Disposition of Intravenous Pyrimethamine in Healthy Volunteers

Abstract

A proportion of patients with AIDS and toxoplasmic encephalitis (TE) sustain low plasma pyrimethamine concentrations during oral treatment, possibly because of incomplete and variable bioavailability. We wanted to develop a safe, practicable intravenous (i.v.) formulation of pyrimethamine and characterize its disposition in healthy volunteers. A neutral, aqueous, sterile solution of pyrimethamine was produced and presented in sealed glass ampoules. Pyrimethamine (1 mg/kg) was given to eight healthy male volunteers by i.v. infusion over 2 h, and blood was sampled over a 2 week period. Pyrimethamine levels in plasma were measured by high-performance liquid chromatography. The drug was well tolerated by all volunteers, and there were no changes in vital signs, electrocardiogram, hematology, or biochemical parameters. The maximum pyrimethamine concentration of 2,089 ± 565 ng ml⁻¹ (mean ± standard deviation) was achieved shortly after the end of the infusion; thereafter, concentrations declined in a log-linear manner, with a half-life of 140 ± 31 h.

Toxoplasma gondii encephalitis (TE) is the second most common AIDS-related opportunistic infection of the central nervous system, occurring in up to 50% of patients with AIDS who are seropositive for antibodies to the parasite and have CD4⁺ T-lymphocyte counts of <100/mm (11). Chemoprophylaxis, usually with co-trimoxazole, and highly active antiretroviral therapy reduces the incidence of TE in industrialized nations (8) but not in the developing world, where these are often less readily available (7). The treatment of first choice for TE is pyrimethamine in synergistic combination with either sulfadiazine or clindamycin (4). The clinical response to pyrimethamine-sulfadiazine is highly variable: in one study, 50% of patients responded by day 3, 85% responded by day 7, and 90% responded by day 14, but about 10% of patients with verified TE (by biopsy or necropsy) showed no response (9). Treatment failure may be due to subtherapeutic drug concentrations, variability in parasite chemosensitivity, or host factors such as degree of immunosuppression. The first of these factors deserves focused attention, as it is amenable to adjustment.

In AIDS patients being treated for TE with enterically administered (oral or nasogastric) pyrimethamine, concentrations in plasma vary markedly (14, 15), and we have hypothesized that differences in oral bioavailability account for this. Furthermore, it is possible that particularly low pyrimethamine concentrations are sustained by the sickest patients who are incapable of swallowing the drug, to whom it is given via nasogastric tubes (15). In these cases, sulfadiazine and clindamycin may be given intravenously (i.v.), but there is no i.v. formulation of pyrimethamine. The intramuscular formulation of pyrimethamine with sulfadoxine (as Fansidar parenteral; Hoffman LaRoche) is not suitable for use in TE: repeated intramuscular injection is impracticable, while sulfadoxine has low potency against T. gondii (2) and can cause severe allergic reactions. We therefore wanted to develop an i.v. preparation of pyrimethamine and describe its disposition in healthy volunteers.

(This work was presented in part at the Federation of Infectious Diseases Society, Manchester, United Kingdom, 29 November 1997.)

Pyrimethamine free base was a gift from Glaxo-Wellcome plc (Uxbridge, United Kingdom) and sulfuric acid BP was purchased from Thornton & Ross Ltd. (Huddersfield, United Kingdom). To pyrimethamine powder (12.5 g) in distilled water (3.5 liters) was added sulfuric acid (0.2 M; 250 ml); the solution was stirred until the pyrimethamine had dissolved. The final volume was adjusted to 5 liters with distilled water (final pH, 3.0 to 6.0); this was sterilized by filtration (0.22-μm-pore-size filter) and aseptically transferred to glass ampoules (10 ml containing pyrimethamine 25 mg of the base).

The Ethics Committee of the Royal Liverpool University Hospital gave approval for the study. Eight healthy male volunteers, between the ages of 26 and 43, were recruited and gave written informed consent. (Written informed consent was obtained from all the human volunteers in the present study in accordance with the guidelines of the Ethics Committee of the Royal Liverpool University Hospital.) Volunteers were excluded if there was (i) a history of adverse reaction to pyrimethamine, (ii) abnormality of hematological or biochemical parameters (including serum folate), (iii) history of cardiac disease or abnormality of the resting electrocardiogram, (iv) regular consumption of concomitant medication, (v) abnormality of physical examination, or (vi) known positive test for human immunodeficiency virus.

