The dolutegravir/valproic acid drug–drug interaction is primarily based on protein displacement Free

Abstract

Introduction

Concurrent use of anticonvulsants and combined ART (cART) in people living with HIV (PLHIV) warrants careful evaluation of drug–drug interactions (DDIs) as both treatments are known for their potency to alter drug exposure. Dolutegravir-containing cART is a first-line treatment option for PLHIV.¹ Valproic acid is a stabilizing agent for treatment of bipolar disorder and a widely used anticonvulsant drug. Palazzo et al.² described an unexpected DDI between dolutegravir and valproic acid in two PLHIV, causing >80% reduction in total dolutegravir exposure. A therapeutic drug monitoring (TDM) registry analysis also identified lower total trough dolutegravir concentrations in individuals treated with valproic acid (132 ng/mL) versus individuals not treated with valproic acid (760 ng/mL).³ Competition for binding sites on plasma proteins has been suggested as the underlying mechanism for the low total dolutegravir plasma concentrations, as dolutegravir and valproic acid are both highly protein bound (>90%). This seems plausible, as valproic acid displaces some other highly protein bound drugs, such as phenytoin.⁴ However, the exact mechanism for this putative DDI between dolutegravir and valproic acid and its clinical relevance has yet to be proven.⁵

Total drug concentrations (sum of plasma protein bound and unbound drug concentrations) can serve as a predictor for efficacy outcomes, provided that an adequate exposure–response relationship is established. However, only the proportion of the drug that is unbound to plasma proteins can diffuse through membranes and exert its biological effect.⁶ In a situation of potential protein displacement, total drug concentrations can decrease, but unbound drug concentrations may not be affected. This may lead to unnecessary advice for dose increase. In such a scenario, unbound drug concentrations give more accurate information about drug exposure for highly protein bound drugs. Current literature provides no information about unbound dolutegravir concentrations in individuals on valproic acid. These data are needed to support the protein displacement theory and to rule out the involvement of other mechanisms, such as increased metabolism or excretion. In this prospective substudy we aimed to unravel the pharmacological mechanism(s) behind the dolutegravir/valproic acid DDI by measuring total and unbound dolutegravir plasma concentrations and dolutegravir’s main glucuronide metabolite in samples from PLHIV who were treated concurrently with dolutegravir and valproic acid.

Methods

Within the ‘Latency reversing agents United as a Novel Anti-HIV strategy’ (LUNA)⁷ trial we conducted a pharmacokinetic (PK) subanalysis in which we quantified in plasma total and unbound dolutegravir concentrations, as well as dolutegravir-glucuronide, dolutegravir’s main inactive metabolite.⁸ The LUNA trial is an open-label randomized controlled trial evaluating the latency reversing effects of pyrimethamine and valproic acid on the HIV reservoir in adults successfully treated with cART (trial registration: ClinicalTrials.gov NCT03525730). Eligible PLHIV were randomized to a 2 week treatment with valproic acid or pyrimethamine alone, both or neither. Patients were treated with 30 mg/kg/day valproic acid (Depakine enteric^®) orally (divided into two equal doses) and/or once-daily pyrimethamine orally in the morning from day 0 to day 13. On day 0, cART and study medication intake was supervised. Since previous literature showed dolutegravir exposure to increase up to 66% when taken with a meal,⁹ participants on once-daily 50 mg dolutegravir-based cART were instructed to take this in the morning with food. Self-reported adherence was assessed at each study visit and empty pill strips were collected after the 2 week treatment. The valproic acid dosage used was based on the dosage that is expected to result in a drug exposure associated with anti-latency effects and is comparable to the maintenance dosage of valproic acid for the chronic treatment of epilepsy.¹⁰ In the PK substudy, dolutegravir PK was compared between the two LUNA treatment arms receiving valproic acid (group I) and, as a control, the two treatment arms not receiving valproic acid (group II). Participants in both group I and group II could be on pyrimethamine, but since it is not known for clinically relevant DDIs with valproic acid or dolutegravir we did not expect this to impact the results of our analysis. Group I also allowed for a within-subject comparison of dolutegravir PK with concurrent valproic acid (days 1, 7 and 14) versus without valproic acid (days 0 and 42). Trough (C_trough,t = 24 h) blood samples for quantification of total and unbound dolutegravir and dolutegravir-glucuronide plasma concentrations were collected on days 1, 7, 14 and 42. On day 0, one sample was collected before intake of valproic acid with cART (t = 0 h) and one sample 6 h after intake. Dolutegravir plasma concentrations were quantified with a validated UPLC–MS/MS bioquantification method.¹¹ Unbound concentrations were obtained by ultrafiltration [lower limit of quantification (LLOQ)  = 0.5 μg/L]. Dolutegravir-glucuronide concentrations (LLOQ = 0.005 mg/L)¹² and the molar metabolic ratio for the main metabolite versus parent drug [dolutegravir-glucuronide (molecular weight = 594.51)/total dolutegravir (molecular weight = 419.38)] were determined. Unbound dolutegravir concentrations are reported in μg/L, while dolutegravir-glucuronide and total dolutegravir plasma concentrations are reported in mg/L. Descriptive statistics for dolutegravir plasma concentrations and geometric means (GMs) with coefficient of variation (CV, %) were calculated with Phoenix WinNonlin 64^® version 8.1 (Certara Inc., Princeton, NJ, USA). Due to the low sample size, no formal statistical test was done. Adverse events were graded according to Common Terminology Criteria for Adverse Events (‘CTCAE’) version 5.0. The DDI was considered not clinically relevant if ≥90% of all individual trough unbound dolutegravir plasma concentrations remained above the proposed in vitro minimal effective concentration (EC₉₀) of 0.85 μg/L for unbound dolutegravir concentrations.¹³

