Introduction

For over 25 years, inhaled Nitric Oxide (iNO) therapy has been used as a rescue treatment to improve arterial oxygenation in Acute Respiratory Distress Syndrome1,2, a major complication of Viral Pneumonia (VP). Nitric Oxide (NO) is a gas produced from arginine in mammalian cells by three NO synthase (NOS) enzymes: neuronal, endothelial, and inducible NO synthase (iNOS)3. Endogenous NO is an endothelium-derived relaxing factor that plays key roles in vascular signalling and blood flow regulation, induces vasodilation, and host defence against various microbial pathogens including bacteria, viruses, fungi, and parasites4,5,6. Following infection or cytokine stimulation, pulmonary iNOS expression is upregulated in macrophages and neutrophils, which play an important protective role against infectious organisms7.

At high doses, NO demonstrates antimicrobial properties against a variety of infectious microorganisms. Pre-clinical and clinical evidence suggests iNO presents favourable clinical applications as a treatment of lung infections, due to multiple therapeutic properties including the potential reduction in microbial load8,9,10,11,12,13.

VP is a serious threat to global health, especially after the outbreak of the coronavirus disease during 2019 (COVID-19) pandemic. Approximately 6 million cases of community-acquired pneumonia occur annually, with over 20% requiring hospitalisation15 and a similar percentage presenting severe-critical pneumonia16. Along with the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), noteworthy causes of VP are respiratory syncytial virus and influenza17,18. VP treatment options are mainly based on symptom relief and supportive care such as oxygen (O2) supplementation and hydration, highlighting the unmet need for the exploration of additional treatment strategies.

Low-dose treatment of iNO (up to 20 ppm), is approved both for newborns with hypoxemic respiratory failure with persistent pulmonary hypertension (US and EU)19,20 and for critically ill adult patients with hypoxemic respiratory failure and post-operative cardiac patients (EU). Nonetheless, higher doses of iNO (160–200 ppm) have been previously claimed to exert a positive clinical effect for patients with viral respiratory infections, including those caused by SARS-CoV-2, thought to be at least partially mediated by the molecule’s antimicrobial properties21,22,23,24,25,26,27,28. This antimicrobial effect of NO at high doses, was demonstrated in-vitro against various types of viruses including influenza A, B, and Coronavirus, at different stages of replication and in a dose-dependent manner12,13,29,30. Specifically, NO inhibited the replication cycle of severe acute respiratory syndrome Coronavirus in-vitro12,13. The in vivo results of iNO murine models point to the balance between exposure time and dose as a key factor in iNO antimicrobial effectiveness24,32. Further research is required to determine the most efficacious iNO treatment dose and schedule against viruses that cause respiratory infections. To date, the literature is lacking adequate controlled clinical trials.

While this trial shares similar attributes with previous clinical trials studying high doses of iNO, it was designed to evaluate the safety of 150-ppm intermittent iNO treatment administered for 40 min, 4 times a day, via a novel NO generator, for clinically diagnosed VP, intended to further explore the optimal regimen under controlled setting (Fig. 1). We hypothesized that high-dose iNO is safe and beneficial in viral-related lower respiratory infections. The present study objective was to evaluate the safety profile of 150-ppm intermittent iNO as a therapy for subjects with VP. The primary safety endpoint was assessed by the incidence of Serious Adverse Events (SAEs) and by monitoring safety parameters during treatment. The secondary efficacy endpoint was measured by improvement in clinical outcome based on (1) number of subjects requiring admission to the Intensive Care Unit (ICU), and (2) number of subjects with unstable saturation, reaching stable room oxygen saturation (SpO2) of ≥ 93% or baseline saturation (whichever is lower) as well as the length of time it took to reach it during hospitalisation. Additional data collected included hospital stay duration, time to fever resolution, time until patient no longer required oxygen, C-Reactive Protein (CRP) and other signs of respiratory distress such as tachypnoea.

Figure 1
figure 1

Study design. Study design included a daily treatment regimen for up to 7 days. Follow-up timepoints: 15, 30, 90, 180 days. R randomisation, NO nitric oxide, iNO inhaled nitric oxide, SST standard supportive treatment.

Full size image

Results

Subject disposition and baseline characteristics

A total of 40 subjects were enrolled, after exclusion of 5 screen failures, and equally randomised between the Treatment-Group (n = 20), receiving iNO treatment in addition to Standard Supportive Treatment (SST), and the Control-Group, receiving SST alone (n = 20). After randomisation, 4 subjects withdrew from the Treatment-Group and one from the Control-Group without completing one inhalation (Fig. 2), while one Treatment-group subject withdrew consent after receiving one inhalation; therefore, the Intention-To-Treat (ITT)-population included 16 subjects in the Treatment-Group and 19 subjects in the Control-Group. Four subjects, 2 from each group, were excluded from the Per-Protocol (PP) efficacy analysis due to major protocol violations (Supplementary Table S4). As the enrolment rate remained low due to COVID pandemic environmental conditions with a parallel reduction of infections by other respiratory viruses, the study terminated early without reaching the recruitment goals.

