Introduction

Nirmatrelvir–ritonavir reduces the risk of hospitalization among unvaccinated individuals with COVID-19 [1]. Large observational studies using administrative databases have similarly demonstrated reductions in hospitalization with nirmatrelvir–ritonavir in vaccinated individuals, however these lack detailed data on adverse drug events (ADE) including those that do not require medical attention and potentially rare effects such as liver toxicity, medication adherence and mitigation strategies for drug-drug interactions (DDI) [2–5]. In addition, these studies are prone to misclassification when determining reason for hospitalization [6]. To complement existing studies and address the potential harms of nirmatrelvir–ritonavir detailed, individual-level clinical information is necessary.

In Ontario, Canada nirmatrelvir–ritonavir is funded for outpatients based on clinical criteria determined by the Ministry of Health, including individuals who were; immunocompromised, older than age 70, age 60 and older with fewer than 3 vaccine doses, age 18 or older with less than 3 COVID-19 vaccine doses and one or more risk conditions (obesity, diabetes, heart disease, hypertension, congestive heart failure, chronic respiratory disease including cystic fibrosis, cerebral palsy, intellectual or developmental disability, sickle cell disease, moderate or severe kidney disease, moderate or severe liver disease and pregnancy) [7]. Third doses of COVID-19 vaccines were first available in August 2021 to select vulnerable populations, expanded in September 2021 to individuals at highest risk of severe illness before being available to all adults in December 2021 [8–11]. The Ontario COVID-19 Science Advisory Table was the main source of clinical guidance on the use of COVID-19 therapeutics including dosing and other clinical aspects such as how to manage potential drug-drug interactions [12]. The primary objective of this study is to describe and quantify potential drug-drug interactions (DDI) and the associated mitigation strategies employed. The secondary objectives are to characterize treatment adherence, prevalence of adverse events and describe clinical outcomes of patients who received nirmatrelvir–ritonavir.

Methods

Results

Overall, 637 individuals were prescribed nirmatrelvir–ritonavir, with 8 patients receiving a second course for a subsequent infection. The median age was 70 (IQR 51–78) and 56% (356/637) were females. The median number of risk factors for severe disease was 2, most commonly age 70 or older (52%, 332/637) and a cardiovascular condition (50%, 316/637); 45% (285/637) were immunocompromised (Table 1). Very few patients were unvaccinated (8%, 48/637) as most had received 3 or more vaccine doses (82%, 523/637). The median time from symptom onset to treatment was 2 days (IQR 1–3) with 82% (527/645) starting treatment within 3 days. Dose reduction for renal dysfunction was required for 29% (185/645) of individuals (Table 1). Those who received a second course were younger, (median age 47) and 88% (7/8) were immunocompromised; the median number of days between courses was 110.5 (IQR: 88–128).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Characteristics of study population.

https://doi.org/10.1371/journal.pone.0293302.t001

The median number of concurrent medications was 7 (IQR: 5–10) with 70% (450/645) of patients having at least 1 DDI, increasing to 82% (273/332) for older adults 70 years and older (Table 2). Of 840 DDIs, 95% (795/840) were categorized as level 2 clinically significant requiring intervention. The most common interacting co-medications were cardiovascular medications (55%, 460/840) with the vast majority prescribed for management of hypercholesterolemia, hypertension and benign prostatic hyperplasia (49% 411/840). The next most common categories were central nervous system (19%, 155/840) and oral antithrombotic agents including 27 direct oral anticoagulants (5%, 43/840) (S2 Table). Out of 829 mitigation strategies, the co-medication was held in 56% (467/829) with the majority of these being HMG CoA reductase inhibitors (246/467), dose adjusted in 22% (179/840) and an alternative drug started in 8% (64/840) of cases.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Prescribing & clinical outcomes.

https://doi.org/10.1371/journal.pone.0293302.t002

A virtual follow-up assessment was completed for 89% (572/645) of patients; those unable to be reached for follow-up were excluded in analysis of ADE, treatment adherence and clinical outcomes. Those lost to follow-up were younger, but otherwise similar. At follow-up, 62% (351/572) reported one or more ADE potentially related to nirmatrelvir–ritonavir, most commonly dysguesia (39%, 223/572), diarrhea (23%, 131/572) and nausea or vomiting (19%, 98/572). Despite ADEs, 95% (542/572) completed therapy as prescribed and 87% (495/572) stated that they were willing to take nirmatrelvir–ritonavir again in the future if offered (Table 2).