Ethics

The study was approved by the Institutional Review Board (MEC-2017-476) and all patients provided written informed consent.

Results and discussion

Ten male individuals on dolutegravir-based cART enrolled in the PK substudy: six in group I (valproic acid) and four in group II (control). The median (IQR) age was 54 (40–68) years and the median (IQR) weight was 82.4 (66.5–116) kg. Participants did not use co-medication known for relevant interactions with either valproic acid or dolutegravir. Albumin levels were within normal ranges and eGFR (CKD-EPI) was >50 mL/min/1.73 m². Two participants stopped valproic acid after 4 and 7 days because of grade 1 and 2 adverse drug events (ADEs); valproic acid dosage was reduced to 50% from day 2 in one participant because of grade 1 and 2 ADEs. Table 1 presents dolutegravir PK results at each timepoint. Unbound concentrations decreased during valproic acid treatment (Figure 1a), but remained above the proposed in vitro EC₉₀ of 0.85 μg/L for protein unbound dolutegravir concentrations in 5/6 (83%) participants and in 11/13 (85%) samples collected on days 1, 7 and 14 (Table 1).

Figure 1.

Mean dolutegravir unbound (a) and total (b) plasma concentrations with and without valproic acid (with SD) during the study. DTG, dolutegravir; VPA, valproic acid.

Open in new tabDownload slide

Total dolutegravir trough concentrations decreased sharply during valproic acid treatment, while in the control group total dolutegravir concentrations remained stable (Figure 1b). In group I (valproic acid) we observed a parallel increase in the unbound fraction of dolutegravir during valproic acid treatment: 0.28% on day 0 and 0.25% on day 42 versus fractions between 0.46% and 0.59% on days 1, 7 and 14 (Table 1). No such relevant alterations were seen in the unbound fractions in group II (control) (median = 0.25%). During valproic acid co-administration the dolutegravir-glucuronide molar metabolic ratio appeared to increase, while in the control group it remained stable. Our results show high inter-subject variability, most prominently for unbound and dolutegravir-glucuronide concentrations.

Two samples from one participant were below the EC₉₀ for unbound concentrations. This participant reported to be adherent, but suffered from repeated vomiting and nausea during the first 2 weeks of treatment. After discontinuation of valproic acid these symptoms resolved, but dolutegravir unbound concentration recovered to just 0.83 μg/L (C_{trough, total} 0.41 mg/L) on day 42, which is still slightly below the EC₉₀. This indicates that factors other than valproic acid treatment or vomiting may have contributed to the lower dolutegravir levels in this patient during the trial.

Total trough dolutegravir concentrations determined in samples without concurrent valproic acid are in line with population reference values for trough concentrations [GM (CV%) = 1.20 (62) mg/L for PLHIV treated with once-daily 50 mg dolutegravir].¹⁴ Furthermore, dolutegravir unbound fractions without valproic acid use are in line with available previously reported data.¹⁵ In all participants, plasma HIV-RNA remained <50 copies/mL throughout the study.

