Point-of-care testing (POCT) for blood glucose is widely used in different clinical settings and, thanks to its rapid turnaround time (TAT), it is commonly recommended in patients with critical conditions calling for urgent monitoring. However data in the literature on its reliability and safety are scarce. The aim of the present study was therefore to investigate the type, frequency and clinical risks associated with POCT using portable glucose meters in the hospital setting by evaluating the post-analytical phase when the data are manually transcribed in patients' record. We checked all registrations performed by nurses of 2 selected wards of the hospital during a period of 30 consecutive days, identifying 1966 blood glucose determinations. We found that data obtained were often not reported in patients' files (12.1% of results are missed), the time of blood sampling was recorded in imprecise manner in 7.2% of cases and the glucose concentration was incorrectly reported in 3.2% of results. Although in the post-analytical phase the frequency of incomplete or incorrect data was high, no adverse events were found to be related to this type of error.