Abstract
Dexamethasone (DEX) is currently the treatment of choice for patients with oxygen‐dependent COVID‐19. It has been observed, primarily in vitro, that dexamethasone induces the expression of CYP3A and the ABCB1 gene, which encodes P‐glycoprotein (P‐gp). This has raised concerns about potential interactions between DEX and substrates of CYP3A and P‐gp, such as direct oral anticoagulants (DOAC). Currently, there is limited robust evidence to support a clinically significant interaction between DEX and DOAC. Using physiologically based pharmacokinetic modeling (PBPK), we investigated the impact of DEX administered in the context of SARS‐CoV‐2 infection on the pharmacokinetics of apixaban (APX) and rivaroxaban (RVX). After validating the induction effect of the DEX compound on two CYP3A4 substrates using the limited available studies, we optimized the compound in a COVID‐19 patient population, where significantly higher DEX plasma concentrations were observed compared to healthy volunteers. Our PBPK‐based PK simulations showed a 20% decrease in the AUC of APX and RVX in a worst‐case scenario and when DEX was administered at 6 mg PO for 10 days. This finding confirms the limited clinical data currently available and supports the use of APX and RVX with DEX in COVID‐19 patients at low‐risk for thrombo‐embolism. In addition, our results suggest that prednisone (PRED), when used at an equipotent dose, could serve as a viable alternative to DEX, given its less pronounced induction effect on APX and RVX. Further research is needed to validate these findings and to explore the clinical implications of using PRED in place of DEX in such scenarios.
Study Highlights.
WHAT IS THE CURRENT KNOWLEDGE ON THE TOPIC?
Dexamethasone (DEX) is currently the treatment of choice for patients with oxygen‐dependent COVID‐19. It has been observed, primarily in vitro, that dexamethasone induces the expression of CYP3A and the ABCB1 gene, which encodes P‐glycoprotein (P‐gp). This has raised concerns about potential interactions between DEX and substrates of CYP3A and P‐gp, such as direct oral anticoagulants (DOAC).
WHAT QUESTION DID THIS STUDY ADDRESS?
Using physiologically based pharmacokinetic modeling (PBPK), we investigated the impact of DEX administered in the context of SARS‐CoV‐2 infection on the pharmacokinetics of apixaban (APX) and rivaroxaban (RVX).
WHAT DOES THIS STUDY ADD TO OUR KNOWLEDGE?
Based on our PBPK analysis, which likely represents a worst‐case scenario, DEX was found to have a limited impact on the exposure of APX and RVX. Our results suggest that prednisone (PRED), when used at an equipotent dose, could serve as a viable alternative to DEX, given its less pronounced induction effect on APX and RVX.
HOW MIGHT THIS CHANGE CLINICAL PHARMACOLOGY OR TRANSLATIONAL SCIENCE?
As a result, the concomitant use of APX and RVX with DEX in COVID‐19 patients could be considered in specific clinical situations, particularly those involving low‐risk scenarios such as long‐term management of atrial fibrillation or venous thromboembolism.
Dexamethasone (DEX) is currently the treatment of choice for patients with oxygen‐dependent COVID‐19. 1 It has been observed, primarily in vitro, that DEX induces the expression of CYP3A and the ABCB1 gene, which encodes P‐glycoprotein (P‐gp). 2 This has raised concerns about potential interactions between DEX and substrates of CYP3A and P‐gp, such as direct oral anticoagulants (DOAC). 3 Since apixaban (APX) and rivaroxaban (RVX) are CYP3A substrates and all DOAC are P‐gp substrates, the use of P‐gp and CYP3A inducers with DOACs should be avoided, as this could compromise their efficacy in preventing ischemic events. 4
Some guidelines recommend against the combined use of DEX with DOACs, favoring the combination of DEX with heparins instead. 5 , 6 However, the parenteral administration of heparin can sometimes be problematic and may even delay the patients' discharge from the hospital. Consequently, we sought to evaluate whether DOACs and DEX could be safely administered together.
