Pharmacokinetics of Antimalarials in Pregnancy
- First Online:
- Cite this article as:
- Wilby, K.J. & Ensom, M.H.H. Clin Pharmacokinet (2011) 50: 705. doi:10.2165/11594550-000000000-00000
- 177 Views
Malaria is a serious parasitic infection, which affects millions of people worldwide. As pregnancy has been shown to alter the pharmacokinetics of many medications, the efficacy and safety of antimalarial drug regimens may be compromised in pregnant women. The objective of this review is to systematically review published literature on the pharmacokinetics of antimalarial agents in pregnant women. A search of MEDLINE (1948-May 2011), EMBASE (1980-May 2011), International Pharmaceutical Abstracts (1970-May 2011), Google and Google Scholar was conducted for articles describing the pharmacokinetics of antimalarials in pregnancy (and supplemented by a bibliographic review of all relevant articles); all identified studies were summarized and evaluated according to the level of evidence, based on the classification system developed by the US Preventive Services Task Force. Identified articles were included in the review if the study had at least one group that reported at least one pharmacokinetic parameter of interest in pregnant women. Articles were excluded from the review if no pharmacokinetic information was reported or if both pregnant and non-pregnant women were analysed within the same group. For quinine and its metabolites, there were three articles (one level II-1 and two level III); for artemisinin compounds, two articles (both level III); for lumefantrine, two articles (both level III); for atovaquone, two articles (both level III); for proguanil, three articles (one level II-1 and two level III); for sulfadoxine, three articles (all level II-1); for pyrimethamine, three articles (all level II-1); for chloroquine and its metabolite, four articles (three level II-1 and one level II-3); for mefloquine, two articles (one level II-1 and one level III); and for azithromycin, two articles (one level II-1 and one level III). Although comparative trials were identified, most of these studies were descriptive and classified as level III evidence. The main findings showed that pharmacokinetic parameters are commonly altered in pregnancy for the majority of recommended agents. Importantly, first-line regimens of artemisinin-based compounds, lumefantrine, chloroquine and pyrimethamine/sulfadoxine may undergo significant changes that could decrease therapeutic efficacy. These changes are usually due to increases in the apparent oral clearance and volume of distribution that commonly occur in pregnant women, and may result in decreased exposure and increased therapeutic failure. In order to assess the clinical implications of these changes and to provide safe and effective dosage regimens, there is an immediate need for dose-optimization studies of all recommended first- and second-line agents used in pregnant women with malaria.
Malaria is currently one of the most important parasitic infections. In 2009, the estimated number of cases worldwide was 225 million, with 781 000 deaths. Although the estimated number of infections decreased from 244 million in 2005, malaria remains one of greatest infectious disease burdens in countries worldwide. While malaria affects many different populations, pregnant women are at especially high risk, as malaria in pregnancy is associated with low birth weight, increased anaemia, severe malaria and death.
In addition to determining the safety and efficacy of antimalarials in pregnancy, it is essential to characterize any pharmacokinetic changes that occur. As pregnancy is associated with changes in drug disposition (including, for example, increases in the apparent clearance after oral administration [CL/F] or apparent volume of distribution after oral administration [Vd/F]), pregnant women may need dosage adjustments and regimen alterations in order to ensure safe and effective treatment. It is therefore important to recognize the changes in pharmacokinetic parameters that occur in pregnancy, in order to maximize efficacy and minimize toxicity in this particularly vulnerable population. Recent reviews on this topic have concluded that there is limited research available to make dosage adjustments, and the lack of pharmacokinetic, efficacy and safety data are major barriers to providing adequate treatment.[5,6]
Interestingly, there has been a recent increase in the amount of literature published to help answer these questions and to begin to characterize the pharmacokinetic changes that occur in pregnant women. As study design and quality can influence study findings, it is important for this literature to be evaluated systematically. The US Preventative Services Task Force classification system recognizes that study design is an important measure to assess the validity of study results. It involves assessment and evaluation of study design, based on a level-of-evidence hierarchy, thereby allowing clinical decision makers to make informed decisions based on the highest level of evidence available. The objective of this review is to summarize and systematically evaluate the published literature reporting the pharmacokinetic parameters of the current medications used to treat malaria in pregnant women.
1. Search Strategy
A search of MEDLINE (1948–May 2011), EMBASE (1980–May 2011), International Pharmaceutical Abstracts (1970–May 2011), Google and Google Scholar was conducted for articles describing the pharmacokinetics of antimalarial drugs in pregnancy. The following search terms were combined: ‘antimalarial OR anti-malarial OR malaria OR falciparum OR atovaquone OR proguanil OR malarone OR artemether OR artemisinin OR artesunate OR dihydroartemisinin OR lumefantrine OR quinine OR sulfadoxine OR pyrimethamine OR clindamycin OR azithromycin OR antibiotics OR mefloquine OR chloroquine’ with ‘pregnancy’, ‘pharmacokinetics’, ‘drug metabolism’ and ‘human’. The reference lists of identified articles were manually searched for pertinent articles not identified in the electronic search. Identified articles were included if the study had at least one group that reported at least one pharmacokinetic parameter of interest in pregnant women. Articles were excluded if no pharmacokinetic information was reported or if both pregnant and non-pregnant women were analysed within the same group.
