Georg Ivanovas From Autism to Humanism - systems theory in medicine
2.5 The limits of evidence based medicine
Side effects, even fatal, are quite common and highly underestimated. Estimates judge that in Sweden fatal side effects even sum up to 3% of all deaths in the general population (Wester et al. 2008). The situation might be worse in countries like Greece or France where drug consumption is much higher (Mölstad et al 2002). Only about 6 % of the adverse effects are related to the medical treatment (Ebbesen et al 2001, Schnurrer 2003). Mostly they are not brought into connection with a certain therapy (Cuervo/Clarke 2003).
Often severe side effects are not detected by RCTs but by observation, as rhabdomyolysis through Baycol/Lipobay (Vandenbroucke 2004a). Trails to prove or disprove side effects are normally initiated only if there is a certain suspicion as with acute myocardial infarction through COX (cyclooxygenase-2) inhibitors (Solomon et al 2004). But in this case it happened after more than 80 million patients had taken this drug (Topol 2004) and a lot of RCTs proving the effectiveness of COX inhibitors had been published.
As rare side effects, according to the non-trivial status of man are often not predictable, RCTs are not helpful (Vandenbroucke 2004b). In contrary, rare side effects slip through the meshes of statistical evaluation. EBM is a tool to dilute the appearance of rare side effects and make them seem as accidental. It took about 100 years to find a correlation between aspirin and pancreatic cancer in women (Schernhammer et al 2004) as this cancer is quite rare and no model had been developed to correlate the two. In contrary, findings in hamsters and in vitro suggested that aspirin has an antineoplastic effect on pancreas (Baron 2004). Even the prevalence of Aspirin induced asthma is much higher than expected before (Jenkins et al 2004). As this concerns one of the most used drugs in medical history, one could have expected that its side effects are well known. But this is not the case.
Another reason for the inability of RCTs to discover side effects might be that most studies exclude persons disposed to suffer side effects (e.g., old persons) which in general practice are treated with the drug (Dieppe et al 2004).
Moreover, the effect of the combination of drugs is rarely investigated, although they are common use. In a study with more than 7000 cardiovascular patients taking aspirin, the additional consumption of ibuprofen (common in practice) correlated with an increase of 50% of deaths in eight years (MacDonald/Wei 2003).