Influence of a Short Course of Ritonavir Used as Booster in Antiviral Therapies Against SARS-CoV-2 on the Exposure of Atorvastatin and Rosuvastatin

Nirmatrelvir/ritonavir (Paxlovid®) was approved in January 2022 for the early treatment of SARS-CoV-2-infected patients at high risk of severe courses. Its component ritonavir is used as a boosting agent to reach and maintain therapeutic drug concentrations of nirmatrelvir throughout the dosing interval by inhibiting a variety of cytochrome P450 (CYP) isozymes including CYP3A4, which is the main CYP involved in nirmatrelvir metabolism. Older patients or patients with chronic conditions such as cardiovascular disease or diabetes have a higher risk of suffering a severe course of SARS-CoV-2 and often concurrently take additional medications [1]. Intake of nirmatrelvir/ritonavir within 3 days after being positive for SARS-CoV-2 significantly reduces the risk in this population [2]. Complicating matters further, patients with mild symptoms of SARS-CoV-2 infection are usually treated at home, where surveillance is limited and adverse effects from drug-drug interactions (DDIs) may go unnoticed. Recommendations for the management of DDIs between nirmatrelvir/ritonavir and other medicines [3] are based on assumptions derived from data available from long-term treatment with ritonavir as it is used in life-long antiviral therapy for patients infected with the human immunodeficiency virus [4]. However, previous DDI studies often used higher ritonavir doses or longer treatment periods, so their results may not adequately reflect the magnitude of the interaction observed with short-term treatment with small booster doses.

Ritonavir primarily and irreversibly inhibits CYP3A4 and to lesser extent CYP2D6, CYP2C19, CYP2C8, and CYP2C9 [5, 6]. It also inhibits hepatic uptake transporters such as several organic anion-transporting polypeptides (OATP) and efflux transporters such as breast cancer resistance protein (BCRP) and P-glycoprotein [7]. In addition, it induces CYP1A2, CYP2B6, CYP2C9, and CYP2C19 [8]. The net sum of the effect within the first 5 days remains unclear but is likely dominated by rapid inhibition [9], while treatment appears to be too short for induction of CYP3A albeit possibly not for all CYP isozymes [10].

Statins are widely used as cholesterol-lowering drugs and are one of the mainstays in the prevention and treatment of atherosclerotic vascular disease, as they reduce the risk of major cardiovascular and cerebrovascular events and cardiovascular mortality [11]. The different statins have very different metabolic pathways, which determine their risk profile for DDIs. Atorvastatin is mainly metabolized by CYP3A4 to form its two active metabolites 2-hydroxy atorvastatin and 4-hydroxy atorvastatin [12], while CYP-mediated metabolism of rosuvastatin to its active metabolite N-desmethyl rosuvastatin is small (10% of the dose) [13]. In addition to a reduction in CYP activity, functional genetic polymorphisms [14] and inhibition of hepatic uptake transporters such as several OATPs or the efflux transporter BCRP can also increase systemic exposure and the risk of adverse events [15]. Rhabdomyolysis due to statin therapy is a rare but serious adverse event [16], and the risk increases with exposure, which is why high statin doses [17] and impaired drug clearance increase its likelihood [18]. Even though rhabdomyolysis often evolves over a time course of months, there are case reports describing the onset of rhabdomyolysis within 1 week [19, 20]. Strong inhibition of statin clearance by ritonavir therefore likely poses a risk particularly in patients on high statin doses.

We therefore aimed to assess the DDI potential of a short-term ritonavir treatment course on the exposure of atorvastatin and rosuvastatin. In addition, a microdose of midazolam was administered to evaluate the contribution of CYP3A4 to the overall extent of the DDI.

In 2021, atorvastatin was the most commonly prescribed statin in Germany (over 160 million defined daily doses), while rosuvastatin saw the largest increase in prescriptions [36]. The widespread use of statins makes their combination with nirmatrelvir/ritonavir probable in clinical routine. Although ritonavir is known to alter the activity of numerous CYP isozymes and drug transporters, the effects and time-course of short-term, low-dose ritonavir treatment have not been studied in detail, and the expected DDIs and recommendations for outpatient management of potential DDIs are mostly based on studies with higher doses and longer treatment courses of ritonavir.

In our trial, the administration of 100 mg ritonavir twice daily for 5 days increased atorvastatin AUC_∞ 4.76-fold, while both metabolites dropped below the LLOQ. This suggests at first glance inhibition of hepatic CYP3A4 metabolism. But surprisingly, there was no correlation between midazolam and atorvastatin ratios, and t_1/2 did not change, which would have been expected if the increase in atorvastatin AUC was caused mainly by the inhibition of hepatic CYP3A4 metabolism. Ritonavir is also a potent inhibitor of drug transporters in vitro and in vivo and inhibits the efflux transporter p-gp [37] and the hepatic uptake transporters OATP1B1 and OATP1B3 [38], which mainly mediate hepatic atorvastatin uptake [39] but are not involved in midazolam disposition [40]. In addition, it has also previously been shown that the hepatic clearance of atorvastatin is mainly determined by the hepatic uptake via OATPs and not by (subsequent) CYP3A4 metabolism [41]. This indicates that effects of ritonavir causing irreversible intestinal inhibition of CYP3A4 [31], reversible inhibition of p-gp [42], and substantial blockage of hepatic atorvastatin uptake by OATP inhibition best explain the increase in atorvastatin exposure. The OATP effect is thought to be reversible as it disappeared after a wash-out of the drug in vitro [43].