Each study was ranked on the basis of the quality of evidence it provided, as per the classification system developed by the US Preventive Services Task Force (1996 version). Level I studies are randomized controlled trials. Level II-1 studies are controlled (with patients acting as their own controls or with a parallel control group included). Level II-2 studies are defined as cohort or case-control studies. Level II-3 studies are multiple time series studies or exceptional descriptive studies. Level III studies are descriptive studies and case reports.
The information extracted from each study included the study design, number of participants, baseline parasitaemia status, treatment regimen (formulation, dosage and duration), blood sampling times, pharmacokinetic model, and pharmacokinetic parameters (the total body clearance [CL], CL/F, area under the concentration-time curve [AUC], volume of distribution [Vd], Vd/F, elimination half-life [t½], maximum drug concentration [Cmax] and time to reach the Cmax [tmax]). The results of the search are described below and are listed with respect to drug class.
1.1 Results of Literature Search
Our search produced 21 articles,[8–28] of which 20 were included in the review. One article was excluded because it included patients throughout the peri-partum period. Six of the studies reported the results of two medications of interest, as these medications are commonly administered as fixed-dose combinations to treat malaria.[12,15–19] For quinine and its metabolites, there were three articles[9–11] (one level II-1 and two level III); for artemisinin compounds, two articles[12,13] (both level III); for lumefantrine, two articles[12,14] (both level III); for atovaquone, two articles[15,16] (both level III); for proguanil, three articles[15–17] (one level II-1 and two level III); for sulfadoxine, three articles[18–20] (all level II-1); for pyrimethamine, three articles[18–20] (all level II-1); for chloroquine and its metabolite, four articles[21–24] (three level II-1 and one level II-3); for mefloquine, two articles[25,26] (one level II-1 and one level III); and for azithromycin, two articles[27,28] (one level II-1 and one level III). Six articles reported results from healthy volunteers,[17,20,23,24,26,28] four from patients with or without parasitaemia but no active infection[18,19,21,27] and ten from patients with acute infection.[9–16,22,25]
2.1 Level II-1
2.2 Level III
The pharmacokinetic parameters of quinine and its metabolites were assessed in nine pregnant and eight non-pregnant Sudanese women presenting with P. falciparum infection, but this well designed descriptive study did not include comparisons between the two groups. All patients received a single dose of quinine (10 mg/kg) as a 2-hour intravenous infusion (phase 1). Blood samples for plasma were obtained before quinine administration and at 0.5, 1, 2, 4, 12, 24, 48 and 72 hours after administration. The protocol was repeated 1 week later to assess changes dependent on infection status (phase 2). Samples were analysed for quinine, 3-OHQ, (10R)-DOHQ and (10S)-DOHQ, and the pharmacokinetic parameters were determined using non-compartmental methods. The parameters from phase 1 (active infection) are described in table II. All patients (pregnant or non-pregnant) cleared the infection by the fourth day following treatment initiation, and none developed parasitaemia or fever during the 28-day follow-up period. When the effect of infection status was assessed, no statistical differences were found in pharmacokinetic parameters between phases 1 and 2 in pregnant patients.
In another descriptive study, quinine was administered as an intravenous infusion over 4 hours in ten pregnant patients with P. falciparum infection. Eight patients received an initial dose of 10 mg/kg and two patients received an initial dose of 20 mg/kg. All patients then received 10 mg/kg doses every 8 hours by 4-hour intravenous infusion or oral tablets for 7 days. Blood samples for plasma were obtained frequently throughout the 4-hour quinine infusion and for 6 hours afterwards. The pharmacokinetic parameters (mean ± standard deviation), calculated after administration of this loading dose, were CL 1.22 ± 0.77 mL/kg/min, Vd 0.96 ± 0.27 L/kg and t½ 11.3 ± 4.3 hours.
3. Artemisinin Compounds
3.1 Level III
One study described the pharmacokinetics of artemether/lumefantrine in pregnant patients. This study was part of a larger randomized controlled trial, in which pregnant women received either 7 days of artesunate or 3 days of the fixed-dose combination artemether/lumefantrine. The pharmacokinetics of artemether and its active metabolite dihydroartemisinin (DHA) were assessed. The results for lumefantrine are reported in section 4.1.
