Main

In Africa, over one million deaths per year are due to malaria, which occurs mostly in children under 5 y of age(1). Quinine is the treatment of choice because it is inexpensive, can be administered via the parenteral route, and is effective against chloroquine-resistantPlasmodium falciparum(2–5). However, as the toxic threshold of quinine (above 10 mg/L)(6) is very close to the therapeutic dose [between 5 and 10 mg/L(3–5) or above 10 mg/L in cerebral malaria according to some authors(2)], it is very important to identify factors, such as nutritional status, which may be suspected of altering the pharmacokinetics of quinine(7).

Salako et al.(8) showed that elimination the half-life is increased and clearance decreased in children suffering from kwashiorkor. However, kwashiorkor is characterized by protein deficiency, and clinical manifestations include inadequate growth, loss of muscular tissue, hepatomegaly, crazy pavement dermatosis, discolored hair, edema, and hypoalbuminemia(9). Moreover, many cases of malnutrition in Africa and in other parts of the world are caused not only by protein deficiency but also by global malnutrition(10–12). Clinical manifestations of global malnutrition are weight loss, absence of subcutaneous fat, muscle atrophy, emaciation, and old man's face without edema(9). According to the Wellcome Party Classification(13), global malnutrition is classified as undernutrition if weight is between 60 and 80% of that expected for age, and marasmus if weight is less than 60%.

In the rat, protein deficiency causes lower metabolic clearance of many drugs, as in kwashiorkor(14). On the other hand, prolonged global food restriction increases both in vivo andin vitro activity of several drug metabolizing enzymes(15). These experimental data could be of clinical importance for the treatment of malaria in malnourished children because malaria and malnutrition often occur together in developing countries. This present study was carried out to investigate the pharmacokinetics of quinine in normally fed and globally malnourished children.

METHODS

Subjects. Fifteen Gabonese children aged from 9 to 60 mo, undergoing treatment for malaria at the Hôpital Provincial, Port-Gentil, Gabon, were enrolled in the study. All of the children had a fever of over 38.5 °C but no evidence of bacteriologic infection (angina, otitis, pneumonia, or urinary infection). The aims and procedures were carefully explained to parents, and informed consent was obtained. For ethical reasons, children with a hematocrit less than 25% were not included in the study.

To assess the influence of nutritional status on the pharmacokinetics of quinine, one investigator measured and recorded the weight, height, head circumference, left upper mid-arm circumference, and skin fold thickness of the posterior side of the left mid-arm of each child. Children were classified into one of the two groups, depending on nutritional status, which was defined using their left mid-arm circumference/head circumference ratio. This ratio is independent of age between 3 and 60 mo, correlates highly with other anthropometric indices, and is very simple to use as it requires only a tape measure(16–18). Normal children (group 1) were defined by a ratio greater than 0.28, and malnourished children (group 2) were defined by a ratio less than 0.279(16). Seven children were included in group 1 (ratio between 0.30 and 0.32) and 8 in group 2 (ratio between 0.20 and 0.27). None of the children had clinical manifestations of kwashiorkor. We also assessed other parameters of nutritional status which included: weight for height and height for age according to the Waterlow classification(19, 20), as well as weight for age, and arm muscle circumference, which correlates highly with protein body mass(21). These data were expressed as mean percentages for age or height and sex (National Center for Health Statistics data). Laboratory examinations revealed the following (mean± SD): total proteins (71 ± 6 g/L), serum albumin (29 ± 5 g/L), prealbumin (0.11 ± 0.03 g/L), transferrin (2.3 ± 0.4 g/L), AGP (1.24 ± 0.34 g/L), creatinine (37 ± 7 mg/L), and hemoglobin(9.6 ± 1.2 g/L). No significant differences were found between the two groups (Table 1). In particular, AGP (with quinine high affinity binding sites) was 1.33 ± 0.33 g/L in the control group and 1.28 ± 0.42 g/L in the malnourished group. Thick and thin blood films of each child were sent to a parasitologic laboratory for Plasmodium research(Pr. Richard Lenoble, Libreville).

Table 1 Comparison of biologic data in control (group 1) and malnourished (group 2) children
Full size table

Study design. Each child was given a loading dose of 16 mg/kg quinine base [25 mg/kg quinine resorcine hydrochloride, Quinimax (Sanofi, Paris, France)] by deep intramuscular injection into one anterior thigh, followed 12 h later by a second dose of 8 mg/kg in the second thigh. Venous blood samples were taken at 0, 1, 2, 4, 6, 12, and 24 h after the first injection. Blood was taken using a butterfly needle with a heparin lock and was collected into lithium heparin tubes. It was then centrifuged at 1000× g for 10 min. Plasma was stored at -20 °C until assay.

