Biomarker-guided clinical decisions: for patients, health economists or neither?

Abstract

In this issue of the European Respiratory Journal , Albrich et al. [1] report on a randomised trial. Their hypothesis is that triage decisions in patients with respiratory infections, guided by adding measurements of the calcitonin-superfamily gene derived peptide, mid-regional proadrenomedullin, are more effective than if guided according to clinical assessment alone. Length of hospital stay is, amongst other end-points, compared between the groups. The authors should be commended for their effort in conducting a randomised trial with a biomarker-strategy; this implies a much increased workload when compared to epidemiological studies. Adrenomedullin is a 52 amino acid polypeptide. It is formed by amino acids 95–146 of “pre-proadrenomedullin” and is involved in multiple effects, which include vasodilation [2, 3], osteoblast activity [4], renoprotection [5] and angiogenesis [6]. Mid-regional proadrenomedullin (amino acids 45–92 of “pre-proadrenomedullin”) is produced in 1:1 ratio with adrenomedullin and is not a pro-hormone of adrenomedullin. The physiological role, if any, of mid-regional proadrenomedullin is largely unknown. Mid-regional proadrenomedullin levels in blood have been shown to be related to the prognosis of the patient. In a study of 228 patients with community acquired pneumonia, a high ability to predict a Pneumonia Severity Index score of 4–5 ( versus 1–3), was reported (area under the curve of the receiver operating characteristic curve: mid-regional proadrenomedullin 0.81, procalcitonin 0.62, C-reactive protein 0.59) [7]. However, knowledge of prognosis indicated by a biomarker is not always translated into …