Currently, there is limited robust evidence to support a clinically significant interaction between DEX and DOAC. A prospective observational study involving a small group of COVID‐19 patients (n = 33) found no significant differences in plasma concentrations of DOACs when administered with DEX. 7 However, the limited power of this study, combined with the known high inter‐individual variability of DOACs, means that no definitive conclusion can be drawn. In addition, a retrospective study based on a large cohort of COVID‐19 patients reported no increased risk of thromboembolic events in patients treated with both DOACs and DEX. 8
There is therefore some evidence suggesting that the concomitant use of DOACs and DEX is possible, but few clinicians currently feel confident in this approach. This hesitancy is partly due to the poorly characterized in vivo inducing power of DEX, which has been explored in only a limited number of methodologically questionable studies. 2 , 9 , 10 , 11 Using physiologically based pharmacokinetic modeling (PBPK), we investigated the impact of DEX administered in the context of SARS‐CoV‐2 infection on the pharmacokinetics of APX and RVX. Additionally, we conducted several sensitivity studies, considering a worst‐case scenario regarding the inducing power of DEX on CYP3A and P‐gp. Given that prednisone (PRD) is often used as an alternative to DEX due to its presumed lesser inducing effect on CYP3A and P‐gp, we also conducted analyses with prednisone in the presence of APX and RVX.
METHODS
PBPK simulation: Generalities
The absorption, distribution, metabolism, and excretion (ADME) simulator SimCYP® version 22 software (CERTARA UK Limited, Simcyp. Division, Sheffield, UK) was used as the platform for PBPK simulation.
The default “Sim Healthy Volunteers” population in SimCYPwas used for the simulations. However, the CYP3A4 first‐order degradation rate constant (k deg) was modified from 0.0193 to 0.008/h, as this adjustment has been shown to better represent both the magnitude and time‐course of CYP3A4 induction with rifampin. 12 Indeed, Prueksaritanont et al. 12 have shown that the time course of return to baseline for midazolam PK following 28 days of rifampin administration to healthy volunteers suggest that a CYP3A4 kdeg of 0.008/h is required to describe the time course of the “de‐induction.”
The area under the curve (AUC) and maximum concentration (C max) were reported in ng/mL*h and ng/mL, respectively. All concentrations were simulated in plasma. For each clinical scenario, 10 trials with 100 participants each were simulated. The induction effect of the perpetrator compound was measured by calculating the ratio of the victim drug's AUC in the presence and absence of the perpetrator.
DEX compound development: Validation of DEX induction effect on CYP3A substrates in healthy volunteers
The compound SV‐dexamethasone in the SimCYP® simulator was used as the basis for developing the DEX compound (Table S1 ). The general work‐flow for DEX validation can be found in Figure S1 . According to a previous study by Ke and Milad, 13 who developed a dexamethasone PBPK model for pregnant women, the estimated fraction metabolized (fm)CYP3A4 was 95%. This estimate is based on a sensitivity analysis where the fmCYP3A4 value was varied between 90% and 98% to match the observed AUC and peak plasma concentration (C max) ratios following itraconazole treatment. 14 The results indicated that an fmCYP3A4 value of 95% best fit the observed dexamethasone AUC and C max ratios. Consequently, the same fmCYP3A4 value was used, and a CYP3A4 intrinsic clearance (Clint) of 0.15 μL/min/pmol was calculated using the retrograde Simcyp® calculator, based on an in vivo intravenous clearance (CLiv) of 15.63 L/h. 13 Using this model as a basis, the predicted‐to‐observed AUC ratio (AUCpred/AUCobs) of 0.88 (90th CI: 0.82–0.95) met bioequivalence criteria in healthy volunteers, as shown in Figure S2 , Table S2 .
Only a few in vivo studies have investigated the induction effect of DEX. In one study using the erythromycin breath test 9 in 12 healthy volunteers, dexamethasone (8 mg administered orally twice a day for 5 days) increased CYP3A4 activity by 25%. To validate this induction effect, DEX was tested on erythromycin using the SimCYP® compound SV‐erythromycin, with age and sex matched to the original study. In the simulation replicating the original trial, erythromycin (500 mg PO single dose) was administered concurrently with the ninth dose of DEX previously (8 mg bid IV for 5 days). Using the half‐maximal induction (IndC50) of 51.22 μM and the highest maximal induction (Ind Max) of 6.6 for CYP3A4 observed in vitro from the same study, simulated erythromycin plasma exposure remained largely unchanged (Table 1 ). A sensitivity analysis was then conducted to determine the IndC50 and the Ind Max values that would result in a 25% decrease in erythromycin AUC (Figure 1 ). An Ind Max of 11 and an IndC50 of 0.2 μM were found to best describe the decrease in erythromycin AUC, with a predicted decrease of 24% (Table 1 ). These values are significantly higher and lower respectively than the in vitro values commonly reported in the literature and the Drug Interaction DataBase (DIDB®). 15 Most of the decrease in AUC was attributed to an increase in hepatic clearance of erythromycin, as expected for a molecule primarily metabolized by the liver 16 (Table 1 ).
