Meta-Analysis in the Mirror of Its Quotations: Science, Scepticism, Scorn, and Sarcasm

Franz H. Messerli MD; Sripal Bangalore; Adrian W. Messerli


Eur Heart J. 2019;40(40):3290-3291. 

In This Article


Given these many pitfalls, physicians, and authors of guidelines should not be blamed for their reluctance to change long-standing practices on the basis of a single meta-analysis. In this context, one may recall a deceptive and defective meta-analysis in 1995 purporting to document that calcium channel blockers directly increase mortality in a dose-dependent way.[7] Ever since the Syst-Eur study, we know that calcium channel blockers are one of the best-documented drug classes to reduce morbidity and mortality in hypertensive cardiovascular disease. The above misleading and flawed meta-analysis actually provoked the dictum in the Lancet[6] 'A meta-analysis is like a bouillabaisse—no matter how much fresh seafood will be added, one rotten fish will make it stink!'.

Mutatis mutandis, The Lancet's Stuart Spencer countered the bouillabaisse principle with 'A spoonful of port may make a poor wine drinkable'.[9] Drinkable it may be for those desperate enough to do so, but this does not mean palatable, not to speak of enjoyable for those of us less audacious. A fine glass of Cabernet Sauvignon, i.e. a simple prospective randomized trial will often taste better than a poorly executed Meritage.

And on the subject of drinking, the dictum of the eminent British physicist and Nobel Laureate Ernest Rutherford comes to mind: 'An alleged scientific discovery has no merit unless it can be explained to a barmaid'.[10] Little question that a barmaid would easily grasp the bouillabaisse and the port principles although this may not help her much to comprehend and evaluate the quality of meta-analyses in general.