Subjects in the artemether/lumefantrine arm were given four tablets of artemether/lumefantrine 20 mg/120 mg twice daily (total dose 80 mg/480 mg twice daily) for 3 days. Doses were given with 250 mL of chocolate milk containing 7 g of fat, as lumefantrine absorption increases 16-fold when administered with a meal rich in fat. Thirteen women (median gestational age 23 weeks) completed the study and had drug concentrations drawn at baseline and at 0.5, 1, 2, 4, 6, 8, 12, 24, 48, 72, 96, 120, 144 and 168 hours after the last dose was given. Both artemether and DHA concentrations were measured. The pharmacokinetic parameters are reported in table II. No serious adverse effects were reported.
A second study described the pharmacokinetics of DHA in pregnant patients. Twenty-four pregnant patients (median gestational age 28.5 weeks) were administered 4 mg/kg/d of oral artesunate, 20 mg/kg/d of oral atovaquone and 8 mg/kg/d of oral proguanil for 3 days. Doses were administered with 200 mL of chocolate milk to increase the absorption of atovaquone. Blood samples for plasma were taken prior to the third dose and at 0.5, 1, 2, 3, 6, 8, 12, 24, 48, 72, 96, 120, 144 and 168 hours after the third dose. The pharmacokinetic parameters of DHA were determined using non-compartmental analysis and are listed in table II. Because of the rapid hydrolysis of artesunate to DHA, artesunate parameters could not be determined. No serious adverse effects were reported.
4.1 Level III
Two studies reported the pharmacokinetic parameters of lumefantrine. One study was embedded within the randomized controlled trial described in section 3.1. The pharmacokinetic parameters of lumefantrine were estimated using non-compartmental analysis. As the objective of the study was to characterize the elimination phase of lumefantrine, blood samples for plasma were collected starting at 60 hours after the first dose was given and continuing until 168 hours after the last dose. Pharmacokinetic parameters are reported in table II. A unique finding of this study was that smoking significantly affected the concentration-time profile of lumefantrine, with the AUC from 60 hours until infinity (AUC60h–∞) being decreased by 30% in smokers compared with non-smokers.
The second study was a nested population pharmacokinetic study, using data from the randomized controlled trial described in section 3.1. In this analysis, 103 Burmese and Karen women from Thailand in the second and third trimesters of pregnancy were treated with oral artemether/lumefantrine 80 mg/480 mg twice daily for 3 days. A population approach was used to estimate the pharmacokinetic parameters. Plasma concentration-time data were best described by an open two-compartment disposition model with first-order absorption and lag time. The plasma lumefantrine concentrations showed large interindividual variability in all pharmacokinetic parameters. A linear covariate relationship was included to reduce the interindividual variability in the apparent central volume of distribution after oral administration (V1/F). The pharmacokinetic parameters estimated from this model are shown in table II. Efficacy endpoints were reported, and 17% of patients experienced recrudescence by day 42.
In order to address decreased efficacy, dosage regimen simulations using the final two-compartment model were performed. The results from this analysis showed increased exposure to artemether and DHA and increased day 7 concentrations of lumefantrine when the same total dose was given over 5 days, as compared with the traditional 3-day regimen.
5. Atovaquone and Proguanil
5.1 Level II-1
One controlled study assessed the pharmacokinetics of proguanil in ten pregnant women of Karen ethnicity living on the Thai-Burmese border. The first phase of this two-phase study occurred during the last trimester of pregnancy; patients were then asked to return to the clinic 2 months after delivery to complete the second (control) phase. For the pregnancy phase, all subjects had negative blood smears for malaria. The women were administered a single oral dose of proguanil 200 mg, and blood samples for plasma were taken at 0, 0.5, 1, 1.5, 2, 3, 4, 5, 6, 8, 10, 24, 32 and 48 hours after dosing. Pharmacokinetic parameters were obtained using the method of residuals, utilizing either one- or two-compartment models. The parameters of proguanil and its active metabolite cycloguanil embonate are reported in table II. The t½ of proguanil was significantly shorter in pregnant patients, and the Cmax and AUC values of cycloguanil embonate appeared to be higher in non-pregnant patients.
5.2 Level III
Two prospective uncontrolled studies described the pharmacokinetic parameters of atovaquone and proguanil in pregnant patients.[15,16] One study enrolled 24 Karen women infected with an uncomplicated P. falciparum infection or mixed infection (P. falciparum and Plasmodium vivax) that had failed to resolve after 7 days of quinine therapy. The women received combination therapy for 3 days of oral atovaquone 20 mg/kg/d plus oral proguanil 8 mg/kg/d plus oral artesunate 4 mg/kg/d. All doses were administered with 200 mL of chocolate milk (8% fat). Blood samples for plasma were obtained at baseline, prior to the third dose and then following the third dose at times of 0.5, 1, 2, 4, 6, 8, 12, 24, 48, 72, 96, 120, 144 and 168 hours. The pharmacokinetic parameters of atovaquone and proguanil were determined using non-compartmental analysis. The pharmacokinetic characteristics were also examined using a compartmental analysis, and the different models were assessed for the best fit.