Teflon cannulae were inserted into the forearm veins of both arms: one for sampling, the other for the infusion. Pyrimethamine (1 mg of the base per kg of body weight) was diluted in 0.9% saline to a final volume of 50 ml; this was infused by an electronic syringe driver at a constant rate over 2 h. Pulse rate, blood pressure, oral temperature, and the infusion site were checked before the start of the infusion and at 30-min intervals until 1 h after the completion of infusion.

Subjects were removed from the study if (i) consent was withdrawn, (ii) there was extravasation of the infusion, (iii) systolic or diastolic blood pressure fell by more than 20 mm of Hg, (iv) consciousness was perturbed, (v) ECG complex or rhythm abnormalities developed, (vi) there was pruritus or rash, or (vii) oral temperature rose above 37.5°C.

Blood (10 ml) was drawn immediately before the dosing and at and at 30, 60, 120, 125, 130, 135, 140, 150, and 160 minutes and 4, 6, 8, 12, 24, 48, 72, 96, 120, 168, and 336 hours after the start of the infusion. Blood was taken into lithium heparin-impregnated tubes and was then centrifuged (about 1,000 × g for 15 min) within 1 h of sampling. Plasma was transferred into anticoagulant-free tubes and stored at temperatures below −20°C until assay.

Assay was by high-performance liquid chromatography using a previously validated method (2). The system used consisted of a SP8700 quaternary-gradient high-performance liquid chromatography pump connected to a SP8000 autosampler. Detection was achieved with a SP100 UV-visible variable wavelength absorbance detector operating at 254 nm (a suitable wavelength giving near-maximal absorbance without interference), with the results interpolated using a SP4400 computing integrator. Chromatographic separation was achieved at ambient temperature using a reversed-phase phenyl μ-Bondapak cartridge (inner diameter, (100 by 8 mm; 10-μm particle size) in a Waters radial compression Z-module, with a CN-Guard Pak precolumn to protect the analytical column. Separation of pyrimethamine and proguanil (the internal standard for the assay) was achieved using a mobile phase consisting of water-acetonitrile-methanol (60:30:10, vol/vol) containing 1-octane sulphonic acid (5 mM) buffered to pH 2.5 with hydrochloric acid, flowing at 3.0 ml min⁻¹. The approximate operating back pressure under these conditions was 750 ± 50 lb/in². Pyrimethamine and proguanil were resolved to baseline with symmetrical peaks and corresponding retention times of 8.5 and 12 min, respectively. Chromatograms obtained from extractions of drug-free plasma showed no evidence of interfering peaks at the retention times described. The analytical recoveries of pyrimethamine and proguanil under the conditions described above were 92 and 87%, respectively, compared with authentic aqueous standard solutions injected directly onto the chromatograph. The inter- and intra-assay coefficients of variation were determined at a pyrimethamine concentration of 50 ng ml⁻¹. The mean (n = 6) coefficients of variation were 3.8 and 4.5% respectively. The lower limits of detection were 5 ng ml⁻¹ (pyrimethamine) and 1 ng ml⁻¹ (proguanil).

The maximum plasma concentrations and the times at which these were achieved were obtained by inspection of concentration-time profiles. Data were entered into a nonlinear regression program (Topfit; Schering, Munich, Germany). A two-compartment model provided the closest iterative fit, and all pharmacokinetic variables were derived from this.

The infusion was well tolerated; blood pressure, temperature and ECG parameters remained acceptable during the infusion and for 1 h thereafter. No subjective adverse events were reported, and hematological and biochemical parameters were unchanged when checked after completion of the follow-up period. All eight subjects completed the study. Figure 1 shows mean plasma pyrimethamine concentrations over the study period.

FIG. 1.

Mean (± standard deviation [error bars]) plasma pyrimethamine concentration versus time postdose (0 to 336 h) following i.v. administration of pyrimethamine (1 mg kg⁻¹).

