Discussion
We developed a population pharmacokinetic model to describe the pharmacokinetics of esomeprazole in pregnant women with preeclampsia. To our knowledge, this is the first population pharmacokinetic model describing esomeprazole pharmacokinetics during pregnancy. We found that clearance was 42.2% (14.9% – 61.6%) lower in pregnant women with preterm preeclampsia after a single dose of esomeprazole compared to non-pregnant individuals. In non-pregnant individuals, clearance was 54.9% (48.2% – 63.5%) lower in extensive metabolizers and bioavailability was 33% (10.0% – 52.0%) higher after repeated dosing compared with single dose.
Increases in concentration of hormones such as oestrogen and progesterone during pregnancy could downregulate CYP2C19.6,21 The lower clearance in the pregnant patients is likely to be due to this downregulation of CYP2C19. Previous studies reported that CYP3A4 is upregulated during pregnancy,21,22 and since CYP3A4 is abundantly present in the gut one would expect a lower bioavailability. We did not find this effect in our model.
The lower clearance in extensive metabolizers in non-pregnant individuals with repeated esomeprazole dosing is likely to be due to auto-inhibition of CYP2C19 while the higher bioavailability could be because of a decreased first-pass effect associated with CYP2C19 auto-inhibition.2,5 Moreover, it has previously been reported for omeprazole that a decrease in intragastric acidity with repeated doses could lower its degradation in the stomach, improving its absorption.23 Hence, with repeated administration of esomeprazole, there could be improved bioavailability due to increase in pH and lower degradation in the stomach.
Esomeprazole disposition has mostly been described in literature with one-compartment kinetics for oral data and two-compartment kinetics for IV data.24,25 Two-compartment disposition showed better fit in our model, similar to a model by Standing et al .26 We estimated a typical clearance which was more than three times higher in extensive metabolizers compared to poor metabolizers This finding is consistent with previous reports that poor metabolizers have up to three times higher exposure than extensive metabolizers.3,4 Our estimate of clearance in poor metabolizers agrees with models by Nagase et al .24 and Standing et al .26The typical central apparent volume of distribution of 14.9 L in our study is similar to that found in healthy individuals (~16 L).24,27 High variability in speed of absorption was observed in our model, likely due to differences in gastric acidity between-individuals and occasions.24
Our study shows that esomeprazole clearance is lower during pregnancy which is probably due to CYP2C19 downregulation. This has several implications: first, metabolism during pregnancy may be more dependent on CYP3A4, which could mean there is less need for CYP2C19 genotyping during pregnancy, which has previously been suggested for proton pump inhibitors due to genotype-dependent variations in exposure and therapeutic/adverse outcomes.28 This could also mean less drug-drug interactions that involve CYP2C19. Second, the nonlinearity in esomeprazole pharmacokinetics with repeated dosing and with dose increases would be expected to be less in pregnancy. Specifically, the increase in exposure with repeated dosing, which is due to CYP2C19 autoinhibition, would not be expected to be as high as in non-pregnant. Esomeprazole exposure increases more than dose-dependently with dose increases above 20 mg, due to saturation of CYP2C19-based clearance and first pass effect.5,29 We don’t expect as high of a nonlinear increase in exposure during pregnancy since metabolism could be CYP3A4-dependent, which is generally considered linear.5,24,29
It was a limitation in our study that we couldn’t isolate the actual effect of pregnancy and repeated dosing because there were other factors that were different between the studies, such as mealtimes and formulation, which could affect the absorption of esomeprazole. However, we believe that with the available data, our model can adequately describe changes during pregnancy and repeated dosing. Only single dose data was available for pregnant patients and effect of repeated dosing during pregnancy could not be investigated. Nevertheless, esomeprazole metabolism seems to be less dependent on CYP2C19 during pregnancy, and we expect pregnant patients to have similar exposure after repeated doses as after single dose.
We used the first oral esomeprazole data from pregnant patients generated by the PIE trial to describe esomeprazole pharmacokinetics in pregnancy. PIE had found similar esomeprazole exposure in pregnant women to non-pregnant. In our model, we identified a lower clearance in pregnant women, but no significant change in bioavailability. Richer data from pregnant patients including healthy, non-pregnant controls and with CYP2C19 genotyping are needed to further investigate this. PIE reported no clinical benefit of esomeprazole for preeclampsia. Preclinical studies are needed to identify the pharmacodynamic target for preeclampsia and how esomeprazole acts on this target. Further clinical trials are also needed to investigate whether the preclinical efficacy can translate into human efficacy and our model can help to inform the design of these studies as well as to establish the pharmacokinetic metric that relates with pharmacodynamic markers for preeclampsia.