The enrolled women had a median age of 20 years (range 16–37 years) and a median estimated gestational age of 28.5 weeks (19.1–35.9 weeks). The pharmacokinetic parameters of atovaquone, proguanil and cycloguanil embonate are listed in table II. Atovaquone was best fitted to a one-compartment model with the CL/F and Vd/F fitted as random effects and the absorption rate constant (ka) as a fixed effect. The CL/F and Vd/F both increased significantly with higher initial parasitaemias, and the Vd/F also increased significantly with increases in the gestational age. Proguanil was best fitted to a one-compartment model with the ka, CL/F and Vd/F all fitted as random effects. The Vd/F increased significantly with increasing gestational age. Of the women who completed follow-up at day 28 (75%), 100% had obtained a clinical cure. No serious adverse events were noted.
A second descriptive study enrolled pregnant women from both Thailand and Zambia. Subjects were in the third trimester of pregnancy (gestational age 24–34 weeks) and were experiencing an acute, uncomplicated P. falciparum infection. Subjects received four oral tablets of the fixed-dose combination atovaquone/proguanil 250 mg/100 mg once daily (total daily dose of 1000 mg/400 mg) for 3 days with 250 mL of water and a meal with 20–25% fat content. Blood samples for plasma were taken immediately prior to the third dose and at 2, 3, 4, 6, 8, 12, 24, 48, 72 and 120 hours afterwards. Additional samples were taken on days 10 and 14 (after the last dose) for determination of atovaquone concentrations only. Pharmacokinetic parameters were analysed using compartmental methods.
The pharmacokinetic parameters determined for atovaquone and proguanil are reported in table II. There were no significant differences in any of the pharmacokinetic parameters of the compounds between the patients from Thailand and Zambia. An efficacy assessment showed that the 28-day cure rates were 100%. The most commonly reported adverse event was abdominal pain.
6.1 Level II-1
Three controlled studies compared the pharmacokinetics of sulfadoxine and pyrimethamine in pregnant and non-pregnant patients. The first study was conducted in Papua New Guinea and enrolled 30 pregnant women (second and third trimesters) and 30 non-pregnant women. The women were eligible regardless of baseline parasitaemia status but were excluded if they presented with severe malaria. All patients received a single oral dose of the fixed-dose combination of pyrimethamine/sulfadoxine 75 mg/1500 mg and a 3-day course of chloroquine. All drugs were administered as full doses under direct supervision. Blood samples for plasma were obtained at baseline, at 1, 2, 4, 6, 12, 18, 24, 30, 48 and 72 hours, and at 7, 10, 14, 28 and 42 days post-dose. Concentration-time profiles were analysed by nonlinear mixed-effects modelling to develop a population-based model. The sulfadoxine model was based on the pharmacokinetic parameters of both sulfadoxine and its major metabolite N-acetylsulfadoxine.
No differences were found between pregnant and non-pregnant patients at baseline, and no adverse events were reported. For sulfadoxine, a one-compartment model was developed with first-order absorption, first-order conversion of the parent drug to the metabolite and first-order elimination of both the parent drug and the metabolite. Interestingly, only the effect of pregnancy on the CL/F of sulfadoxine significantly improved the model. For pyrimethamine, a final two-compartment model was determined, and the effect of pregnancy on the CL/F, V1/F and apparent peripheral volume of distribution after oral administration (V2/F) produced significant improvement in the base model.
The post hoc Bayesian-predicted pharmacokinetic parameters are listed in table II for both sulfadoxine and pyrimethamine. Compared with non-pregnant patients, pregnant patients had a significantly smaller AUC, faster CL/F and shorter t½ of sulfadoxine; and a smaller AUC, faster CL/F, larger V1/F, V2/F and apparent volume of distribution at steady state after oral administration (Vss/F), and longer t½ of pyrimethamine. Parameters were also reported for N-acetylsulfadoxine. When compared with non-pregnant patients, pregnant patients had a faster CL/F (0.567 vs 0.338 L/h), similar Vd/F (3.3 vs 3.1 L) and shorter t½ (4.1 vs 6.5 hours). A 38% failure rate at day 28 was reported for patients with P. falciparum infection.
The second study enrolled 33 pregnant women from Kisumu, Kenya, with or without P. falciparum parasitaemia but without clinical infection. Women had an uncomplicated singleton pregnancy of 16–28 weeks gestation and a haemoglobin level >8 g/dL. Subjects acted as their own non-pregnant controls and returned to the clinic to repeat the study during the first 3 months of the post-partum period.
Participants were administered a single dose of the fixed-dose combination of pyrimethamine/sulfadoxine 75 mg/1500 mg (three 25 mg/500 mg tablets). Blood samples were obtained at baseline and at 6, 24, 48, 72, 96, 120 and 240 hours after administration. All concentrations were measured using a high-performance liquid chromatography assay, and parameters were determined from whole-blood samples. Pharmacokinetic parameters were determined using a one-compartment model for both sulfadoxine and pyrimethamine. The same protocol was repeated during the post-partum period (median 10 weeks post-partum), with 11 patients returning to complete this part of the analysis.
