Search

Biomarkers are becoming increasingly important in the process of development of new drugs. However, as they are often not able to reflect the actual clinical value and safety of a drug, their role in the framework of the licensing of drugs is limited. For the health authorities, proof of a positive benefit/risk ratio based on hard clinical end points is generally the decisive factor. Only in very few indications do validated surrogate end points exist that show a high predictive value for a relevant clinical end point – the granting of marketing authorization on the basis of a surrogate end point alone is therefore the exception. In principle, this is also true for drugs for the treatment of serious, life-threatening diseases, as the present accelerated approval procedures for the granting of a license on the basis of provisional data and without direct proof of a clinical benefit are used by the authorities only in isolated cases. Only if new biomarkers can be validated as surrogate end points for hard clinical end points can their value for the licensing of new drugs increase in the future.

Within the framework of evidence-based medicine, meta-analyses represent a methodological tool to consolidate the state of knowledge on particular medical questions. However, meta-analyses per se already raise some fundamental problems. If quality criteria are not taken into account in the setting up of meta-analyses, this can result in methodological bias, so that in the final analysis, the intervention efficacy can be clearly underestimated or overestimated. Unfortunately, despite the existence of guidelines for their preparation, many published meta-analyses show considerable shortcomings as far as the methodology is concerned, and their value for the drawing up of diagnostic or therapeutic guidelines therefore has to be considered as questionable. Only a limited number of collaborative groups, for example the Cochrane Collaboration, set up meta-analyses with a high level of methodology. While their systematic reviews are to be recommended, the large number of published meta-analyses have to be viewed with scepticism, since often their results and conclusions cannot be comprehended objectively. In those areas of medicine where data and evidence are scanty, even meta-analyses cannot work wonders, because the value of a meta-analysis depends entirely on the value of the individual studies. Therefore, randomised, controlled studies will in future continue to play the decisive role in finding answers to the existing medical questions – and meta-analyses can help to formulate the right questions.