Table 1.
Geometric mean values of predicted AUC (96–108 h) and hepatic clearance for systemic plasma concentration of erythromycin (500 mg PO, single dose) when administered concurrently with the 9th dose of DEX (8 mg IV BID for 5 days)
| Without DEX | With DEX (Ind Max 6.6; IndC50 51.22 μM) | Ratio | With optimized DEX (Ind Max 11; IndC50 0.2 μM) | Ratio | |
|---|---|---|---|---|---|
| Predicted erythromycin AUC (96–108 h) (ng/mL.h) (CI 90th) | 6,673.53 (6,235.23–7,142.63) | 6,667.05 (6,229.41–7,135.44) | 1.00 | 5,074.46 (4,765.90–5,402.99) | 0.76 |
| Predicted erythromycin hepatic clearance (CI 90th) | 23.30 (22.12–24.55) | 23.32 (22.14–24.57) | 1.00 | 30.23 (29.03–31.49) | 1.30 |
Figure 1.
Sensitivity analysis for erythromycin AUC ratio for different values of DEX CYP3A4 IndC50 and Ind Max.
The effect of DEX with the optimized value for induction was then tested on a typical CYP3A substrate, triazolam (TRZ). In a previous study, 17 the effect of DEX (1.5 mg PO for 5 days) was tested in 10 healthy volunteers who received a single dose of TRZ (0.5 mg PO) on day 5. The study found that DEX had no statistically significant effect on the pharmacokinetics of triazolam.
Using PBPK modeling, we sought to determine if the same conclusion could be reached.
The previously developed DEX compound with the optimized induction value was used in combination with the original SimCYP® compound Wsp‐triazolam without any modifications. Age and sex were matched to the original study, and the dosage and clinical trial parameters were simulated as per the original study. The simulation showed that DEX 1.5 mg PO had little effect on TRZ pharmacokinetics, consistent with the observed results from the original study 17 , which reported a 19% decrease in TRZ AUC but no significant difference between the placebo and DEX groups (Table S3). A sensitivity analysis showed that increasing the DEX dose tenfold (15 mg DEX) decreased the TRZ AUC ratio to 0.7 (Figure S3 ). In line with previous findings, an Ind Max of 11 and an IndC50 of 0.2 μM were maintained for the remainder of the analysis. This approach ensures a worst case scenario, as these values are close to the in vitro values commonly observed for rifampicin and used in validated PBPK models. 18
DEX compound development: Validation in COVID‐19 patients
Following the successful simulation in healthy volunteers, which qualified the original DEX compound for further testing, the compound was assessed in a population of COVID‐19 patients. We then tested the compound's performance in a population of 30 adult COVID‐19 patients treated with DEX from a previous study. 19 However, the initial simulation in COVID‐19 patients revealed that the compound largely underpredicted DEX plasma exposure (Figure 2 ). A middle‐out approach was used to optimize the compound with adjustments made to the absorption constant (K a), lag time and oral clearance (Clpo), using parameters obtained from a previously published population pharmacokinetic (PopPK) model derived from the same COVID‐19 patient population. 20 After these optimizations, the PBPK model was re‐evaluated, resulting in an AUC ratio that met the bioequivalence criteria (Figure 2 , Table 2 ). The input parameters for the final DEX PBPK model are shown in Table S4 .
Figure 2.
Mean values of simulated systemic plasma concentration of dexamethasone (DEX) over time with the modified and the base model (bright and dark green line respectively), and observed (orange dots) systemic plasma concentration of single dose DEX 6 mg PO over time in COVID‐19 patients. Observed values were obtained from Abouir, K. et al. 19
Table 2.
