Background and importance Osmolarity is one of the pharmaceutical controls carried out on the nominative parenteral nutritions (NPN) produced at the pharmacy for a given patient (magistral formula). According to a previous method validation, we use the Pereira Da Silva equation (PDS) when the theoretical osmolarity (TO) determined by this model is greater than 1000 mosmol/L and we use the manufacturers’ data (MD) equation when the TO according to the PDS equation is below 1000 mOsmol/L. This method is associated with an osmolarity nonconformity (NC) rate of 7.0%.

Aim and objectives To determine the best predictive model to calculate TO of the NPN to decrease osmolarity NC rate.

Material and methods A retrospective analysis of measured osmolarities was carried out from 1 June 2018 to 7 September 2021 to determine the osmolarity classes most affected by NC. The MO (measured osmolarity) was compliant if it is between –10% and +10% of the TO. Different models (modification of the limit value; choice of the two models within a range of osmolarities) were tested and the one with the lowest NC rate was chosen.

Results A total of 3795 MO were analysed. Table 1 shows the distribution of MO, their NC rate according to the current model and their NC rate according to PDS and MD equations.

By increasing the osmolarity limit value to 1150 mOsmol/L (instead of the current 1000 mOsmol/L), the NC rate decreases to 5.2%. By allowing both equations for a TO according to the PDS equation between [900;100]; [800–1200] or [700–1300], the NC rate decreases to 4.8%; 4% and 3.5%, respectively.

Conclusion and relevance The optimal theoretical model depends on the TO PDS. Allowing both PDS and MD equations, we reduced the NC rate. Therefore, the actual TO method should be revised in favour of the MD equation for TO under 700 mOsmol/L, both MD and PDS equation for TO between 700 and 1300 mOsmol/L and PDS equation over 1300 mOsmol/L.

Conflict of interest No conflict of interest