Most patients reported symptom resolution after treatment, however 32% (184/572) experienced lingering symptoms that was most commonly fatigue followed by cough. While 11% (60/572) experienced worsening of symptoms 7 days or less after completing treatment, most patients reported that their rebound symptoms were milder and none were severe enough to require retreatment. Within 28-days, 9% (54/572) visited the ED, 8% (47/572) visited a family physician or walk-in clinic and 3% (19/572) were hospitalized. Of the ED visits that did not require admission, 54% (19/35) were related to COVID-19 symptoms with most presenting within 7 days of starting nirmatrelvir–ritonavir. The other visits were unrelated to COVID-19 and occurred more than 10 days after staring therapy. Only 37% (7/19) of hospital admissions were related to COVID-19; of these, 5 were due to progression of COVID-19 symptoms (including 1 patient with rebound). Two individuals were admitted to hospital after experiencing severe gastrointestinal symptoms leading to impaired oral intake (S3 Table). No deaths occurred.

Discussion

A retrospective cohort study of 637 outpatients prescribed nirmatrelvir–ritonavir when BA.1/2 and BA.4/5 were the prevailing variants found a high prevalence of DDIs that required intervention. Nearly 70% of patients had more than one clinically significant DDI. This increased to 82% in older adults, which is similar to studies using administrative databases which found that 2/3 of older adults had ≥ 1 potential DDI [5, 20]. Almost all DDIs (98.6%) required pharmacist intervention, which likely contributed to the low rate of severe adverse events in our cohort. While a formal time-motion study was not conducted as part of this study, pharmacists typically spent 30–60 minutes per patient performing assessment, counselling and documentation activities. However, patients requiring comedication dosage changes and/or initiation of new medications while on therapy often required more attention especially in instances where changes to compliance packaging were required (i.e. blister packs). Additionally, individuals on immunosuppressive drugs (e.g. sirolimus, tacrolimus) may require therapeutic drug monitoring to guide dose adjustment following nirmatrelvir-ritonavir therapy which requires additional coordination [13].

The high frequency of ADEs (62%) did not translate into discontinuation of therapy as 95% of patients were adherent to therapy. Dysgeusia and diarrhea were most frequent at 39% and 23% respectively, which was much higher than the 6% for dysgeusia, 3% for diarrhea, and 1% for nausea observed in EPIC-HR. The higher rates in this cohort may be due to a combination of our cohort being older, more comorbid and biases related to retrospective data collection when collecting specific ADEs [1, 21]. Nevertheless, while the real-world gastrointestinal ADEs are more prominent than originally reported, they rarely led to severe ADE and hospitalization.

By 28 days following treatment, 32% (184/572) reported persistent COVID-19 symptoms. Studies conducted before the Omicron variant appeared estimated that 10–35% of individuals have symptoms that persist for at least 28 days post COVID-19 infection [22, 23]. While we were unable to perform longer term assessments, the prevalence of lingering symptoms at 28 days appears similar to those seen with long COVID. However, without a comparison group it is not possible to determine whether nirmatrelvir–ritonavir led to a faster resolution of symptoms or a reduction in long COVID. Consistent with other studies, 11% experienced rebound symptoms [24–26]. In our study, no patients with rebound were severe enough to require retreatment however, one individual required hospitalization for progression of COVID-19 symptoms.

Overall, 3% of the cohort were hospitalized within 28 days with 1% deemed related to COVID-19 progression and there were no known deaths. A recent population-based cohort study in Ontario found that individuals dispensed nirmatrelvir–ritonavir from a community pharmacy had a rate of hospitalization from COVID-19 or all cause death rate within 30 days of 2.1% as compared to 3.7% for those that did not receive nirmatrelvir–ritonavir, which is higher than that observed in our cohort. It is important to note, however, that >30% of patients in the provincial cohort were long-term care residents, which are at higher risk of COVID-19 progression and were not included in this cohort [5]. Our rate of COVID-19-related hospitalization, however, was higher than that seen in EPIC-HR (1.2% vs 0.8%) likely related to differences in patient populations.

Supporting information