Mean values of predicted and observed C max and AUC (0–8 h) for systemic plasma concentration of single dose dexamethasone 6 mg PO in healthy volunteers
| Mean predicted (90th CI around the geometric mean) | Mean observed | Predicted/observed ratio | |
|---|---|---|---|
| C max (ng/mL) | 139.06 (124.36–139.11) | 133.86 | 1.04 (0.93–1.04) |
| AUC (0–8 h) (ng/mL.h) | 689.61 (647.68–698.53) | 754.02 | 0.91 (0.86–0.93) |
Observed values were obtained from Abouir, K. et al. 19
Testing the inducing effect of DEX on apixaban (APX) and rivaroxaban (RVX)
After the successful validation of the DEX model against a classical CYP3A probe drug, we proceeded to test the induction effect of DEX on two DOACs, apixaban and rivaroxaban (APX and RVX). The compounds used were previously validated in a cohort of hospitalized patients 21 and largely based on Otsuka's models, 22 which were developed and validated for drug–drug interaction (DDI) purposes. Detailed input parameters for these compounds are available in the referenced study. 21 In brief, the model accounted for CYP3A hepatic metabolism and P‐gp intestinal transport for both APX and RVX, whereas P‐gp renal and hepatic transports were only considered for RVX. 21 , 22 In this analysis, an induction effect of DEX on intestinal and liver P‐gp was added but not for the kidney. 23 , 24 This decision was based on a rodent model study where high dose of DEX (1 mg/kg) slightly reduced P‐gp expression in the kidney 24 (Table S3 ). As no IndC50 for DEX is available for the ABCB1 gene, and no in vivo studies have investigated the induction effect of DEX on a typical P‐gp substrate, we used the same IndC50 as for CYP3A4. This is consistent with observations made for rifampicin. 18 , 25 The Ind Max values were derived from a study in a rodent model where high dose of DEX (1 mg/kg) was administered, and ABCB1 mRNA expression was measured in different organ tissues. 24 The scenarios tested were based on realistic COVID‐19 treatment regimens clinical scenario: APX 5 mg BID or RVX 20 mg OD, both given concomitantly with DEX 6 mg OD for 10 days. Several sensitivity analyses were conducted on both CYP3A4 and P‐gp. The turnover rate for P‐gp remained unchanged in the SimCYP® healthy volunteer population model (0.054/hour), as it is based on a meta‐analysis of 13 independent published studies from seven laboratories. 25
Test of the inducing effect of prednisone (PRED) on APX and RVC
Given that PRED is often used in the clinic as an alternative to DEX due to its supposedly lesser inducing effect, we tested its induction effect at an anti‐inflammatory dose equivalent to DEX 40 mg OD. The previously validated PRED compound from Mercantonio et al. 26 was used without modification. The same clinical scenarios and sensitivity analyses were performed as for DEX to assess and compare the potential inducing effects of PRED on APX and RVX.
RESULTS
Inductive effect of DEX on APX and RVX
The concomitant administration of DEX 6 mg OD on APX or RVX for 10 days led to a decrease of 21 and 19% in AUCt, and 16 and 11% in C max, respectively (Table 3, Figures S4 and S5 ). The sensitivity analysis for CYP3A4 and intestinal P‐gp (Figure 3 ) for APX indicated that DEX induction had a limited effect up to concentrations of 2 μM, even at high Ind Max values. However, below 2 μM, a strong effect of Ind Max was observed. The sensitivity to P‐gp induction was notably less pronounced compared to CYP3A4 (Figure 3 ). In fact, nearly the maximum induction effect was already achieved for APX when P‐gp had the least influence (high IndC50 and low Ind Max), whereas no reduction in AUC was observed when the CYP3A4 induction values were less pronounced. Similar findings were observed in the sensitivity analysis for RVX (Figure 4 ). Most of the induction effect was mediated by CYP3A4, followed by intestinal P‐gp and liver P‐gp, which had a very limited effect even at very low IndC50 values.
Table 3.
