PSP and NT-proCNP assessed for predisposition to infection

Gouel-Chéron and colleagues have recently published a pilot study [1] on which we would like to offer the following comments.

As the paper currently appears, it is unclear whether in fact the study aimed atevaluating the predictive value of pancreatic stone protein/regenerating protein(PSP) and amino-terminal pro-C-type natriuretic peptide (NT-proCNP) for apredisposition to infection in non-infected systemic inflammatory response syndrome(SIRS) trauma patients under mechanical ventilation, or whether it rather aimed atevaluating the predictive value (that is, prognosis) of the biomarkers for sepsisdevelopment. Indeed, the authors omitted to mention that the assessment was carriedout in patients exclusively before the occurrence of a hospital-acquired infection.This certainly explains why there are only 16 patients plotted in Figure 1 and not21 (36% of 61 patients) as one would have otherwise expected. A non-equivocal use of'prediction' is therefore essential.

Providing these indications would have allowed to put the results into perspectivewith recent studies published in Critical Care on PSP ability todifferentiate patients with sepsis from those with a non-infective SIRS [2] or on NT-proCNP ability to confirm sepsis and correlate with markers ofbacterial infection [3].

Instead of analyzing the role of PSP and NT-proCNP in the prediction of sepsis, theauthors seem to have actually predicted future hospital-acquired infection in theintensive care unit. Stating that the study assesses the prediction of sepsis istherefore clearly a misleading shortcut, since in fact the occurrence of infection,sometime later than the time-point of analysis of the biomarkers, will classifythese SIRS patients as septic by definition.