Establishment of a mathematical prediction model for voriconazole stable maintenance dose: a prospective study

Abstract

Background: In patients with invasive fungal infection (IFI), the steady-state serum trough concentration (Cmin) of voriconazole (VCZ) is highly variable and can lead to treatment failure (Cmin < 0.5 mg/L) and toxicity (Cmin ≥ 5.0 mg/L). However, It remains challenging to determine the ideal maintenance dose to achieve the desired Cmin level quickly. Aims: This randomized, prospective observational single-center study aimed to identify factors affecting VCZ-Cmin and maintenance dose and create an algorithmic model to predict the necessary maintenance dose. MeThe study enrolled 306 adult IFI patients, split into two groups: non-gene-directed (A) (where CYP2C19 phenotype is not involved in determining VCZ dose) and gene-directed (B) (where CYP2C19 phenotype is involved in determining VCZ dose). Results: Results indicated that CYP2C19 genetic polymorphisms might significantly impact VCZ loading and maintenance dose selection. CYP2C19 phenotype, C-reaction protein (CRP), and average daily dose/body weight were significant influencers on VCZ-Cmin, while CYP2C19 phenotype, CRP, and body weight significantly impacted VCZ maintenance dose. A feasible predictive formula for VCZ stable maintenance dose was derived from the regression equation as a maintenance dose (mg) =282.774-0.735×age (year)+2.946×body weight(Kg)-19.402×CYP2C19 phenotype (UM/RM/NM:0, IM:1, PM:2)-0.316×CRP (mg/L) (p < 0.001). Discussion: DiThis formula may serve as a valuable supplement to the Clinical Pharmacogenetics Implementation Consortium (CPIC®) guideline for CYP2C19 and VCZ therapy, especially for IFI patients with highly variable inflammatory cytokines during VCZ therapy.