In order to demonstrate the relationship between performance characteristics of laboratory tests and clinical outcomes, diabetes seems to represent a paradigmatic disease: diagnosis, monitoring of therapeutic efficacy and prognosis are adequately achieved by means of laboratory testing. Starting from a simple molecule, glucose, used for the diagnosis of diabetes, continuing with creatinine, used for monitoring renal function in diabetic patients and concluding with cardiac troponins, a recognised gold standard for the diagnosis and risk stratification of cardiovascular diseases, several criticisms may be stressed considering the current methodological state-of-the art. Finally, an often overlooked aspect of performance, the analytical interferences, being responsible of unexpected results, that in turn depend from unknown or undisclosed factors will be discussed, concerning in particular, in our paper, the macroprolactin and the heterophilic antibodies aspects.