A pharmacokinetic-pharmacodynamic assessment of the hepatic and bone-marrow toxicities of the new trypanoside fexinidazole

Fexinidazole is a novel oral treatment for Trypanosoma brucei gambiense human African trypanosomiasis: g-HAT. Fexinidazole is also active against other kinetoplastid parasites, notably T. cruzi the causative agent of Chagas disease.During the course of a dose ranging assessment in chronic indeterminate Chagas disease, delayed neutropenia and significant increases in hepatic transaminases were observed and clinical investigations were suspended. We retrospectively analyzed all available pharmacokinetic and pharmacodynamic data on fexinidazole in normal healthy volunteers and in patients with chronic Chagas disease and g-HAT, in order to assess the determinants of toxicity. A population pharmacokinetic model was fitted to plasma concentration data on the bioactive fexinidazole sulfone metabolite from three phase 1 studies, two g-HAT phase 2/3 field trials and one Chagas phase 2 field trial (462 individuals in total). Bayesian exposure-response models were then fitted to hematological and liver related pharmacodynamic outcomes in chronic Chagas patients. Neutropenia, reductions in platelet counts, and elevations in liver transaminases were all found to be exposure and thus dose dependent in patients with chronic Chagas disease. Clinically insignificant transient reductions in neutrophil and platelet counts consistent with these exposure-response relationships were observed in the g-HAT trials. In contrast, no evidence of hepatotoxicity was observed in the g-HAT trials. Fexinidazole treatment results in a dose dependent liver toxicity and transient bone marrow suppression in Chagas disease. Regimens of shorter duration should be trialled for Chagas. The currently recommended regimen for sleeping sickness provides exposures within a satisfactory safety margin for bone marrow suppression and does not cause hepatotoxicity.

: Summary of the available pharmacokinetic data. These relate only to the quantification of the metabolite M2. The phase 1 data were used to evaluate the best structural model, and the full dataset was used in the pharmacokinetic analyses in order to estimate the drug exposures in fexinidazole treated chronic Chagas patients and fexinidazole treated g-HAT patients.  Comparison between the observed concentrations of the fexinidazole sulfone metabolite M2 and the conditionally weighted residuals of the final pharmacokinetic model fit to all M2 data from phase 1 trials (blue triangles), g-HAT sleeping sickness trials (green crosses) and asymptomatic Chagas disease trials (red circles).
elevations of ALT and AST were observed between 50 and 100 days after the start of treatment. 80 For the patients whose levels rose above 3 times the ULN (for males: 53 and 46 units/L for ALT and  with the exposure intervals in the g-HAT regimen ( Table 2). For AST, the 80% credible intervals 93 of the marginal posterior distribution over the ED 50 are above the exposure intervals in the g-HAT 94 regimen ( Table 2). The relationship between AUlogC and mg/kg dose in the Chagas trial is shown

134
Median hemoglobin counts had also dropped by 1.3 g/dL by day 67, the population median 135 nadir value (Fig 5, bottom right panel). In contrast, lymphocyte counts showed no clear trend 136 (Fig 5, bottom left panel).

145
For reductions in neutrophil counts, the credible intervals over the EC 50 parameters in both 146 the absolute and relative models are above g-HAT exposures ( Table 2). However for the reductions 147 in platelet counts, the absolute model suggests an EC 50 within the g-HAT exposures ( Table 2).  whereas in all field trials the drugs were administered after food). The final estimation of the 385 model parameters used all data from Phase 1, g-HAT and Chagas field trials. Variation between these phase 1 and treatment trials (study and disease effects) were considered only as changes to 387 the relative "absorption" parameter F (scaling parameter on F). The data from both the g-HAT 388 and Chagas field trials are too sparse to estimate changes in absorption rate, clearance, or volume  To estimate the exposure-response curves for the hematological parameters and the liver toxicity 433 outcomes, we fitted four parameter sigmoid functions, defined as: where f is the pharmacodynamic outcome of interest modeled as a function of drug exposure x; 435 E min is the baseline mean outcome under no or negligible drug exposures; E max is the asymptotic 436 maximal effect for high drug exposures; EC 50 is the drug exposure corresponding to Emax+Emin  This was a post-hoc PK-PD analysis and as such was prone to data dependent analyses and false 446 positive results (12). We attempted to minimize this danger by avoiding an analysis contingent on This is an intuitive measure of the overlap between two arbitrary densities of same support f 457 and g, with values varying between 1 (complete agreement) and 0 (complete disagreement).

458
Predicting toxicity in the fexinidazole regimen recommended for g -HAT

459
The trial of fexinidazole in chronic Chagas disease had regular full blood counts (weekly up to 460 week 10, then monthly) providing reasonable confidence that the observed time-series patterns 461 for hematological parameters and liver function are indicative of the true underlying patterns.

462
For the three g-HAT field trials, weekly blood counts in the weeks following treatment were not  (14)) we cross predicted 469 using only the relative models (nadir or peak values divided by baseline value).  The blue dots show the relationship between mg/kg total dose received and the AUlogC exposure.     Watkins for review of liver safety findings.