The pharmacokinetic parameters that were obtained are presented in table II. For sulfadoxine, pregnant patients had a decreased AUC, increased CL/F and shorter t½, as compared with the post-partum period. No parameters were significantly different between periods for pyrimethamine.
A third study enrolled 98 pregnant women from Mali, Mozambique, Sudan and Zambia. Subjects were 18–45 years old with a gestational age of 15–36 weeks and were non-parasitaemic at baseline. Subjects from Mali and Zambia were encouraged to return to the clinic 6–8 weeks after delivery to repeat the study and act as the non-pregnant controls.
Drugs were administered as a single oral dose of three fixed-dose combination tablets (total doses of sulfadoxine 1500 mg and pyrimethamine 75 mg). Blood samples were obtained immediately prior to dosing, at 3, 6 and 12 hours, and at 1, 3, 7, 14, 21 and 28 days after dosing at the Mali and Zambia sites, and immediately before dosing and at 1, 2, 3, 7, 14, 21 and 28 days after dosing at the Mozambique and Sudan sites. The same procedures were followed for those patients who returned to the clinics during the post-partum period.
The pharmacokinetic parameters of sulfadoxine and pyrimethamine, both calculated using a one-compartment model across all study sites, are listed in table II. When compared with post-partum data, pregnant patients had a higher Cmax, smaller AUC, faster CL/F, smaller Vd/F and shorter t½ of sulfadoxine. For pyrimethamine, pregnant patients had a higher Cmax, larger AUC, slower CL/F and longer t½.
7.1 Level II-1
Three controlled studies compared the pharmacokinetics of chloroquine. In one study, 30 pregnant and 30 non-pregnant patients from Papua New Guinea with or without parasitaemia were given 1350 mg (three oral tablets of 450 mg) of chloroquine once daily as the base regimen for three days, with a single dose of oral pyrimethamine/sulfadoxine 75 mg/1500 mg on the first day. Blood samples for plasma were taken at baseline, at 1, 2, 4, 6, 12, 18, 24, 30, 48 and 72 hours post-treatment, and at 7, 10, 14, 28 and 42 days post-treatment. A nonlinear mixed-effects modelling approach, allowing for the assessment of covariates, was used to analyse the concentration-time profiles from plasma samples.
The data that were obtained were fitted to a two-compartment model, and pharmacokinetic parameters were determined for both chloroquine and its metabolite desethylchloroquine (DECQ). In the covariate analysis, the effect of pregnancy on the CL/F of both chloroquine and DECQ, as well as the effect of haemoglobin on the V1/F of chloroquine, were found to be significant. The pharmacokinetic parameters of chloroquine and DECQ are listed in table II. For chloroquine, pregnant patients had a faster CL/F, larger Vss/F, smaller AUC and shorter t½ than non-pregnant patients. Significant differences were also found with DECQ, including a faster CL/F, smaller AUC and shorter t½.
A 65% response rate was obtained for those who presented with P. falciparum at baseline. The subjects presenting with P. vivax and Plasmodium malariae did not redevelop parasitaemia. An interesting finding of the simulation analysis of this study was that a chloroquine regimen of four tablets once daily for 3 days in pregnancy would raise the AUC of chloroquine to that of non-pregnant patients; however, a five-tablet regimen would be needed to raise the AUC of DECQ to non-pregnant levels.
Another study enrolled 12 pregnant and 13 non-pregnant Karen women from Thailand with acute P. vivax infection. An oral dose of 25 mg/kg of chloroquine base was given over 3 days as 10 mg/kg, 10 mg/kg and 5 mg/kg at 0, 24 and 48 hours, respectively. Blood samples were taken at baseline, at 0.25, 0.5, 1, 1.5, 2, 3, 4, 6, 8 and 24 hours, and at 2, 3, 7, 14, 21, 28, 35 and 42 days. Non-compartmental analysis was used to estimate the pharmacokinetic parameters of chloroquine and DECQ from whole-blood samples, which are reported in table II. No differences were found for any of the pharmacokinetic parameters that were assessed (the Cmax, tmax, AUC, CL/F and t½) for chloroquine or DECQ. No efficacy data were reported. Interestingly, there was a significant reduction in the chloroquine AUC with increasing gestational age; however, this decrease did not become significant when compared with values in non-pregnant patients.
The third study enrolled five pregnant healthy volunteers in the early third trimester and five non-pregnant healthy volunteers. Patients received a single oral dose of 600 mg of chloroquine base, and blood samples for plasma were obtained at baseline and at 1, 2, 4, 6, 8, 24 and 48 hours after dosing. The pharmacokinetic parameters of chloroquine and DECQ, determined from plasma samples, are listed in table II. There were no differences in the tmax, Cmax and AUC of chloroquine between pregnant and non-pregnant patients. However DECQ demonstrated increased Cmax and AUC values in pregnant patients.