Quinine assay. Quinine and 3-hydroxyquinine(22) were assayed in plasma using HPLC with fluorimetric detection(23). Plasma (0.2 mL) was extracted with methylenechlorideisopropanol (80:20) (10 mL) after adding sodium hydroxide, 0.1 N, and internal standard (cupreine, 20 μg). The orgainc phase was taken to dryness, and the residue was dissolved in the mobile phase (250 μL). The mobile phase consisted of acetonitrile-acetic acid-water, 30:20:450, all by volume. The flow rate was 2 mL/min. Separation was performed using aμBondapak C18 column (300 × 4.7 mm, particle size 10 μm) and fluorimetric detection (excitation 250 nm, cut-off filter 389 nm, ABI SF980, Perkin-Elmer, Saint-Quentin, France). Retention times were 4.8, 7.4, and 15.8 min for hydroxyquinine, internal standard, and quinine, respectively. Interassay variability was 3.5-7.3% for quality controls 10 and 3 μg/mL quinine. The minimum level of quantification was 0.25 mg/L for quinine and 0.1 mg/L for hydroxyquinine. As 3-hydroxyquinine is not commercially available, 3-hydroxyquinidine was used.

Quinine binding in plasma was measured in the sample collected 1 h after the first injection using equilibrium dialysis with Dianorm cells (OSI, Paris, France) and Wisking cellophane membrane (Bioblock, Paris, France). A volume of 0.5 mL of plasma was dialyzed against 0.5 mL of Sorensen's phosphate buffer, pH 7.4, for 6 h at 37 °C. After equilibration, quinine concentrations in plasma (B) or buffer (T) were measured as already described, except that calibrations and quality control measurements were performed with either plasma or buffer with quinine added. Percentage of binding was calculated as (B - T)·100/T.

Pharmacokinetic calculations. Peak plasma concentrations(Cmax) and time to reach the peak (Tmax) were the recorded values. The elimination rate constant (k) was estimated by nonlinear regression analysis of the terminal phase of plasma concentrations; t1/2 was the elimination half-life(t1/2 = 0.693/k). AUC was calculated by the linear trapezoidal rule and extrapolated to infinity. Body clearance was Cl/F = D/AUC where D is the injected dose. Volume of distribution was calculated asVd/F = D/(AUC ×k) (F is the factor of bioavailability).

Statistical analysis. The averages of clinical and laboratory recordings were compared using the Mann-Withney test. Correlation coefficients were computed using the Spearman correlation rank. Probability (p) values less than 0.05 were considered statistically significant. Values are given as mean ± SD.

RESULTS

All the anthropometric measurements, except height for age, differed significantly between the two groups and were significantly correlated with arm/head circumference ratio (Table 2). Skin fold thickness which reflects total body fat(21) and muscular perimeter, indices of protein body mass, were both lowered in group 2. Blood film was positive for P. falciparum for only one child in each group.

Table 2 Comparison of anthropometric data in control(group 1) and malnourished (group 2) children, correlation with left mid-arm/head circumference ratio
Full size table

The mean profile of plasma concentration for each group is shown inFigure 1. Tmax was significantly longer in normal children (2.2 ± 1.2 h) than in malnourished children (1.1± 0.4 h) (Table 2). No difference was observed between the two groups for Cmax, volume of distribution, or protein binding. In the malnourished children, clearance was significantly faster (4.4 ± 3.6 versus 2.3 ± 1.4 mL·min-1/kg, p < 0.05) and half-life shorter (6.3± 1.8 versus 10.1 ± 3.4 h, p < 0.05) than in the normally fed children. Concentration at 12 h was lower in malnourished children (3.3 ± 1.6 mg/L) than in control children (5.3 ± 1.6 mg/L). There was a significant correlation between elimination half-life and left mid-arm<ead circumference ratio (Fig. 2). The pharmacokinetic parameters of 3-OH quinine are shown in Table 3. The ratio between AUC of hydroxyquinine and AUC of quinine(Rm) was significantly higher in group 2 compared with controls and was significantly correlated with left mid-arm/head circumference ratio (r = 0.71, p < 0.05)(Fig. 3).

Figure 1
figure 1

Concentrations of quinine (mean ± SD); □, control children; ▪, malnourished children.

Full size image
Figure 2
figure 2

Relationship between half-life and mid-arm/head circumference ratio; ○, control children; •, malnourished children.

Full size image
Table 3 Pharmacokinetics of quinine and hydroxyquinine in control (group 1) and malnourished (group 2) children
Full size table
Figure 3
figure 3

Relationship between AUC of hydroxyquinine/AUC of quinine ratio (Rm·10-3) and left mid-arm/head circumference ratio; ○, control children; •, malnourished children.

Full size image

There was a negative correlation between AGP concentration and volume of distribution per kg (r = -0.71, p < 10-3)(Fig. 4), but not between mid-arm/head circumference ratio and volume of distribution (Fig. 5). Age was not significantly correlated with pharmacokinetic parameters.

Figure 4
figure 4

Relationship between the AGP level and volume of distribution (Vd/F, with F factor of bioavailability); ○, control children; •, malnourished children.

Full size image
Figure 5
figure 5

Relationship between left mid-arm/head circumference ratio and volume of distribution (Vd/F with F factor of bioavailability); ○, control children; •, malnourished children.

