Antimalarial drugs are usually given on a mg kg−1 basis. We have reported previously the population pharmacokinetics of mefloquine derived from data of patients treated for uncomplicated Plasmodium falciparum malaria as part of trials conducted between 1990 and 1995 [1]. The patients were from the Karen ethnic minority who lived in camps for displaced persons situated along the western border of Thailand. For the studies conducted after 1994 the number of precise mefloquine tablets given was recorded. The patients received one of two mefloquine (Lariam, 250 mg base tablets, Roche Pharmaceuticals) dosing regimens in combination with artesunate (12 mg kg−1 over 3 days): 25 mg kg−1 (single dose) or 25 mg kg−1 (split dose: 15 mg kg−1 on day 0 followed by 10 mg kg−1 24 h later). Using the actual dose (mg) in the population pharmacokinetic model as opposed to the weight adjusted dose (mg kg−1), the relationship between the derived pharmacokinetic parameters and body weight was investigated.
A one-compartment model with first-order absorption and first-order elimination was found to describe the data adequately [1]. As very few mefloquine concentrations were recorded during the absorption phase, the absorption rate constant was set to a constant value of 7.0/day derived from a pharmacokinetic study [2] of patients with malaria from the same population. The fundamental pharmacokinetic parameters used to characterize the one-compartment model were apparent clearance (CL/F) and apparent volume of distribution (V/F). The nonlinear mixed effects modelling procedure [3] of the S-PLUS data programme (S-PLUS 4.5 for Windows, Mathsoft, Inc., Cambridge, Massachusetts) was used to calculate estimates of the population pharmacokinetic parameters (CL/F and V/F) and their respective interpatient variances. Individual pharmacokinetic parameters were obtained as posterior estimates from the nonlinear mixed effects modelling procedure. The model building process and the procedures for accepting and rejecting models were identical to those published previously [1]. Measurements of whole blood mefloquine concentration were performed by h.p.l.c. as described previously [4]. The interassay coefficients of variation were 8.3% and 5.7% at 100 and 1000 ng ml−1, respectively.
For those patients given a single dose of mefloquine as part of combination therapy (n = 74) the posterior patient specific estimates for CL/F and V/F were significantly positively correlated with body weight. Incorporating body weight as a covariate for both CL/F and V/F in the model reduced the interpatient variability from 66 to 49% for CL/F and from 50 to 34% for V/F. At the minimum (13 kg), mean (25.6 kg) and maximum (43 kg) body weights the respective population estimates of CL/F would be 13.34, 33.0 and 60.14 l day−1, and for V/F would be 192.3, 383.9 and 648.3 l, indicating a linear relationship with body weight. Similar results were obtained for patients receiving a split dose of mefloquine as part of combination therapy (n = 31) (Table 1).
Table 1.
Mefloquine population pharmacokinetic parameters for the final models where adjusting for body weight gave improved fits.
| Combined therapy and single dosing (74 patients) Estimate (s.e.) | 90% prediction intervals | Combined therapy and split dosing (31 patients) Estimate (s.e.) | 90% prediction intervals | |
|---|---|---|---|---|
| CL/F (l day−1)* | 33.0 (2.4) | 14.9, 73.3 | 23.8 (2.2) | 12.9, 43.9 |
| Change in CL/F (l day−1) for anincrease of 1 kg in body weight | 1.6 (0.3) | 0.9 (0.3) | ||
| V/F (l)* | 383.9 (18.3) | 220.50, 668.24 | 330.6 (19.2) | 217.4, 502.9 |
| Change in V/F (l) for an increase of 1 kg in body weight | 15.2 (2.1) | 11.9 (2.7) | ||
| ke (elimination rate)* | 0.09/day | 0.07/day | ||
| t½ (elimination half life)* | 8.1 days | 9.6 days | ||
| σε (ng ml−1) | 295.6 | 259.2 |
For a person having a mean body weight; s.e., standard error.
The original decision to adjust the dose of mefloquine according to body weight was an empirical decision that had not been tested formally. There is no evidence that the clearance of mefloquine is dose-dependent [5]. These data suggest a linear relationship between clearance, and volume of distribution, and body weight over the range between 13 and 43 kg.
Previous population pharmacokinetic modelling [1], where we used the weight adjusted dose (mg kg−1), found no correlation between body weight and the posterior patient specific estimates for both CL/F and V/F. Thus our present and previous findings confirm that the current recommendations for mefloquine administration, both for nonobese adults and children, that the dose of mefloquine should be adjusted for body weight, are correct. Further studies on heavier adults are needed to determine if this linear adjustment is still appropriate for overweight individuals.
References
- 1.Simpson JA, Price R, ter Kuile F, et al. Population pharmacokinetics of mefloquine in patients with acute falciparum malaria. Clin Pharmacol Ther. 1999;66:472–484. doi: 10.1016/S0009-9236(99)70010-X. [DOI] [PubMed] [Google Scholar]
- 2.Nosten F, ter Kuile F, Chongsuphajaisiddhi T, et al. Mefloquine pharmacokinetics and resistance in children with acute falciparum malaria. Br J Clin Pharmacol. 1991;31:556–559. doi: 10.1111/j.1365-2125.1991.tb05581.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Lindstrom MJ, Bates DM. Nonlinear mixed effects models for repeated measures data. Biometrics. 1990;46:673–687. [PubMed] [Google Scholar]
- 4.Edstein MD, Lika ID, Chongsuphajaisiddhi T, Sabcharoen A, Webster HK. Quantification of Fansimef components (mefloquine+sulfadoxine+pyrimethamine) in human plasma by two high performance liquid chromatographic methods. Ther Drug Monit. 1991;13:146–151. doi: 10.1097/00007691-199103000-00010. [DOI] [PubMed] [Google Scholar]
- 5.Karbwang J, White NJ. Clinical pharmacokinetics of mefloquine. Clin Pharmacokinet. 1990;19:264–279. doi: 10.2165/00003088-199019040-00002. [DOI] [PubMed] [Google Scholar]