7.2 Level II-3
Twenty-five pregnant women from Tanzania were given 310 mg of chloroquine base orally once weekly as prophylaxis throughout pregnancy, as well as a daily dose of 200 mg ferrous sulphate and 5 mg folic acid. Subjects were healthy volunteers and were excluded if they presented with positive malaria smears at enrolment. Samples were taken in weeks 26 (the second trimester) and 36 (the third trimester). Blood samples were taken prior to the dose and at 1, 2, 3, 4 and 7 days afterwards. Whole-blood samples were used for the analysis. The mean CL/F of chloroquine increased significantly from the second to the third trimesters, resulting in CL/F values of 160 mL/min and 180 mL/min, respectively (p = 0.01). However, the AUC of DECQ did not change significantly from the second to the third trimester (80 361 nmol · h/L and 77 824 nmol · h/L, respectively).
8.1 Level II-1
One controlled study enrolled nine pregnant and eight non-pregnant women from Thailand. Seven of the women were in the last trimester, and two were in the first trimester. All patients were infected with P. falciparum. Oral mefloquine was given at a dose of 15 mg/kg to the nearest 125 mg (i.e. halved 250 mg tablets) as a single oral dose. Blood samples were obtained at 0.5, 1, 2, 3, 4, 5, 6, 8, 10, 12, 24 and 48 hours, then daily until discharge and at weekly intervals thereafter (if possible) in pregnant patients. Non-pregnant patients were sampled at 0.5, 1, 2, 4, 6, 8, 10, 12, 16, 18, 20, 24, 28, 36 and 48 hours, and then at 3, 4, 7 and 14 days. The pharmacokinetic parameters are reported in table II. The differences observed between groups included a significantly lower Cmax and a larger Vd/F in pregnant patients.
8.2 Level III
A descriptive study enrolled 20 healthy pregnant Karen women from Thailand who were all near the beginning of the third trimester of pregnancy. Subjects were randomized to receive either 250 mg or 125 mg of oral mefloquine once weekly. During the seventh study week, blood samples were obtained prior to dosing, at 1, 2, 3, 4, 6 and 8 hours, and at 2, 3, 4, 6 and 8 days. The data were best fitted to either a one- or two-compartment pharmacokinetic model, individualized for each patient.
Eighteen patients were included in the final analysis. The pharmacokinetic parameters are reported in table II. In addition, separate AUC values were reported for each dosage and were 2251 ± 256 ng · d/mL in the 125 mg group and 3955 ± 1214 ng · d/mL in the 250 mg group.
9.1 Level II-1
Antibacterial drugs can be used to enhance the efficacy of antimalarial agents. Macrolide agents, primarily azithromycin, inhibit protein synthesis in the plasmodial apicoplast. Their antimalarial effect is slow and weak, as a result of acting against the progency of parasites after exposure. However, they may have synergistic or additive effects when combined with other antimalarials.
One study assessed azithromycin in pregnant and non-pregnant patients in Papua New Guinea. Thirty-one pregnant and 29 non-pregnant patients were enrolled and received oral azithromycin 2 g at enrolment and again 24 hours later. Patients were also randomized to receive either oral pyrimethamine/sulfadoxine 75 mg/1500 mg as a single dose or oral chloroquine 450 mg once daily as a base regimen for 3 days. Only seven patients in the pregnant group and two patients in the non-pregnant group were infected with P. falciparum (seven patients) or P. vivax (two patients) at enrolment. The median gestational age was 21–24 weeks in the pregnant patients. Blood samples for plasma were taken at 1, 2, 3, 6, 12, 24, 32, 40, 48 and 72 hours, and at 4, 5, 7, 10 and 14 days after the initial azithromycin dose. Pharmacokinetic parameters were determined by nonlinear mixed-effects modelling, and the influence of covariates was also assessed.
A three-compartment model was best to fit the data, and pregnancy on V1/F was the only covariate that significantly decreased the objective function value. The pregnant patients had a significantly higher Vd/F and a longer absorption half-life (t½abs). These differences, however, did not result in significant differences in the AUC from time zero to infinity (AUC∞) [28 713 vs 31 781 µg · h/L] or the t½ [78.2 vs 77.1 hours] in pregnant versus non-pregnant patients. Therapeutic efficacy was not assessed in this study.
9.2 Level III
In one study, a 1g oral suspension dose of azithromycin was administered to 20 healthy pregnant women at intervals of 6, 12, 24, 72 or 168 hours pre-Caesarean section. Blood samples for plasma were obtained, and the t½ was estimated to be 12 hours in both the maternal and fetal serum. The t½ in maternal myometrium and adipose tissue was reported to be >70 hours. No other pharmacokinetic parameters were reported.