Geometric mean values of predicted C max and AUCt for systemic plasma concentration of apixaban (5 mg bid) and rivaroxaban (20 mg bid) given concomitantly with dexamethasone (6 mg OD) or prednisone (40 mg OD) for 10 days in healthy volunteers
| Geometric Mean predicted without inhibition (90th CI) (day 10) | Geometric Mean predicted with DEX (90th CI) (day 10) | DEX ratio | Geometric Mean predicted with PRED (90th CI) (day 10) | PRED ratio | |
|---|---|---|---|---|---|
| Apixaban C max (ng/mL) | 168.74 (162.94–174.76) | 141.42 (135.59–147.49) | 0.84 | 166.55 (160.73–172.57) | 1.00 |
| Apixaban AUCt (ng/mL.h) | 1305.23 (1255.43–1356.96) | 1031.43 (983.04–1082.22) | 0.79 | 1292.81 (1243.32–1344.26) | 1.00 |
| Rivaroxaban C max (ng/mL) | 206.16 (192.52–217.37) | 183.27 (174.06–192.96) | 0.89 | 205.75 (195.17–216.61) | 1.00 |
| Rivaroxaban AUCt (ng/mL.h) | 1489.56 (1386.89–1599.82) | 1203.55 (1123.09–1289.78) | 0.81 | 1487.10 (1384.75–1597.00) | 1.00 |
Figure 3.
Sensitivity analysis for apixaban (APX) AUC ratio as a function of dexamethasone (DEX) IndC50 and Ind Max values for CYP3A4 (left) and intestinal P‐gp (right).
Figure 4.
Sensitivity analysis for rivaroxaban (RVX) AUC ratio as a function of dexamethasone (DEX) IndC50 and Ind Max for CYP3A4 (up), intestinal P‐gp and hepatic P‐gp (bottom).
Inductive effect of PRED on APX and RVX
Administration of PRED at 40 mg OD for 10 days had no significant effect on APX or RVX AUCt (Table 3 ). Sensitivity analyses for CYP3A4 and intestinal P‐gp IndC50 and Ind Max (Figure S6 ) showed that PRED had a less pronounced effect on the APX AUC ratio compared to DEX. When using similar CYP3A4 IndC50 and Ind Max values as DEX (~ 0.2 and 12 μM, respectively), a 10% decrease in APX AUC ratio was observed. For RVX, sensitivity analysis yielded results for CYP3A4 consistent with those previously seen (Figure S7 ). Intestinal P‐gp was less sensitive to PRED induction, and hepatic P‐gp was almost insensitive to PRED induction, even at low and high values of IndC50 and Ind Max, respectively.
DISCUSSION
In the present study, we investigated the impact of DEX and PRED on APX and RVX exposure using PBPK modeling and in a clinical scenario mimicking the dose and timing of administration in COVID‐19 patients. After validating the induction effect of the DEX compound on two CYP3A4 substrates using the limited available studies, we optimized the compound in a COVID‐19 patient population, where significantly higher DEX plasma concentrations were observed compared to healthy volunteers. 19 Our PBPK‐based PK simulations show a 20% decrease in the AUC of APX and RVX when DEX was administered at 6 mg PO for 10 days. This result likely represents a worst‐case scenario, as the IndC50 and Ind Max values used in the DEX compound were comparable to those observed in vitro for rifampicin, a compound with a more important inductive property on CYP3A4 than DEX. 18 , 27 The actual in vivo inductive effect of DEX has been debated, 3 partly because some in vitro studies have used DEX concentrations far exceeding those typically encountered in clinical settings (as high as 50 or 100 μM), 28 , 29 while others have found a significant effect at much lower concentrations (0.01 μM). 30 However, most of the observed values for IndC50 are much higher than those used to optimize the DEX compound in our study. In vivo studies are very limited 9 , 11 and not always well designed to allow the DEX‐inducing effect to occur, primarily due to the short study duration or the use of DEX dose that are too low. 17
Therefore, our input parameters for induction were mainly based on the study by McCune et al, 9 which did not examine the effect of DEX on erythromycin plasma exposure directly, but using the erythromycin breath test, a validated method for assessing in vivo CYP3A activity. 31 However, an overestimation of the DEX induction effect on CYP3A by this method is not excluded for two main reasons. First, there is considerable inter‐individual variability, and the DEX induction effect is dependent on the CYP3A4 basal activity in each subject. 