Exploring variability in paromomycin pharmacokinetics in East African visceral leishmaniasis patients
Luka Verrest (1), Monique Wasunna (2), Gilbert Kokwaro (3,4), Rashid Aman (4), Ahmed Musa (5), Asrat Hailu (6), Fabiana Alves (2), Thomas Dorlo (1)
(1) Department of Pharmacy and Pharmacology, Antoni van Leeuwenhoek Hospital, Amsterdam, (2) Drugs for Neglected Diseases initiative (DNDi), Geneva, Switzerland, (3) KEMRI-Wellcome Trust Programme, Nairobi, (4) African Centre for Clinical Trials, (5) Institute of Endemic Diseases, University of Khartoum, Sudan, (6) Addis Ababa University, Ethiopia
Objectives: There is a high need for new therapies for the neglected tropical parasitic disease visceral leishmaniasis (VL), as effective, safe and affordable treatments are still lacking. Paromomycin sulphate (PM) has been shown to be effective in Indian VL patients[1] and is favourable over other therapies because of its affordability and its reasonable safety profile. However, a similar PM dosing regimen of 15 mg/kg/day for 21 days resulted in a lower efficacy in East Africa[2] and a dose increase or a combination therapy was required to achieve adequate efficacy[3]. In order to obtain a better understanding of the differences between populations, a population PK model of PM in different African patient populations was developed.
Methods: PM PK data was available from a multi-centre randomized controlled trial (RCT) in VL patients from Kenya, Sudan, and Ethiopia[3]. Intramuscular PM monotherapy (20 mg/kg/day for 21 days) was compared to PM plus intravenous sodium stibogluconate (SSG) combination therapy (PM 15 mg/kg/day and SSG 20 mg/kg/day for 17 days). Paromomycin plasma concentrations were obtained using HPLC-UV. In Kenya and Sudan samples were obtained frequently at the first and last day of treatment. In Ethiopian study sites a sparse sampling scheme was used. A population PK model of PM was developed using NONMEM (v 7.3). A total of 388 concentrations from 74 patients were included in the PK analysis. Tested covariates included study site, country, treatment group (monotherapy or combination therapy with SSG), creatinine plasma levels, glomerular filtration rate (GFR), and albumin plasma levels. To evaluate the model fit, a visual predictive check (VPC) was performed.
Results: A one-compartment model with first-order absorption was found to best describe PM in plasma. Typical parameter estimates (CV) were a clearance of 3.86 L/h (8%), central volume of distribution (Vc) of 13.1 L (11%), and an absorption rate constant (ka) of 1.31 h-1(16%). Body weight was included on clearance and Vc, with fixed powers of 0.75 and 1.00, respectively. Ethiopian patients exhibited deviating concentration-time profiles which were best characterized by a 1.94 (25%) times higher bioavailability (FEth) and a 3.19 (20%) times slower ka. Between-subject variability (BSV) was included on clearance (41%) and ka(52%), and an additional BSV for FEth(150%) described the overall higher variability in Ethiopian patients. Additionally, for all patients, there was a decrease in clearance over time amounting to -33.2% between start and end of treatment (day 21), which could not be explained by either GFR, creatinin or albumin. AUC0-tau,SS for 15 mg/kg/day (median [SD]) was significantly higher in Ethiopia (218.2 µg·h/mL [1724.1]) compared to Kenya and Sudan (165.7 µg·h/mL [53.1]). AUC0-tau,SS for 20 mg/kg/day was not significantly different between Ethiopia (240.6 µg·h/mL [1849.8]) and Kenya and Sudan (258.0 µg·h/mL [193.1]). The high variability in Ethiopia was mainly caused by 6 patients with extremely high AUC0-tau,SS levels of >2000 µg·h/mL. These patients were all treated at one of the two study sites in Ethiopia. The VPC showed that the model could adequately predict PM in the different countries, as well as the differences in variability.
Conclusions: The developed PK model of PM in East African VL patients showed a slower ka and higher bioavailability in Ethiopia, compared to Kenya and Sudan, as well as a decrease of clearance over time for all patients. Accordingly, daily exposure in Ethiopian patients was estimated to be up to 30% increased. In order to explain these differences mechanistically, creatinine, GFR, and albumin were tested on different parameters in the covariate analysis, but no significant covariates could be identified. The difference in exposure between the Ethiopian study sites might suggest a drug related difference, although batch numbers were not different between countries. Other possible causes were considered, e.g. different dose calculation, or administration-related differences such as site of injection. However, these factors could not be verified.
Besides the observed differences in PK between East African countries, differences in clinical efficacy with Indian patients were observed. To further understand this diversity, it is planned to compare the PK of PM in East African populations with the Indian population.
References:
[1] Sundar, S. et al. N. Engl. J. Med. 356, 2571–2581 (2007).
[2] Hailu, A. et al. PLoS Negl. Trop. Dis. 4, (2010).
[3] Musa, A. et al. PLoS Negl. Trop. Dis. 6, (2012).