Our search identified the available literature describing the pharmacokinetic changes that occur with antimalarial drugs in pregnancy. The limited number of trials that we identified is probably due to frequent exclusion of pregnant women from clinical trials, ethical considerations pertaining to fetal exposure to medications, and lack of funding for malaria research in endemic populations. The available evidence varies widely in quality (as evaluated by the level-of-evidence classification) and study design. However, we identified many studies that allowed for conclusions to be drawn regarding pharmacokinetic changes during pregnancy and associated clinical implications.
Quinine is one of the first-line agents recommended for use during the first trimester of pregnancy. According to the best available evidence (level II-1), the pharmacokinetics of quinine and its metabolites do not appear to be significantly altered in pregnancy. A recent review of parasitological efficacy reported conflicting results, with failure rates of 0–40% being reported in studies assessing efficacy in pregnant patients. However, it is possible that the high failure rates may have been due to non-adherence, and thus the true rate of therapeutic failures is unknown. For malaria treated in a hospital setting with a standard quinine regimen, pharmacokinetic changes will probably not influence its efficacy. However, this should be evaluated in combined pharmacokinetic-efficacy trials if quinine is to remain a first-line agent for use during pregnancy.
Artemisinin-based compounds have been studied in level III studies. When compared with studies of non-pregnant patients from similar populations, it appears that the Cmax and AUC values of DHA are lower in pregnant women.[12,13,33] Lower concentrations achieved in pregnant women may lead to decreased clinical efficacy and increased morbidity and mortality. Higher failure rates have been shown to occur in pregnant women with both artemether and artesunate. Although the exact mechanism is unknown, these discrepancies may be due to pharmacokinetic changes that occur in pregnancy, such as increases in the CL/F and Vd/F. It is essential that dose-optimization studies of artemisinin-based compounds are performed in pregnant women, in order to maximize the clinical efficacy and tolerability of these regimens.
Although there were no controlled trials identified for lumefantrine, it is speculated that differences exist in its pharmacokinetic parameters in pregnant patients because of increased failure rates and comparisons with historical controls. The significantly increased V1/F with each 1-week increase in the estimated gestational age is interesting and confirms that changes occur in pregnant women with respect to lumefantrine disposition. In addition, these changes could account for the lower day 7 concentrations reported in the studies that we assessed, and could thus be important factors in clinical failure. According to a unique simulation model developed in pregnant patients, it is possible that a same-total-dose 5-day regimen of artemether/lumefantrine may offer adequate exposure to these medications (on the basis of population estimates) by providing artemisinin exposures for three parasite cycles and boosting the day 7 concentrations of lumefantrine.
Unfortunately, there is limited information on the variable absorption of lumefantrine, a characteristic that complicates its pharmacokinetic profile; this has important implications for pregnant women. For example, the patients in the studies described here received doses with a high-fat beverage in order to maximize absorption of the lipophilic molecule.[12,14] When malaria is treated routinely in endemic regions, it is unknown how often lumefantrine is administered with a high-fat meal. It is therefore worrisome that any pharmacokinetic changes occurring in pregnancy may result in a further increase in therapeutic failure if patients are unable to maximize the bioavailability of lumefantrine. Interestingly, smoking status was also found to be a significant covariate, according to a population model. These results confirm the highly variable pharmacokinetic profile of lumefantrine, thereby supporting the case for dose-optimization trials in pregnant women.
The changes in the pharmacokinetics of chloroquine and mefloquine that occur in pregnancy are not well characterized, but the available evidence suggests that significant changes may occur. For chloroquine, level II-1 evidence suggests that the CL/F and Vd/F may be increased, while the AUC may be decreased in pregnant patients. Decreases of 25% and 45% in the AUCs of chloroquine and DECQ, respectively, suggest lower exposure that may compromise therapeutic efficacy. However, a study reporting parasitaemia clearance showed 100% success at 96 hours in all patients treated with 10 mg/kg loading doses at 0 and 24 hours and then 5 mg/kg dose at 48 hours. In order to achieve target concentrations and increase the likelihood of therapeutic efficacy, a population model suggested that a 4- or 5-tablet daily regimen of chloroquine may increase the AUCs of chloroquine and DECQ to values similar to those seen in non-pregnant patients. The authors of this study recommended that a 4-tablet regimen be used routinely in pregnant patients. While this recommendation may theoretically increase the exposure of chloroquine and DECQ, clinical studies evaluating safety and potential harm to the fetus are needed. Although the available evidence suggests that plasma concentrations of mefloquine may be altered in pregnancy (most likely because of the increased Vd/F), this drug has fallen out of favour because of an increased risk of stillbirths observed in Thai patients.
