2.1 Study Design and Setting

The study had a noninterference policy, and any treatment decisions about patients in the study were at the discretion of the treating physicians, which was approved by the Shanghai Changhai Hospital Ethics Committee (CHEC 2023-100). In addition, patients needed to meet several criteria for the following inclusion: patients with the coronavirus disease 2019 (COVID-19) [6] received nirmatrelvir/ritonavir treatment; blood and bronchial lavage fluid (BALF) samples were obtained from clinicians in the course of routine clinical care, which were collected at pre-dose trough concentrations after administration of NMV; bronchoalveolar lavage was collected within 1 h after blood samples; All patients were required to provide oral informed consent and be at least 18 years old. A total of 7 patients were eligible during a ten-month study period between January 2023 and October 2023. Detailed data on treatment history, treatment patterns, and outcomes were collected. (See Table S1 and Figure S1).

2.2 Determination of Nirmatrelvir in Human Plasma and Bronchoalveolar Lavage Fluid by Liquid Chromatography Coupled to Tandem Mass Spectrometry

Procedures for determining NMV in biological samples have been previously published [7, 8]. The system consisted of high-performance liquid chromatography (model 20AD; Shimadzu, Kyoto, Japan) with a QTrap hybrid quadrupole linear ion trap mass (model AB Sciex3200; Foster City, CA, USA). The NMV was detected in the positive ion mode using multiple reaction monitoring of the transitions at m/z 500.2 → 319.3. A gradient mobile phase consisting of Mobile phase-A: 0.1% formic acid solution and Mobile phase-B: acetonitrile was employed. Isocratic elution was performed with 20% mobile phase A (0.1% formic acid in water) for a total run time of 8 min. The prepared sample was pumped through a Zorbax SB-Aq C₁₈ column (250 × 4.6 mm I.D., 5 μm particle size) at a flow rate of 1.0 mL·min⁻¹ at 30℃ of column temperature.

2.3 Calculated NMV in the ELF by Urea as an Endogenous Marker

Since urea diffuses readily throughout the body, the urea concentrations in ELF and plasma are the same. We used urea as an endogenous marker of ELF dilution to quantify the apparent epithelial lining fluid concentration (C_ELF), as described previously [9]: Concentration in ELF = Concentration in BALF × Urea in serum/Urea in BALF. The urea content of BALF samples was determined by using routine biochemical tests.

2.4 Samples Processing

Blood samples: The blood samples were collected for routine laboratory tests, which were separated into two tubes and then centrifuged for plasma and serum. The plasma was collected for our testing (determination of nirmatrelvir concentration). The urea in serum was performed in the routine laboratory.

Bronchoalveolar lavage fluid: 10 mL aliquots of 0.9% normal saline were instilled into the area of pneumonia, with the fluid being immediately aspirated and placed on ice after each aliquot. The sample was sent to the laboratory and immediately centrifuged at 10,000 g for 5 min. The supernatant was separated into two independent samples for determining urea and nirmatrelvir concentrations, which were collected and frozen until the assays could be performed.

2.5 Statistical Analysis

All statistical analyses and graphing were performed using GraphPad Prism 8. Correlational analyses were performed using Pearson's correlation analysis to further examine the relationship between the plasma concentration of nirmatrelvir and its concentration in the epithelial lining fluid.

3.1 Baseline Characteristics of Study Subjects

Seven patients (male) with severe acute respiratory disease coronavirus 2 were enrolled in this study (22 samples in total, including 11 blood samples and 11 BALF samples). Six patients were classified as critically ill with being not administered orally and given drugs through a nasal cannula. The glomerular filtration rate was calculated: 2/7 （28.5%）of the patients had kidney damage with increased GFR (eGFR> 120 mL/min/1.73 m²), one patient (14.3% of all patients) had a normal function (eGFR ≥ 90 and < 120 mL/min/1.73 m²), and four patients (57.1% of all patients) had a moderate impairment (estimated glomerular filtration rate [eGFR] ≥ 30 and < 60 mL/min/1.73 m²). The mean ± SD age and weight of the subjects were 83.29 ± 10.72 years and 62.71 ± 12.71 kg, respectively. Further details can be found in Table S1, and the treatment course and therapeutic drug monitoring results of nirmatrelvir/ritonavir in plasma and BALF are shown in Figure S1.

3.2 Precision and Accuracy of the Analytic Method

The assays were linear (R = 0.96) over the 0.05–10 mg/mL range for plasma and 0.02–4 mg/mL for the BALF. The intra-day and inter-day quality control samples had a coefficient of variation of 8.1% for all matrixes.

