Article Text

Download PDFPDF
Design and validation of a predictive equation to estimate unbound valproic acid concentration
  1. Silvia Conde Giner,
  2. Maria Dolores Belles Medall,
  3. Raul Ferrando Piqueres
  1. Pharmacy Department, Hospital General de Castellón, Castellon de la Plana, Spain
  1. Correspondence to Silvia Conde Giner, Pharmacy Department, Hospital General de Castellón, Castellon de la Plana, Comunidad Valenciana, Spain; silviacondeginer{at}gmail.com

Abstract

Objectives Total plasma levels of valproic acid (VPA) may mask an increased risk of adverse effects in hypoalbuminaemic patients since, in these patients, the free fraction is higher. The aim of this study is to analyse the relationship between plasma levels of total and free VPA (FVPA) in hypoalbuminaemic patients and define an equation that allows the estimation of FVPA concentration, as well as to validate the obtained equation.

Methods This is a retrospective observational study conducted between January 2015 and January 2020. Hypoalbuminaemic adult patients with normal renal function were included. Serum VPA levels were determined using an automated enzyme immunoassay technique with a pre-treatment of the sample by ultrafiltration for the quantification of FVPA. Patients’ determinations were randomised into two groups: first, to calculate the FVPA estimation equation (regression group) by multiple linear regression analysis; and second to validate the equation (validation group), calculating the agreement between experimental and estimated FVPA concentrations using Lin’s coefficient and a Bland and Altman analysis.

Results We included 51 determinations, corresponding to 33 patients: 26 in the regression group, and 25 in the validation group. The multiple linear regression analysis showed a statistically significant relationship between FVPA concentration (Y), total VPA concentration (X1) and albumin level (X2), explained by the equation Y=11.882 + 0.216*X1–4.722*X2. Pearson’s correlation coefficient was 0.798 (p<0.001). Lin’s coefficient was 0.82 (95% CI 0.63 to 0.92). The Bland and Altman analysis showed a bias of 0.32 mg/L, and the concordance limits were between −3.80 and 4.44.

Conclusions The calculated equation adequately predicts FVPA concentration, with a high degree of correlation between the variables. Despite Lin’s coefficient outcome, Bland and Altman analysis showed a minimum bias that slightly underestimates FVPA concentration, positioning the calculated equation as a useful and validated estimation tool in hypoalbuminaemic patients with normal renal function.

  • therapeutic drug monitoring
  • pharmacokinetics
  • pharmacy service
  • hospital
  • safety
  • neurology

Data availability statement

Data are available upon reasonable request.

Statistics from Altmetric.com

Data availability statement

Data are available upon reasonable request.

View Full Text

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.