and RR=1.04 respectively (meta-regression p trend=0.036).



In summary, multiple lines of evidence from primate experimental results, observational studies, and randomized trials all support the above effect modification of estrogen's effects on CHD risk. Thus, perhaps the worlds of observational epidemiology and randomized trials unite afterall.



Hormone Therapy and Coronary Heart Disease: The Role of Time since Menopause and Age at Hormone Initiation.
Grodstein F, Manson JE, Stampfer MJ.
J Womens Health (Larchmt). 2006 January/February;15(1):35-44.

[The Editors have no personal or professional conflict of interest in the above editorial]