PT - JOURNAL ARTICLE AU - Pelluz, C Sangrador AU - Pastor, J Pardo AU - Lopez, P March AU - Borge, S Gallardo AU - Capafons, S Redondo AU - Gordon, A Manzaneque AU - Biosca, R Garriga AU - Picó, J Nicolas TI - 4CPS-045 Predictive factors of hyperglycaemia in patients with parenteral nutrition AID - 10.1136/ejhpharm-2018-eahpconf.136 DP - 2018 Mar 01 TA - European Journal of Hospital Pharmacy PG - A61--A62 VI - 25 IP - Suppl 1 4099 - http://ejhp.bmj.com/content/25/Suppl_1/A61.2.short 4100 - http://ejhp.bmj.com/content/25/Suppl_1/A61.2.full SO - Eur J Hosp Pharm2018 Mar 01; 25 AB - Background Hyperglycaemia is the most frequent complication in patients with parenteral nutrition (PN). Numerous factors may favour its appearance.Purpose Identify the predictive factors of hyperglycaemia in patients with PN in order to guide the design of a starting PN.Material and methods Retrospective observational study (January to December 2016) performed in a 450-bed university hospital.All adult patients with central venous PN were included.Recorded variables: sex, age, body mass index (BMI), patient classification (surgical, critical and medical), diagnosis, comorbidities, duration and carbohydrates/kg (HC/kg) provided in PN, glycaemia prior to PN initiation, renal clearance (Clr), presence of sepsis and treatment with potentially hyperglycaemic drugs. Hyperglycaemia was defined as three consecutive blood glucose levels>150 mg/dL or two>180 mg/dL.Descriptive, bivariate and multivariate statistical analysis (binary logistic regression) was performed using the SPSS. v. 24 program.Results Two hundred and thirty-four patients were included, 65.8% male, mean age 65.3±14.2 years and BMI of 26.1±6.0 kg/m2. 66.7% were surgical, 16.2% critical and 17.1% medical patients. 20.1% had diabetes mellitus (DM), 19.2% dyslipidaemia, 10.3% hypertension, 7.3% kidney failure (KF), 4.3% heart failure (HF) and 1.3% hepatic failure. The mean duration of PN was 9.3±7.5 days, with a mean of 3.0±0.7 g HC/kg. Mean pre-glycaemia was 135.7±47.3 mg/dL. 16.7% received corticosteroids, 7.3% octreotide and none immunosuppressants. The prevalence of hyperglycaemia was 44% and 11.1% of sepsis.The predictive factors identified after the multivariate analysis were DM (p<0.001 95% CI: 3.028 to 31.697 OR: 11), previous glycaemia (p<0.001, 95% CI: 1.026 to 1.051 OR: 1), corticosteroid treatment (p=0.023, 95% CI: 1.183 to 9.219 OR: 3.3) and Clr (p=0.010, 95% CI: 0.968 to 0.996 OR: 0.982). No statistical significance was obtained in relation to age, KF, HF, dyslipidaemia, sepsis, PN duration and HC/kg, these variables were significant in the bivariate analysis.Conclusion DM, previous elevated blood glucose levels, critical patient, low Clr and treatment with corticosteroids are predictive factors of developing hyperglycaemia, so it would be convenient to consider them in the design of the PN formula.References and/or Acknowledgements I wish to acknowledge the help provided by the pharmacy membersNo conflict of interest