The current evidence of pregnancy-related alterations in the pharmacokinetics of atovaquone and proguanil is weak and mostly comes from two level III descriptive studies.[15,16] Although these trials were not designed as efficacy trials, both reported cure rates of 100%. Therefore, it is uncertain if dosage adjustments are needed in pregnant women, and this should be further evaluated in larger efficacy trials. To our knowledge, atovaquone and proguanil are not commonly used as treatment for malaria in endemic areas; therefore, we believe that research should be focused on the artemisinin-based compounds, lumefantrine and pyrimethamine/sulfadoxine.
Cycloguanil embonate pharmacokinetic parameters were reported in two studies and were shown to be decreased in pregnant patients when compared with non-pregnant patients. A study designed to assess the effect of pregnancy on conversion of proguanil to cycloguanil embonate found that late pregnancy was associated with decreased conversion to the active metabolite. The authors of this study speculated that this might be due to estrogen inhibition of the cytochrome P450 2C19 metabolic pathway and recommended a 50% increase in the proguanil dosage in pregnant patients. The clinical implications of this regimen need to be assessed before this recommendation is incorporated into clinical practice.
Level II-1 evidence suggests that the pharmacokinetics of sulfadoxine and pyrimethamine, when administered as a fixed-dose combination, are altered in pregnant women.[18–20] Pregnant patients had increased CL/F, decreased AUC and decreased t½ values. The pyrimethamine results were conflicting between studies, but it appears that the CL/F and Vd/F may be increased in pregnancy. One study reported a 38% failure rate with the standard pyrimethamine/sulfadoxine regimen, and high failure rates in pregnant women have also been reported in a recent review. As therapeutic efficacy may be compromised, adequate dose-optimization studies are urgently needed to assess the most optimal dose of pyrimethamine/sulfadoxine in pregnant patients.
Pyrimethamine/sulfadoxine is also used as intermittent preventative therapy for malaria during pregnancy. Only one of the reviewed studies assessed this indication, and it found inconsistent changes between the different geographical populations that were assessed. A recent review has questioned the appropriateness of administering pyrimethamine/sulfadoxine for this indication and called for more research to establish effective dosing strategies. Clinically, it is essential that influence of pharmacokinetics on the efficacy of these medications should be characterized in order to justify exposing pregnant patients to them.
Antibacterials can be used to enhance the efficacy of antimalarial drugs. The best evidence (level II-1) suggests that azithromycin parameters are significantly affected in pregnancy (increased Vd/F and t½abs values) but translation into clinical outcomes is uncertain, as efficacy was not assessed.
The results of our review correlate with a recent review that systematically evaluated the parasitological efficacy of antimalarial regimens in pregnant women with P. falciparum-specific malaria; 60 articles were included in that review of efficacy, safety and pharmacokinetics. Overall, the authors found that 22 of 53 treatment arms (41.5%) achieved a World Health Organization failure rate of <5%. The highest number of failures was seen with the pyrimethamine/sulfadoxine combination, with 58.3% of treatment arms having a failure rate of ≥10%. High failure rates were also seen with chloroquine, quinine and artemisinin-based regimens when given as monotherapy. It is speculated that pharmacokinetic changes occurring in pregnancy could be a major contributor to parasitological failure of these compounds. We agree with this conclusion, considering the pharmacokinetic changes we have observed. Increases in the CL/F and Vd/F and decreases in the AUC may compromise therapeutic efficacy, and dose-optimization studies are urgently needed.
Combination regimens of multiple antimalarials have been suggested to provide acceptable failure rates of <5%. Although only limited evidence is available, it has been suggested that amodiaquine may be a potential candidate for combination regimens. However, studies are needed to assess the safety and efficacy of combinations of any antimalarial medications before recommendations can be made. In addition, there is insufficient information describing the role of dapsone in pregnancy, and thus this agent cannot be recommended at this time.
The major limitation of this systematic review was the difficulty in drawing conclusions on the basis of the studies that were identified. Most of the studies were descriptive and did not include control groups. Although it is possible to make observations when making comparisons with historical controls, variability in assays and populations limit the validity of this method. Moreover, many of the identified studies came from similar populations and may not be reflective of all populations living within malaria zones. Genetic and environmental influences may significantly alter the pharmacokinetics in different populations. Lastly, there are few trials reporting the efficacy of antimalarials in pregnant patients, and this hinders our ability to draw adequate conclusions based on the pharmacokinetic and pharmacodynamic effects.
The studies identified in this review suggest that the pharmacokinetic parameters of agents currently used to treat malaria in pregnant patients may be significantly affected by pregnancy, thereby leading to potential compromises in therapeutic efficacy. Future pharmacokinetic and dose-optimization studies should include therapeutic efficacy outcomes, in order to adequately assess the impact of pharmacokinetic changes on clinical outcomes. Urgent research is needed, especially for artemisinin compounds, lumefantrine, chloroquine and pyrimethamine/sulfadoxine, as these agents are the most commonly used and recommended for malaria treatment.
No sources of funding were used to prepare this review. The authors have no potential conflicts of interest to declare that are directly relevant to the content of this review.