In general, if drug clearance exclusively depends on protein binding, changes in unbound fraction will affect total plasma concentrations, but unbound concentrations will remain stable.⁶ Surprisingly, we found that, along with decreased total concentrations, unbound dolutegravir concentrations during valproic acid co-administration decreased from >3 μg/L to approximately 1 μg/L during 14 days of valproic acid treatment. These findings suggest that other mechanisms might be involved in this DDI, such as induction of metabolizing enzymes leading to increased dolutegravir elimination. Valproic acid is best known as an inhibitor of several uridine diphosphate glucuronosyltransferase and cytochrome P enzymes. However, it can also exert auto-inducing effects on β-oxidation. Via this mechanism, high-dose valproic acid has been reported to mildly reduce exposure of certain antipsychotics, such as aripiprazole, albeit leading to a far smaller decrease than observed in this study.¹⁶ We did observe an increase in the metabolic ratio for dolutegravir-glucuronide during valproic acid treatment, but protein binding is not considered in this ratio. The cause of the increase in metabolic ratio could be secondary to the altered protein binding (more free dolutegravir is available for metabolism) and/or increased formation of the metabolite due to enzyme induction. Previously it has also been suggested that cation excipients in valproic acid formulations may impair absorption of dolutegravir,² but such a valproic acid formulation was not used in the LUNA trial.

Conclusions

Our prospective data on total and unbound dolutegravir concentrations confirm the previously reported decrease in total dolutegravir plasma concentrations when combined with valproic acid. This rapid and consistent decrease in total concentrations can be explained by protein displacement of dolutegravir by valproic acid, as we observed an increased unbound fraction of dolutegravir. Theoretically, this DDI might also be expected to occur between valproic acid and other integrase inhibitors. So far, however, such a DDI has not been described with other integrase inhibitors that are highly (yet not extensively) protein bound.³

In this analysis, unbound dolutegravir concentrations remained above the proposed threshold for inhibition of viral replication in all but one participant and this particular participant remained virologically suppressed. Therefore, although other mechanisms lowering dolutegravir drug exposure could be involved too, this DDI is probably not clinically relevant in adherent patients with no relevant comedication who take dolutegravir with food. In case of doubt, TDM can be considered in PLHIV on dolutegravir-containing cART and valproic acid. In clinical practice, however, measuring plasma HIV-RNA regularly and ensuring antiretroviral adherence may be a more feasible approach, since most TDM services measure total dolutegravir concentrations only. Clinicians and pharmacologists should be aware of this DDI for correct interpretation of TDM results on total and unbound dolutegravir concentrations.

Acknowledgements

We would like to thank the patients for participating in this study. We would also like to thank the members of the Erasmus Medical Center HIV Eradication Group (EHEG), the research nurses and the laboratory personnel at the Department of Virology of the Erasmus Medical Center, as well as the laboratory personnel at the Department of Pharmacy of the Radboud University Medical Center for conceptualizing the study, data collection, sample processing and analysing the PK samples.

Funding

The LUNA trial was financially supported by the Erasmus Medical Center MRace Grant (2016). No additional funding was obtained for this particular PK analysis.

Transparency declarations

D.M.B. has received honoraria and/or study grants from MSD, Gilead and ViiV Healthcare. C.R. has received study grants from Merck, Gilead and ViiV Healthcare. All other authors: none to declare.

Author contributions

C.R., H.A.B.P., P.D.J.B. and D.M.B. designed the substudy. C.R., A.V., H.A.B.P., B.J.A.R., T.E.M.S.d.V.-S., E.v.N., J.N., H.B. and M.d.M.M. recruited patients. H.A.B.P. enrolled participants and collected clinical data. P.D.J.B., A.C., K.V.-G. collected PK data and conducted the statistical analysis. P.D.J.B., H.A.B.P. and A.C. drafted the manuscript (which was revised by C.R., D.M.B., A.V., B.J.A.R., T.E.M.S.d.V.-S., E.v.N., J.N., H.B., M.d.M.M. and K.V.-G.). C.R. and D.M.B. supervised the project. All authors reviewed, critically revised and approved the manuscript.

References

Author notes