9 Erythromycin is not exclusively metabolized by CYP3A, but it is also a substrate for P‐gp and it is highly unstable at both acid and neutral pH, being subject to degradation via non‐metabolic clearance pathways. 31 Our simulations, which likely represent a worst‐case scenario, showed only a 20% reduction in APX and RVX exposure, which may be considered acceptable by clinicians in most clinical situations, except in cases of off‐label underdosing or acute venous thromboembolism, where the exposure‐event association has been less studied compared to atrial fibrillation. 32 This reduction is modest compared to the 50% decrease in AUCs of APX and RVX observed with rifampicin, a known potent inducer of CYP3A. 33 The relatively low dependence of APX and DEX on CYP3A metabolism 33 likely places these molecules at lower risk of significant induction compared to the more typical CYP3A substrates, such as erythromycin or triazolam. The available clinical studies tend not to show an effect of DEX on DOAC plasma concentrations but suffer from significant limitations. 7 , 34 These studies have included a small number of patients, and due to the large interindividual variability and lack of statistical power, definitive conclusions cannot be drawn. In this debated case, PBPK modeling can thus provide valuable support for the use of DOACs with DEX and encourage clinicians to conduct prospective studies.
PRED is thought to have a less pronounced inducing effect on CYP3A compared to DEX. 6 However, the data supporting this hypothesis are very limited. One in vitro study demonstrated a similar inducing effect on CYP3A in human hepatocytes for DEX and PRED at concentrations of 50–100 μM. 29 While there is some indirect in vivo evidence supporting that PRED may induce CYP3A 26 , 35 no studies have specifically tested the inductive effect of PRED at anti‐inflammatory doses equivalent to those of DEX. In vitro studies suggest that the activation of the pregnane X receptor (PXR) is dependent on glucocorticoid receptor activity at concentrations below 10 μM of DEX. 2 This finding implies that PRED could have a similar inducing effect as DEX when administered at equipotent anti‐inflammatory doses. However, our PBPK simulations indicate that PRED has a less pronounced inducing effect on the PK of DOACs such as APX and RVX, even when considering doses equipotent to those of DEX. The observed difference in inductive effect may be attributed to the variations in the Ind Max values reported in the literature. The underlying cellular mechanisms that contribute to the difference in induction between PRED and DEX remain to be fully elucidated. Since PRED is also an available and low‐cost drug, it could represent an interesting alternative.
The present study has several limitations. First, our PBPK simulations were not directly verified through in vivo studies, which limits the ability to fully validate our findings. Nevertheless, the results are consistent with the recent and limited clinical observations reported in the literature, where no significant effect of DEX on DOAC exposure or related clinical events was observed. 7 , 8 The absence of in vivo study specifically addressing P‐gp induction by DEX also constrained our ability to validate the compound for typical P‐gp substrates. To address this limitation, we adopted a conservative approach by using the same IndC50 value for P‐gp as we did for CYP3A4, acknowledging that this would represent a worst‐case scenario. Our simulation indicated that CYP3A4 mediated the majority of the induction effect, suggesting that the impact of DEX on P‐gp substrates may be less pronounced. However, this hypothesis needs further investigation to more accurately determine whether pure P‐gp substrates, such as dabigatran or edoxaban, can be safely co‐administered with DEX.
FUNDING
No funding was received for this work.
CONFLICTS OF INTEREST
All authors declare no competing interests for this work.
AUTHOR CONTRIBUTIONS
J.T., K.A., F.G., Y.D., and C.F.S. wrote the manuscript; J.T. and C.F.S. designed the research; J.T. and K.A. performed the research; J.T. analyzed the data.
Supporting information
Data S1
DATA AVAILABILITY STATEMENT
The compound files and simulation data are available upon request to interested researchers.
