26 Saliva sampling to facilitate pharmacokinetic studies in children: a case study with metamizole

Introduction Pharmacokinetic (PK) studies in young children and infants are challenging, among others because of the invasiveness of blood samples. Using saliva as matrix could increase their feasibility. A recent PK study of the prodrug metamizole (dipyrone), used as off-label intravenous (i.v.) analgesic in infants < 1 year of age, aimed to assess the PK of its main metabolites after a single i.v. dose of 10 mg/kg. It showed increased plasma exposure to the main active metabolite 4-methylaminoantipyrine (MAA) in four infants of 3–12 months compared to children of 1–6 years. We aimed to evaluate the suitability of saliva PK sampling in this study to facilitate further PK studies in a larger infant population for dose optimization. Methodology Saliva was sampled at up to 6 pre-defined time points until 24 h post-dose. Concentrations of four metabolites were quantified with a novel liquid chromatography-mass spectrometry method. Combined plasma/saliva population-PK modeling was performed to investigate mechanistic aspects of the two active metabolites' distribution (MAA and 4-aminoantipyrine [AA]) and to quantify inter- and intraindividual variability of the metabolites' distribution among the matrices. Results Overall, 25 and 16 children provided at least 1 plasma or saliva sample, respectively, and 26 children were included in the analysis (median [interquartile range] age: 57 [22.0–58.8] months, weight 17 [11.3–18.7] kg). No distribution delay of the metabolites between plasma and saliva could be quantified. The fractions of MAA and AA distributing to saliva were estimated to (95% confidence interval [CI]) 0.32 (0.22–0.42) and 0.57 (0.49–0.65), with logit-normally distributed interindividual variability of 0.81 and 0.55, respectively. A strong correlation between the two predicted fractions of 0.89 was observed, but no correlation with age or albumin concentrations. Residual intraindividual variability in saliva was 0.47 and 0.28 versus 0.17 and 0.11 in plasma for MAA and AA, respectively. Conclusion This analysis indicates instantaneous distribution of active metamizole metabolites from plasma into saliva, approximately to the extent of reported unbound fractions in plasma. Due to high variability in saliva, suitability of saliva as single matrix for PK studies in children is compromised. However, rich saliva sampling in combination with reduced plasma sampling could be evaluated to minimize the burden of PK studies in infants and children.