High rifampicin peak plasma concentrations accelerate the slow phase of bacterial decline in tuberculosis patients: Evidence for heteroresistance
Fig 2
Overview of the fitting process and the relationship between the decline rates of bacteria and pharmacokinetic measures.
We have grouped participants (Fig 2A) by forming equal groups within the pharmacokinetic measures. For each group, we pooled the bacterial count measurements (Y-axis) of all participants (dots, here one colour corresponds to one participant) and fitted biphasic curves to them (solid line) (Fig 2B). Next, we analysed the relationship between the groups’ median pharmacokinetic measurements (X axis) and the properties of the fitted curves (Y axis) (Fig 2C). Fig 2D summarizes the dependence of the decline rates in the quick and slow phases on and
. Here, we have normalized the values (by dividing with the maximum), in order to be able to show that both measures produce the same result but with different errors. Fig 2E demonstrates how this affects the bacterial count measurements (X-axis) over time (Y-axis). The curves were made with Eq (1), starting at 106 CFU/ml bacteria, using day 4 as the day of transition, with the obtained fits for decline rates (see Table 1), and Cmax values of 10, 20, 30…80 [mg/l] (in the dataset the Cmax values range from 7.7 to 85.6 mg/ml).