We wanted to develop a safe i.v. pyrimethamine formulation which would be practicable (both from the point of view of in-house formulation and administration) in developing nations. We have described a simple method for the preparation of an aqueous, neutral pyrimethamine sulfate solution and have given this to healthy volunteers as i.v. infusions. The process of formulation should be possible in those national referral hospitals which treat opportunistic infections and could be established within generic drug manufacturing houses throughout the developing world.

Pyrimethamine had not previously been given i.v. to humans, and we were concerned about achieving safe concentrations with the present i.v. dose, the dose size and rate of infusion being the main determinants of this. The dose of a new i.v. drug formulation would usually be derived as a product of oral dose and bioavailability. Unfortunately the oral bioavailability of pyrimethamine is unknown in humans (although it has previously been assumed to be near complete [5]). In animal species bioavailability is thought to range from near-complete (12) to 0.56 (3). Consequently, for the present study, we opted to assume that the oral bioavailability of pyrimethamine is incomplete in humans and fixed the i.v. dose at a cautious 1.0 mg of pyrimethamine base kg⁻¹ i.v., rather than the standard oral loading dose of 2.0 mg kg⁻¹. We fixed the i.v. infusion rate with reference to the rate of absorption of pyrimethamine after intramuscular administration in humans, for which, 2 h after injection of 1.25 mg/kg, concentrations rise to about 75% of the maximum concentration in plasma, i.e., around 300 ng/ml (10, 17). Infusion of 1 mg/kg in 50 ml i.v. over 2 h was well tolerated: there was no local discomfort, subjective or objective adverse effects, or changes in vital signs or ECG parameters.

Plasma pyrimethamine concentration-time profiles showed relatively little intersubject variability after i.v. infusion (the coefficient of variation of the area under the concentration-time curve from 0 h to infinity [AUC_0–∞] was 37%), in contrast to those achieved after oral administration (where the coefficient of variation of AUC_0–∞ is about 62% [10]). Although the present data were derived from healthy subjects and not patients, this relatively low variability is reassuring: the i.v. formulation may, in due course, be relied upon to achieve effective drug concentrations rapidly while avoiding potentially toxic levels in vulnerable patients. The disposition of the present i.v. formulation of pyrimethamine now needs to be studied in patients with AIDS and TE.

The mean elimination half-life of pyrimethamine was approximately 140 h, and this figure is in broad agreement with estimates made after oral dosing, which have ranged from 35 to 175 h (1, 5, 6, 13). The present estimates for AUC_0–∞ (and derived parameters, such as clearance and volume of distribution) vary markedly from those after oral dosing. An i.v. dose of pyrimethamine of 1 mg/kg achieved a value of 332 mg/liter/h in the present study, whereas 0.3 mg/kg achieved 19 mg/liter/h after oral administration (13). We are unable to calculate absolute oral bioavailability from these results but estimate the parameter to be below 0.5 in healthy normal volunteers. If this is correct it lends weight to our previous suggestion that the variability of pyrimethamine concentrations after oral dosing in patients with TE is due to differences in bioavailability; we are in the process of studying this possibility.

TABLE 1.

Calculated pharmacokinetic parameters after i.v. administration of pyrimethamine to eight healthy subjects^a

Subject	Cmax (μg/ml)	Tmax (h)	AUC0–336 (mg/liter/h)	AUC0–∞ (mg/liter/h)	V (liters/kg)	CL (liters/kg/min)	t1/2 (h)
1	1,130	2.25	189	260	1.04	3.98	187
2	2,610	24.0	309	432	0.52	3.36	156
3	2,980	2.03	390	556	0.46	2.13	176
4	1,860	2.12	132	145	0.87	10.8	87
5	2,600	2.53	286	349	0.53	4.83	128
6	2,050	24.0	261	335	0.62	3.18	145
7	1,680	3.0	207	238	0.66	4.83	109
8	1,800	2.45	212	263	0.71	4.05	130
Mean (SD)	2,089 (565)		248 (75)	322 (119)	0.68 (0.20)	4.65 (2.47)	140 (31)
Median (range)		2.49 (2.03–24)

^aC_max, maximum concentration of drug in serum; T_max, time to C_max, V, volume of distribution; CL, clearance; t_1/2, half-life.