Our study suggested that the management system for the appropriate use of NMV-r facilitated the use of appropriate antivirals following the classification of DDIs and the implementation of recommended DDI management. More than 85% of patients with DDIs requiring the consideration of alternative COVID-19 therapy other than NMV-r received MOV (Fig. 2 (c)). The proportion of patients receiving NMV-r increased according to the classification of DDI management. In the NMV-r group, 95.5% of patients with concomitant medications categorized as “temporary withhold/reduction/alteration of concomitant medication” (Table 2 (d)) adhered the recommended DDI management. Even when the management that differed from the recommendations in our list was selected, safe DDI managements, such as temporary withholding or adjustment of concomitant medications, were implemented. Various adverse events caused by DDIs with NMV-r have been reported [13,14,15,16,17,18]. For immunosuppressants or cardiovascular drugs, strategies for their adjustment when used in combination with NMV-r have been proposed [28, 29]. However, no studies have evaluated comprehensive management for DDIs of NMV-r. In addition, the use of NMV-r in Japan was only 7% of that of MOV, which differs from the usage in the United States [19, 20]. In the present study, NMV-r was used in more patients than MOV and most patients with problematic DDIs received the recommended DDI management, suggesting that the management system contributed to these results. In our system, both physicians and pharmacists carefully reviewed DDIs using the constructed list, and the DDI management included those implemented following the suggestion of pharmacists. The dissemination of the list that describes the recommended DDI management and the establishment of a multidisciplinary system for reviewing DDIs may have contributed to the appropriate DDI management.

Immunosuppressants such as tacrolimus or cyclosporine were the most common concomitant medications categorized as “consider alternative COVID-19 therapy other than NMV-r”, and all patients except one case received MOV in this study. Several adverse events associated with markedly elevated blood concentrations of tacrolimus in combination with NMV-r have been reported [13,14,15,16], and management should be implemented to avoid concomitant administration of these medications. NIH guidelines classify immunosuppressants as “temporarily withhold concomitant medication” and warn that these drugs should not be used if therapeutic drug monitoring (TDM) cannot be closely conducted [8]. Another study recommended the discontinuation of tacrolimus during the treatment period of NMV-r, then resumed with dose adjustment while conducting TDM [30]. Visiting the hospital frequently and conducting TDM closely would be difficult for outpatients with mild COVID-19 due to isolation measures. Thus, immunosuppressants were categorized as “consider alternative COVID-19 therapy other than NMV-r” in our list. Other concomitant medications categorized as “consider alternative COVID-19 therapy other than NMV-r” included clonazepam and warfarin in this study. We differently categorized these drugs from the NIH guidelines [8] because frequently performing TDM or laboratory tests is difficult, similarly as in the case of immunosuppressants. A previous study reported a case series of changes in the international normalized ratio with the combination of warfarin and NMV-r [31]. When drugs that require monitoring of blood drug concentrations or pharmacological indicators are used in combination with NMV-r, a setting that allows close monitoring of the effects of DDIs is required.

Of all eligible patients in this study, 60.2% had concomitant medications with DDIs with NMV-r. In large-scale observational studies involving patients with COVID-19 or those with risk factors for severe COVID-19, the prevalence of DDIs has been simulated to be 5–20% [9, 32, 33]. Meanwhile, 60–70% of eligible patients had clinically significant DDIs in single-center retrospective studies involving patients prescribed NMV-r [11, 12]. Our facility is a university hospital and may have had more patients with risk factors for severe COVID-19, such as malignancy or prior solid-organ transplantation, and patients with concomitant medications that cause problematic DDIs with NMV-r. In addition, our constructed list was based on the JSPHCS DDI guidance [24] as well as the NIH DDI guidelines [8], thus drugs, such as etizolam or brotizolam, which are hypnotics and not listed in the NIH DDI guidelines [8], could be also detected. Although 60.2% of eligible patients had DDIs with NMV-r, only 22% (i.e., 5 patients of (a) and (c) of Fig. 2) of them had concomitant medications that required the consideration of alternative COVID-19 therapy other than NMV-r due to DDIs. Although the use of NMV-r in Japan has been substantially limited compared to the United States [19, 20], potentially more patients in Japan could benefit from the use of NMV-r by managing DDIs appropriately, and there is a need to establish a system to promote the appropriate use of NMV-r, as in this study.

