4 Discussion
We successfully developed a pregnancy PBPK model to predict the
disposition of the approved monthly and bimonthly dosing regimens of LAI
CAB and LAI RPV in pregnancy without the oral lead-in components. An
earlier study had shown that Ctrough after the first LAI
CAB dose with or without an oral lead-in were comparable though the
target Ctrough could be achieved faster with the LAI CAB
dose if given with the oral lead-in component [53]. The pregnancy
PBPK model was developed from a validated adult PBPK model by
incorporating pregnancy-induced biological changes that are known to
influence PK such as changes in body weight, relevant enzyme activities
and cardiac output defined by gestational age [17]. The adult PBPK
model was validated with PK data of LAI CAB and LAI RPV in adults
[34]. Similarly, available clinical PK data of oral RPV in pregnancy
were used to validate the pregnancy PBPK model for RPV PK and by
extension, the CYP3A4 activity in pregnant women. The absence of
clinical PK data for oral CAB in pregnancy led to the adoption of a
probe substrate (RAL) to validate UGT1A1 activity in pregnancy. RAL was
a suitable probe substrate in this instance as it is solely metabolised
by UGT1A1 and clinical PK data of oral RAL in pregnancy are available
[52]. The use of a probe substrate to validate enzyme activity for a
different drug with inadequate data has been previously reported in
another study [54].
The model predictions suggest that both dosing regimens of LAI CAB were
predicted to maintain efficacy throughout pregnancy. In contrast, whilst
the monthly dosing regimen of RPV could maintain antiviral efficacy
throughout pregnancy for majority of the population, the model suggests
the need for caution in introducing the bimonthly regimen of RPV to the
pregnant population. Use of the bimonthly regimen therefore requires
careful clinical evaluation, including viral load monitoring and
potentially therapeutic drug monitoring.
In the PBPK model, a simple first-order equation was used to
characterise the absorption/release rate of the drug into the systemic
circulation from the IM depot in the muscle. The mathematical expression
was independent of the size of the patient’s muscle mass which could
explain why the predicted PK of LAI CAB did not vary significantly
between virtual patients with different body mass indices (BMI). Unlike
the predicted PK of LAI CAB, studies in humans have reported that BMI is
a significant covariate for the PK of LAI CAB [55]. The size of
muscle mass might affect the available depot space for the drug in the
muscle which could lead to a faster release of the drug into the
systematic circulation and contribute to a faster decline of the LAI CAB
concentrations in patients with low BMI [55, 56]. Patel et al (2020)
reported higher maximal levels of LAI CAB in the plasma of a study
volunteer with lower BMI compared to two others with higher BMI
[56]. However, the release rates of LAI CAB and LAI RPV used in this
study were fitted into the model with available clinical data [33,
34]. Sensitivity analyses of the plasma concentrations of cabotegravir
and rilpivirine to variations of their release rates are shown in Figure
S1. The PK of monthly and bimonthly LAI CAB and LAI RPV were also
simulated for non-pregnant adult females for comparison with the
pregnant population because LAI CAB PK has been reported to differ
between males and females [55].
Drug transporter activity was not incorporated into the PBPK model
primarily due to lack of data. RPV is not a known substrate of any drug
transporter. On the other hand, CAB is a substrate of Multidrug
resistance protein 1 (P-glycoprotein 1) and breast cancer resistance
protein (BCRP) in vitro [57]. Though pregnancy has been reported to
influence the activity of P-glycoprotein 1 and BCRP in rodents [58,
59], data in humans are not available. Regardless, the influence of
drug transporter activity on the PK of oral CAB appear to be minimal
[57].
A fetal compartment was not included in the female reproductive of the
pregnancy PBPK model. As such, fetal exposure to the LAI CAB and LAI RPV
in pregnancy could not be evaluated. UGT1A is not likely expressed in
fetal liver [60]. Though fetal liver has been reported to express
CYP3A, however, the contribution of the fetal liver clearance to the
overall drug clearance of LAI RPV in the mother are expected to be
minimal [61].
The administration of LAI drugs in pregnancy is not a new paradigm. LAI
antipsychotic drugs have been administered in pregnancy for over two
decades. Despite this long duration, PK data on the use of LAI
antipsychotics in pregnancy has been very limited [62]. Similarly,
outcomes on the safety of LAI antipsychotics in pregnancy have been
inconsistent. Where poor outcomes have been reported in pregnancy after
the use of LAI antipsychotics, there have been insufficient data to
determine if the poor outcomes are due to the illness, class of the drug
or the long-acting formulation [63]. Nonetheless, there have been
strong arguments for LAI antipsychotic use during pregnancy owing to
improved adherence, reduced risk of overdose, and less psychiatric
rehospitalisation compared to oral antipsychotics [64]. Adherence to
antipsychotics is particularly important during pregnancy to prevent
relapses which might lead to poor birth outcomes [64].
In a similar vein, adherence to antiretrovirals in pregnancy is highly
necessary to reduce the risk of vertical transmission of HIV. LAI
antiretrovirals might be a preferred choice throughout pregnancy to
support adherence and to reduce psycho-social challenges relating to
disclosure of HIV status. In addition, the new option of LAI
antiretrovirals might be particularly important in early pregnancy for
women living with HIV that may prefer a non-oral route of drug
administration due to nausea and vomiting. However, there is a need to
frequently monitor pregnant women on LAI antiretrovirals towards
improving available data on safety and efficacy. PBPK modelling readily
overcomes many ethical and logistic challenges associated with
randomised clinical trials in complex populations. It could also prove
useful in exploring PK in complex clinical scenarios and complex
populations.
Since the approval of LAI RPV and LAI CAB for the general adult
population, there have been limited clinical data to guide the dosing of
both LAIs in pregnant women. In this study, we developed a pregnancy
PBPK model to describe plasma concentrations of LAI CAB and LAI RPV in
pregnancy. Based on the model predictions, both the monthly and
bimonthly dosing regimen of LAI CAB could maintain antiviral efficacy
throughout pregnancy without need for adjustments. However, bimonthly
dosing regimen of LAI RPV might be introduced in pregnancy with caution
and adequate monitoring. Future clinical studies in humans are needed to
confirm these model predictions.
Acknowledgements
S.A. was supported by the Duncan Norman Charitable Trust. CW is funded
by the Wellcome Trust [222075_Z_20_Z]. For the purpose of open
access, the author has applied a CC BY public copyright licence to any
Author Accepted Manuscript version arising from this submission.
Conflict of interest statement
Marco Siccardi has received research grant funding from Janssen and ViiV
unrelated to this work. M.S. is currently employed by Labcorp. All other
authors have no potential conflicts of interest to declare.
Funding Information
Wellcome Trust (222075_Z_20_Z)
Data Availability statement
The data that support the findings of this study are available from the
corresponding author upon reasonable request.