References
- 1. Bhimraj, A. et al. Infectious diseases society of america guidelines on the treatment and management of patients with COVID‐19. Clin. Infect. Dis. ciaa478 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Bourdin, V. et al. Drug–drug interactions involving dexamethasone in clinical practice: myth or reality? J. Clin. Med. 12, 7120 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Smythe, M.A. , Burns, C. , Liu, Q. & Garwood, C.L. Potential dexamethasone–direct oral anticoagulant drug interaction: is this a concern in COVID? Ann. Pharmacother. 56, 319–329 (2022). [DOI] [PubMed] [Google Scholar]
- 4. Terrier, J. et al. Drug‐drug interactions with direct Oral anticoagulants: practical recommendations for clinicians. Am. J. Med. 134, 939–942 (2021). [DOI] [PubMed] [Google Scholar]
- 5. Information on COVID‐19 treatment, prevention and research. COVID‐19 Treatment Guidelines at <https://www.covid19treatmentguidelines.nih.gov/>
- 6. Cuker, A. et al. American Society of Hematology 2021 guidelines on the use of anticoagulation for thromboprophylaxis in patients with COVID‐19. Blood Adv. 5, 872–888 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Potere, N. et al. Direct oral anticoagulant plasma levels in hospitalized COVID‐19 patients treated with dexamethasone. J. Thromb. Thrombolysis 53, 346–351 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Kravchenko, O.V. et al. Drug–drug interaction between dexamethasone and direct‐acting oral anticoagulants: a nested case–control study in the national COVID cohort collaborative (N3C). BMJ Open 12, e066846 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. McCune, J.S. et al. In vivo and in vitro induction of human cytochrome P4503A4 by dexamethasone. Clin. Pharmacol. Ther. 68, 356–366 (2000). [DOI] [PubMed] [Google Scholar]
- 10. Roberts, P.J. et al. The influence of CYP3A5 genotype on dexamethasone induction of CYP3A activity in African Americans. Drug Metab. Dispos. 36, 1465–1469 (2008). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Moore, M.J. , Hardy, R.W. , Thiessen, J.J. , Soldin, S.J. & Erlichman, C. Rapid development of enhanced clearance after high‐dose cyclophosphamide. Clin. Pharmacol. Ther. 44, 622–628 (1988). [DOI] [PubMed] [Google Scholar]
- 12. Prueksaritanont, T. et al. Drug‐drug interaction studies: regulatory guidance and an industry perspective. AAPS J. 15, 629–645 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Ke, A.B. & Milad, M.A. Evaluation of maternal drug exposure following the administration of antenatal corticosteroids during late pregnancy using physiologically‐based pharmacokinetic modeling. Clin. Pharmacol. Ther. 106, 164–173 (2019). [DOI] [PubMed] [Google Scholar]
- 14. Varis, T. , Kivistö, K.T. , Backman, J.T. & Neuvonen, P.J. The cytochrome P450 3A4 inhibitor itraconazole markedly increases the plasma concentrations of dexamethasone and enhances its adrenal‐suppressant effect. Clin. Pharmacol. Ther. 68, 487–494 (2000). [DOI] [PubMed] [Google Scholar]
- 15. DIDB® ‐ The drug interaction database. Certara drug interaction solutions at <https://www.druginteractionsolutions.org/solutions/drug‐interaction‐database/>
- 16. Farzam, K. , Nessel, T.A. & Quick, J. Erythromycin (StatPearls Publishing, Treasure Island, FL, 2024). StatPearls. <http://www.ncbi.nlm.nih.gov/books/NBK532249/> [PubMed] [Google Scholar]
- 17. Villikka, K. , Kivistö, K.T. & Neuvonen, P.J. The effect of dexamethasone on the pharmacokinetics of triazolam. Pharmacol. Toxicol. 83, 135–138 (1998). [DOI] [PubMed] [Google Scholar]