In the MOV group, the sum of patients who had two or more concomitant medications that have DDIs with NMV-r was 45.2% in this study (Fig. 3). Patients for whom NMV-r could be selected according to the DDI management system (i.e., (d), (e), and (f) of Fig. 2), included those who had multiple concomitant medications requiring adjustment due to DDIs. A previous study revealed that DDIs with NMV-r were common in older adults with polypharmacy, and many DDIs involved potentially inappropriate medications (PIMs), which were candidates for deprescribing [34]. The use of NMV-r may have been discouraged in patients with polypharmacy due to the wide variety of DDIs, even when there are no concomitant medications that require consideration of alternative drugs. Implementing deprescribing for PIMs and providing a more specific indication for DDI management, such as showing a reduction dose of concomitant medication or safer alternative drugs, may increase adherence to the DDI management system even in patients who had multiple DDIs with NMV-r, and more patients could benefit from the use of NMV-r.

This study evaluated the clinical outcomes of patients with COVID-19 under the control of the management system for the appropriate use of NMV-r. In previous studies of outpatients who received NMV-r with adequate vaccination status, the proportion of hospitalization was 0.5% [95% confidence interval (CI), 0.4–0.5] and the mortality was 0.01% [95%CI, 0.01–0.02] [4], and the incidence of the composite outcome of hospitalization or death was 0.6% [95%CI, 0.4–0.7] [5]. In this study, the 30-day all-cause mortality for patients who received NMV-r was similar to those in the previous studies [4, 5]. Although the proportion of patients requiring oxygen therapy or ventilatory support was not reported in the previous studies [4, 5], the proportion of outcomes including hospitalization (i.e. less severe outcomes) in these studies [4, 5] was lower than the proportion of patients requiring oxygen therapy in this study. This may be due to the differences in patient backgrounds and the inclusion of inpatients in this study. Although MOV demonstrated no effectiveness in general patient populations with adequate vaccination status [6, 7], some observational studies including solid organ transplant recipients, patients with hematological malignancies, and patients aged ≥ 65 years with adequate vaccination status have demonstrated the effectiveness of MOV in reducing the rate of hospitalization or death [35,36,37]. In these studies, the mortality for patients who received MOV were 5.2% [95%CI, 1.9–10.9] and 1.2% [95%CI, 0.9–1.6], respectively [36, 37], indicating similar results in patients receiving MOV in this study. The proportion of patients requiring oxygen therapy or ventilatory support was not reported in these studies [36, 37]. In the present study, patients who received MOV had more risk factors, such as aged ≥ 65 years or prior solid-organ transplant, than patients who received NMV-r, as a result of the selection of the medication for COVID-19 according to the contents of DDIs. The mortality in the present study was similar to those in the previous studies [36, 37] that showed effectiveness in reducing mortality, suggesting that MOV may be a viable treatment option to improve clinical outcomes in patients with high-risk factors for whom NMV-r is unavailable due to DDIs. Although the mortality in patients who received each drug was similar to previous studies [4, 5, 36, 37], it is unclear whether the management system improved clinical outcomes in the overall patient population, because we could not compare the situation without the present management system. Further studies are warranted to determine the impact of the system that promotes the use of NMV-r in appropriate patients while avoiding problematic DDIs on clinical outcomes in patients with COVID-19.

This study has several limitations. First, this is a single-center, retrospective, observational study, and the distribution of DDIs may not represent various medical facilities across Japan. Our facility is a university hospital, with a higher proportion of patients with risk factors for severe COVID-19 than the general population. We could not evaluate the effectiveness of our system in DDI management with comparators. Additionally, some patients could not be evaluated for concomitant medications prescribed at other hospitals and clinical outcomes after treatment. Although the dosage of NMV-r should be assessed using eGFR (mL/min), which is not normalized by body surface area, we assessed it using eGFR (mL/min/1.73m2) because the height or weight of some patients was retrospectively unavailable. Second, the recommended DDI management in our constructed list was based on the NIH DDI guidelines [8] and the JSPHCS DDI guidance [24], nevertheless, the recommendations for several medications differed in their classification. This is because we considered that frequent visits and laboratory tests in outpatients with mild COVID-19 were difficult due to isolation measures, as previously mentioned. Third, this study did not evaluate adverse events due to DDIs. Future studies are needed to evaluate whether appropriate DDI management could reduce adverse events due to DDIs with NMV-r.

In conclusion, our study suggested the effectiveness of the management system for the appropriate use of NMV-r in the DDI management. Our management system promoted the use of appropriate antivirals based on the classification of DDIs with NMV-r and the implementation of recommended DDI management. Although 60% of the eligible patients had problematic DDIs with NMV-r, 22% of the patients required consideration for a change of NMV-r due to DDIs. Methods for the appropriate DDI management should be established and expanded to ensure the appropriate use of NMV-r and improve COVID-19 treatment outcomes.