3.3 The ELF and Plasma Concentration of Nirmatrelvir and Correlation Analysis

Detailed information data on urea, albumin, and nirmatrelvir levels in plasma, BALF, and ELF of the seven patients were presented in Table 1. A scatterplot with the associated linear regression line of the ELF versus plasma nirmatrelvir concentrations is shown in Figure 1 and observed good linear relationships (ELF = 0.4976*PLASMA-204; R = 0.96, p < 0.0001).

The ELF: plasma ratios (RELF:P) of nirmatrelvir were calculated using the plasma as reference (C_ELF/C_plasma) for each sample. As shown in Figure 2, the ELF: plasma concentration RELF: P of nirmatrelvir was in the median range of 0.426 ([0.299–0.524], n = 11).

4.1 The Concentrations of NMV in Pulmonary Epithelial Lining Fluid and Plasma

From Figures 1 and 2, the permeability of lung tissue was low (RELF:P ranged from 0.299 to 0.524), but increasing NMV concentrations in the blood will lead to an increase in pulmonary ELF levels. There was a significant relationship (Range: 230–9010 ng/mL) between NMV levels in plasma and those in ELF (ELF = 0.4976*PLASMA-204;R = 0.96, p < 0.0001), with a correlation whose slope (0.4976) suggested that the blood-to-ELF ratio of drug penetration was 2:1.

4.2 The Analysis of ELF Concentrations and Plasma Exposure

According to our linear equations, a negative value indicates that penetration of NMV through blood vessel walls into ELF is required to achieve a baseline concentration. Hence, we can predict that a plasma target concentration of 292 ng/mL corresponds to an ELF concentration that can not achieve the EC₉₀ value. COVID-19 releases inflammatory mediators such as IL-1β, IL-6 and leukotrienes [12], further amplifying inflammation and disrupting the air-blood barrier [13], which can contribute to increased drug permeability. The degree of permeability is correlated with the severity of lung damage. Indeed, some studies have employed the measurement of albumin concentrations in bronchoalveolar lavage fluid as a means to assess the magnitude of alveolar injury [14]. Notably, NMV with low solubility (0.90–1.21 mg/mL throughout the physiological pH range) and with low permeability of NMV in vitro [15], make it challenging for them to cross the cell membrane. This implies that patients with mild-to-moderate have a lower NMV penetration compared with severe patients.

Pharmacodynamics for NMV targeting respiratory pathogens (COVID-19) should be based on free rather than total drug concentrations in ELF. According to the linear equations (ELF = 0.4976*PLASMA-204; R = 0.96, p < 0.0001), ELF values of 181 nM (90 ng/mL) need plasma concentration values of 591 ng/mL (approximately 2 × EC₉₀), and ELF values of 292 ng/mL need plasma concentration values of 996 ng/mL (approximately 3.4 × EC₉₀). In previous studies [9, 16], small amounts of plasma proteins normally reach the alveolar epithelial surface. A value very similar to the average albumin concentration in the ELF was 12% of its level in plasma [9, 16]. Due to the inability to obtain the protein binding rate of NMV in ELF, we can only speculate that the free NMV concentration in the ELF with the EC₉₀ value needed for the corresponding plasma concentration was > 2 × EC₉₀ and ＜ 3.4 × EC₉₀. Interestingly, our hypothesis could explain why the viral clearance rates were higher on the Day 5 nirmatrelvir minimum plasma concentration (> 3×) EC₉₀ plasma concentration in patients enrolled in EPIC-HR [4, 5]—this concentration ensures that the ELF concentration of nirmatrelvir also reaches or above the EC₉₀, and was similar on the Day 5 nirmatrelvir minimum plasma concentration (3–5×) EC₉₀ plasma concentration and (> 5x) EC₉₀ enrolled in EPIC-HR. The hypothesis can also explain why the viral clearance rates were similar on the Day 5 nirmatrelvir minimum plasma concentration (1–3×) EC₉₀ plasma concentration and placebo group enrolled in EPIC-HR. This was because a plasma concentration in this range may not achieve the EC₉₀ in the epithelial lining fluid, leading to a suboptimal antiviral effect and comparable viral clearance rates to the placebo group. Noted, there is no contradiction between our hypothesis and the nirmatrelvir treatment shows significant clinical efficacy. The data from EPIC-HR demonstrated that a significant portion (93.8%) of available C_trough values were > 3 × EC₉₀.

The main limitation of this study was the small sample size, and the inclusion of participants from only a single center, which might restricted outcomes. In the future, increasing the number of data to verify the feasibility and relationship between NMV levels in plasma and those in ELF, may improve clinical outcomes.