- 18. Almond, L.M. et al. Prediction of drug‐drug interactions arising from CYP3A induction using a physiologically based dynamic model. Drug Metab. Dispos. 44, 821–832 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Abouir, K. et al. Dexamethasone exposure in normal‐weight and obese hospitalized COVID‐19 patients: an observational exploratory trial. Clin. Transl. Sci. 15, 1796–1804 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Świerczek, A. & Jusko, W.J. Anti‐inflammatory effects of dexamethasone in COVID‐19 patients: translational population PK/PD modeling and simulation. Clin. Transl. Sci. 16, 1667–1679 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Terrier, J. et al. Apixaban and rivaroxaban's physiologically‐based pharmacokinetic model validation in hospitalized patients: a first step for larger use of a priori modeling approach at bed side. CPT Pharmacometrics Syst. Pharmacol. 12, 1872–1883 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Otsuka, Y. , Choules, M.P. , Bonate, P.L. & Komatsu, K. Physiologically‐based pharmacokinetic modeling for the prediction of a drug‐drug interaction of combined effects on P‐glycoprotein and cytochrome P450 3A. CPT Pharmacometrics Syst. Pharmacol. 9, 659–669 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. van de Kerkhof, E.G. , de Graaf, I.A.M. , Ungell, A.‐L.B. & Groothuis, G.M.M. Induction of metabolism and transport in human intestine: validation of precision‐cut slices as a tool to study induction of drug metabolism in human intestine in vitro. Drug Metab. Dispos. 36, 604–613 (2008). [DOI] [PubMed] [Google Scholar]
- 24. Demeule, M. , Jodoin, J. , Beaulieu, É. , Brossard, M. & Béliveau, R. Dexamethasone modulation of multidrug transporters in normal tissues. FEBS Lett. 442, 208–214 (1999). [DOI] [PubMed] [Google Scholar]
- 25. Brouwer, K.L.R. et al. Regulation of drug transport proteins‐from mechanisms to clinical impact: a white paper on behalf of the international transporter consortium. Clin. Pharmacol. Ther. 112, 461–484 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Marcantonio, E.E. et al. Prednisone has no effect on the pharmacokinetics of CYP3A4 metabolized drugs – midazolam and odanacatib. J. Clin. Pharmacol. 54, 1280–1289 (2014). [DOI] [PubMed] [Google Scholar]
- 27. Kato, M. , Chiba, K. , Horikawa, M. & Sugiyama, Y. The quantitative prediction of in vivo enzyme‐induction caused by drug exposure from in vitro information on human hepatocytes. Drug Metab. Pharmacokinet. 20, 236–243 (2005). [DOI] [PubMed] [Google Scholar]
- 28. Usui, T. , Saitoh, Y. & Komada, F. Induction of CYP3As in HepG2 cells by several drugs. Association between induction of CYP3A4 and expression of glucocorticoid receptor. Biol. Pharm. Bull. 26, 510–517 (2003). [DOI] [PubMed] [Google Scholar]
- 29. Pichard, L. , Fabre, I. , Daujat, M. , Domergue, J. , Joyeux, H. & Maurel, P. Effect of corticosteroids on the expression of cytochromes P450 and on cyclosporin A oxidase activity in primary cultures of human hepatocytes. Mol. Pharmacol. 41, 1047–1055 (1992). [PubMed] [Google Scholar]
- 30. Shi, D. , Yang, D. & Yan, B. Dexamethasone transcriptionally increases the expression of the pregnane X receptor and synergistically enhances pyrethroid esfenvalerate in the induction of cytochrome P450 3A23. Biochem. Pharmacol. 80, 1274–1283 (2010). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31. Rivory, L.P. , Slaviero, K.A. , Hoskins, J.M. & Clarke, S.J. The erythromycin breath test for the prediction of drug clearance. Clin. Pharmacokinet. 40, 151–158 (2001). [DOI] [PubMed] [Google Scholar]
- 32. Rizos, T. et al. Low exposure to direct oral anticoagulants is associated with ischemic stroke and its severity. J. Stroke 24, 88–97 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33. Foerster, K.I. , Hermann, S. , Mikus, G. & Haefeli, W.E. Drug–drug interactions with direct Oral anticoagulants. Clin. Pharmacokinet. 59, 967–980 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34. Bosch, F.T.M. , Candeloro, M. , Potere, N. , Porreca, E. , di Nisio, M. & Kamphuisen, P.W. Effect of dexamethasone on direct Xa‐inhibitor oral anticoagulant plasma levels in patients with COVID‐19. Thromb. Res. 205, 106–109 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35. Jia, S.‐J. et al. Interactive effects of glucocorticoids and cytochrome P450 polymorphisms on the plasma trough concentrations of Voriconazole. Front. Pharmacol. 12, 666296 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data S1
Data Availability Statement
The compound files and simulation data are available upon request to interested researchers.