Aims: The purpose of this study was to establish the relevance of several clinical factors associated with parenteral nutrition (PN) hypertriglyceridemia and to construct a predictive model for this complication.
Method: This multicenter study included all patients with initial serum triglyceridemia <3 mmol and receiving a minimum of 7 days' PN therapy. The study ended for each patient when hypertriglyceridemia developed or PN was terminated. Two multivariate models were constructed, one to study the clinical factors and the second to predict plasma triglyceridemia. A total of 22 clinical factors studied as independent variables were included in the multiple-step regression models only when they showed a P-value over 0.1. Statistical significance was determined by the confidence interval of the odds ratio (OR) and the partial regression coefficient (b).
Results: The study included 260 patients from 14 hospitals. Lipid administration was 0.83+/-0.37 g/kg/day. Among the total, 68 patients (26.2%) showed hypertriglyceridemia. Variables included in both models were serum glucose (OR, 2.63; b, 0.06), renal failure (OR, 10.56; b, 1.70), corticoid administration >0.5 mg/kg (OR, 7.98; b, 0.97), pancreatitis (OR, 4.38; b, 0.64), sepsis (OR, 4.48; b, 0.24), lipids infused (OR, 3.03; b, 0.24) and heparin administration >3 mg/kg/day (OR, 0.11; b, -1.21).
Conclusion: Although the rate of lipid infusion was low, certain clinical factors modified triglyceridemia. Nevertheless, relatively fast plasma clearance of lipids infused indicates that a reduction in lipid supply could be a quick, effective measure for controlling hypertriglyceridemia. Thus, careful monitoring of patients with clinical factors predicting risk in the model studied, with adjustment of lipid perfusion rates accordingly, is suggested to avoid hypertriglyceridemia.