Full size image

DISCUSSION

P. falciparum malaria is life-threatening(24) and prevalent in Central Africa(25, 26). Reading blood films requires experience and time. In most regions in Africa, children are treated on the basis of clinical suspicion(27). All children with fever and without clinical evidence of bacterial infection receive antimalarial treatment. However, as we found only only two children with parasitemia, we suggest that this approach needs to be reconsidered, with local conditions and seasonal variations being taken into account. Although an interesting side finding, it was beyond the scope of the present study and needs to be investigated in a further study.

The pharmacokinetics of quinine is modified in all patients with fever whatever the cause(28–30). These modifications are caused by quinine fixation on AGP(31, 32) and an increased AGP plasma level in children with infection(33). This is the likely explanation for the correlation we found between AGP and volume of distribution. Mean AGP level was high: 1.24 ± 0.34 g/L (normal level is under 0.7 g/L) and no statistical difference was found for AGP, volume of distribution, or for unbound quinine between the two groups.

However, quinine clearance and the ratios between AUC of hydroxyquinine and quinine were significantly higher in malnourished children. As 3-hydroxyquinine is the major metabolite of quinine(22), an increase in the ratio between AUC of hydroxyquinine and AUC of quinine strongly suggests an increase in hepatic metabolism of quinine. Although a decrease in renal excretion of 3-OH quinine in malnourished children might also account for these observations, this hypothesis is highly improbable because quinine clearance was also faster in malnourished children. Furthermore, levels of urea and creatinine were low in this group and did not differ from control children.

The findings of Salako et al.(8) concerning the pharmacokinetics of quinine in kwashiorkor differed dramatically from the results we obtained. Their study was carried out in seven healthy African children and six children suffering from kwashiorkor (edema and low plasma total protein and albumin levels). None of the infants had malaria. Quinine was administered orally. Elimination half-life was significantly longer (15.0± 4.4 versus 8.0 ± 1.3 h) and oral clearance significantly slower (32 ± 9 versus 108 ± 35 mL min-1) in kwashiorkor than in control subjects. However, the level of AGP and volume of distribution were not assessed. Some authors have shown that AGP is increased in kwashiorkor(7, 34). Therefore, it is very difficult to interpret these results because an increase in AGP could decrease the unbound quinine and consequently its hepatic uptake.

However, there is some evidence in the literature that protein deficiency decreases hepatic metabolism of drugs in human and animals. Buchanan et al.(35) studied the pharmacokinetics of antipyrine in kwashiorkor children. Antipyrine clearance was prolonged at admission and shortened significantly after recovery. Oxidative metabolism of acetanilide was also decreased in children with kwashiorkor(36). Feldman et al.(37) observed a lower clearance of theophylline in children with a low protein diet compared with healthy children. In the rat, total body clearance of theophylline was decreased in protein-restricted animals(38). Total microsomal cytochrome P-450 content and NADPH-cytochrome P-450 reductase activity were also lowered(39, 40). Measured cytochrome P-450 activity was decreased using various substrates: ethylmorphine, phenobarbital, benzo(a)pyrene(14). We found no study in humans which focused on modifications of drug metabolism in children with global malnutrition. In the rat, global food restriction increased total content of cytochrome P450 and also in vivo and in vitro activity of several drug metabolizing enzymes(14). Feed restriction of 45% for 28 d in weaning male rats enhanced hepatic microsomal proteins, total content of cytochrome P-450, NADPH-generating enzymes, several cytochrome P-450 activities (aniline hydroxylase,p- chloromethylaniline-N-demethylase, and decreased hexobarbital sleeping time (hexobarbital is a substrate of cytochrome P-450 monooxygenases)(15). These observations were confirmed by both Chou et al.(41) and Alterman et al.(42). Also, Hong et al.(43) showed that fasting in rats induced a specific form of cytochrome P-450 (p-450j).

Thus, it appears that results obtained in humans are similar to experimental data obtained in animals: hepatic metabolism of quinine is decreased in protein deficiency and increased in global malnutrition. As this is the first description of an increase in drug metabolism in malnourished children, confirmation with other drugs is needed. Notwithstanding, our results indicate that hepatic metabolized drugs should be carefully monitored in children with global malnutrition.

Predicted plasma concentrations of quinine over a 72-h period were computed(Siphar/PC, version 4.0, Simed, Créteil, France) using a one-compartment model(2). In controls, predicted concentrations of quinine ranged between 3.9 and 7.9 mg/L on the 3rd d of treatment after receiving a 16 mg/kg intramuscular quinine-base loading dose followed by 8 mg/kg every 12 h. In globally malnourished children, predicted concentrations ranged between 1.8 and 5.5 mg/L using the same regimen schedule. However, these concentrations are probably not effective againstP. falciparum(30). Therefore as hydroxyquinine has less antimalarial activity than the parent drug(44), the quinine administration interval should be reduced in globally malnourished children, especially in areas where P. falciparum sensitivity to quinine is low. We suggest the reduction of the dose interval in malnourished children rather than increase the dose with unchanged dose interval because the two groups differed for clearance but not volume of distribution. Using our data and the same one-compartment model, an interval of 8 h would appear to be safe. However, these theoretical calculations need to be validated in a larger number of patients.