In contrast, the AUC_∞ of rosuvastatin was unchanged, and only C_max moderately increased, which is in line with previous findings [44, 45]. P-gp does not play a role in rosuvastatin uptake, and CYP metabolism is of limited relevance [46]. This situation may explain the clinically not relevant elevation of rosuvastatin AUC and the minor elevation of rosuvastatin C_max. In heterozygous SLCO1B1 carriers, rosuvastatin PK neither at baseline nor during ritonavir differed from carriers of two wild-type alleles, confirming earlier findings [47]. In contrast, rosuvastatin exposure in the homozygous SLCO1B1 *5/*15 carrier was 2.6-fold higher at baseline than observed in the other participants, consistent with the greater AUC of 19–68% reported in SLCO1B1 521 T > C carriers [48]. Interestingly, in this participant, ritonavir caused a 55% reduction of the AUC_∞. Such a paradoxical reaction has already been reported with a different OATP1B1 inhibitor in the same individual [49]. The mechanism of such a phenomenon is not clear but might indicate the activation of the (defective) transporter by ritonavir. Indeed, the activation of OATP1B transporters by drugs that are not transported has already been demonstrated [50], but to our best knowledge, this has not been investigated for ritonavir.

Serious adverse effects of statins, such as rhabdomyolysis, can occur after only 1 week of high statin exposures [19, 20]. The almost 5-fold increase in exposure to atorvastatin could therefore pose a risk to patients, especially if they take high-maintenance doses or other risk factors are present. On the other hand, acute discontinuation of statins may pose vascular risks, not only in patients with acute vascular syndromes [51, 52] but also in patients with chronically very high cardiovascular risk [53] and in patients with COVID-19 [54]. Temporary discontinuation of statins in patients with COVID-19, as sometimes recommended [3], was shown to be detrimental as discontinuation of statins can lead to higher in-hospital mortality [54]. Based on the observed PK changes, we therefore simulated how atorvastatin treatment schedules could be changed to keep atorvastatin exposure uninterrupted and within safe and effective margins.

Using PK modelling of atorvastatin plasma concentrations, we explored various alternative treatment regimens during ritonavir treatment that are suitable to avoid excessive atorvastatin exposure and associated adverse effects. According to this modelling, abrupt atorvastatin discontinuation rapidly leads to subtherapeutic plasma concentrations, and if atorvastatin is restarted even after a 2-day pause, plasma concentrations are still expected to temporarily double. The latter finding is presumably explained by irreversible CYP3A4 inhibition, which only gradually recovers after the discontinuation of ritonavir [31]. For patients on high potentially dangerous maintenance doses, the best dosing scheme was a step-wise dose reduction followed by an incremental increase to the original atorvastatin dose as follows: administration of the full dose on the first nirmatrelvir/ritonavir treatment day (at the same time as ritonavir), a half of the atorvastatin dose on day 2, and a quarter of the dose on days 3–6. After the discontinuation of nirmatrelvir/ritonavir, the administration of half the dose until day 9 and restart of the original dose from 5 days after discontinuation of nirmatrelvir/ritonavir. The second-best alternative to minimize exposure changes is to reduce the patient’s atorvastatin dose to either half for patients receiving 20 mg or to a quarter for patients receiving 40 mg or higher doses until 2 days after the end of nirmatrelvir/ritonavir treatment and then continue with the original dose. Some of these regimens require that the tablets be divided into quarters, which is unfortunately often not the case [29]. For patients with 10 mg atorvastatin, maintaining the original dose is usually the preferred option as exposure will likely not exceed exposures observed with the maximum licensed dose.

This trial analyzed the change in the exposure of single low doses of atorvastatin or rosuvastatin on the fifth day of ritonavir treatment compared to baseline. While this design assessed maximum inhibition, it cannot inform about the time course of inhibition in the first days of treatment. However, thorough analyses with the CYP3A4 marker substrate midazolam revealed that profound inhibition already occurs after the first ritonavir dose [23]. Furthermore, because the PK of atorvastatin is slightly non-linear [55], the effect of ritonavir on higher doses of atorvastatin may differ slightly from these estimates. In addition, we used a mechanism-based model that plausibly explains prolonged ritonavir effects by irreversible CYP3A inhibition. However, possible additional reversible ritonavir effects on drug transporters could not be distinguished by our trial design.

In conclusion, our trial showed an almost 5-fold increase in atorvastatin exposure in healthy volunteers while rosuvastatin exposure was only mildly altered. Semi-mechanistic PK modelling based on enzyme turnover, to reflect irreversible inhibition, suggests that boosting doses of ritonavir achieve nearly maximum inhibition after 24–48 h. Therefore, temporary dose reductions of atorvastatin are recommended for patients maintained on daily doses exceeding 10 mg. The results of this trial suggest that patients receiving rosuvastatin do not need to change their dose during short courses of ritonavir.