Figure 2
figure 2

Subject disposition. CONSORT (Consolidated Standards of Reporting Trials) 2010 flow diagram. The Intent-to-treat (ITT) population included 35 subjects enrolled into the study who received at least one iNO (inhaled nitric oxide) treatment. SST standard supportive treatment, PP per protocol.

Full size image

Baseline age was similar between groups, but there was a higher percentage of males in the Control-Group compared to the Treatment-Group (89.5% vs. 56.3%; Table 1). Identification of which baseline characteristics correlate with subsequent time-to-event or efficacy outcomes based on the log rank test, did not identify sex as a significant covariate. Baseline SpO2 in room air presented a slight difference between control and treatment groups, however, this measurement could not be obtained in all subjects due to the requirement for supplemental oxygen and reluctance to destabilise subjects to measure saturation. However, supplemental O2 was similar between groups, indicating a similar clinical subject profile. While the underlying cause for admission was predominantly related to COVID-19 (n = 34) vs. VP caused by another virus (influenza; n = 1), the medical history was similar between groups (Supplementary Table S5); a higher percentage of subjects with vascular disease in the Treatment-Group compared to the Control-Group (50.0% vs. 21.1%; respectively) was observed. Further, the percentage of COVID-19 related medications was similar between groups (Table 1).

Table 1 Subject demographics and baseline characteristics of ITT population (n = 35).
Full size table

Safety endpoint evaluation

Primary safety endpoints were analysed for the ITT-population alone (Treatment-Group: n = 16; Control-Group: n = 19).

Adverse events and serious adverse events

The overall Adverse Events (AEs) rate measured during the study period of 180 days, was similar between groups, with 9 subjects (56.3%) from the Treatment-Group vs. 8 subjects in the Control-Group (42.1%). No AEs were determined to be related to study treatment (Table 2). There were two SAEs in the Treatment-Group, one during the hospitalisation period and another during post-hospitalisation period, both accounted for by underlying pre-existing conditions (Bradycardia deterioration and Triple Vessel disease). A total of 3 subjects experienced AEs that led to treatment discontinuation, two in the NO group (hypoxia and bradycardia) and one (hypoxia) occurring in the control group.

Table 2 Number of ITT subjects experiencing an adverse event by study arm.
Full size table

iNO treatment and associated safety parameters

One-hundred and three iNO inhalations started in 17 subjects of which one withdrew consent after a few minutes of iNO inhalation and thus was not included in the ITT-population. Therefore, 16 subjects were included in ITT-population with 102 iNO inhalations initiated. Two iNO inhalations were paused and not resumed because of discharge and hypoxia occurring due to protocol deviation, thus 100 inhalations were completed.

The potential harmful by-products of iNO treatment were continuously monitored during all inhalations. Nitrogen Dioxide (NO2) levels were measured inherently by the device and Methemoglobin (MetHb) levels were measured using an external non-invasive co-oximeter. Out-of-range NO2 values trigger an automatic pause of NO production by the device and out-of-range MetHb values set off an alarm informing the investigator the inhalation is to be stopped.

NO2 concentration in the breathing circuit remained below the pre-determined safety threshold of 5.0-ppm (Treatment-End Mean 2.3-ppm; SD 0.8; max 4.4), and MetHb levels remained below the safety threshold of 10% (Treatment-End mean 2.1; SD 1.1 max 6.8) (Table 3). Nine inhalations reported out-of-range NO values; while one had low SpO2, the majority of cases resolved within seconds, and all inhalations were completed per protocol, once values returned to the allowed range.

Table 3 NO inhalations safety parameters of ITT population (n = 35).
Full size table

Efficacy endpoint evaluation

Efficacy analyses were performed on both the ITT- and on PP-populations.

Although ICU admissions were defined as an efficacy endpoint, no subjects were admitted to ICU, additionally no mortality was observed.

Stable room SpO2 of ≥ 93%

Significantly more subjects from Treatment-Group reached SpO2 ≥ 93% during hospitalization compared to the Control-Group (Hazard Ratio (HR) = 5.4; 95% Confidence Interval (CI) 1.0, 28.8; p = 0.049) and the median time to reach it was 4 days in the treatment-group, but was not calculable for in the control-group since it was not achieved during hospital stay; (Fig. 3; Supplementary Table S6). Time to reach stable room SpO2 was determined based on subject medical charts. This analysis included ITT subjects with SpO2 < 93% during hospitalization (Treatment-Group: n = 13; Control-Group: n = 15). A similar outcome, although in terms of benefit was noted in the PP-population (Supplementary Table S7).

Figure 3
figure 3

Time to reach stable oxygen saturation of at least 93%. Kaplan–Meier Estimates of time to reach stable room air oxygen saturation (SpO2) of at least 93% among the intent-to-treat (ITT) population. P-value = 0.0490 based on Cox Proportional Hazard Model. NO nitric oxide, SST standard supportive treatment.

Full size image

Oxygen support duration

Oxygen support duration was documented for the full study period including follow-up (up to day 180) and was significantly reduced in the Treatment-Group (median, 3.6 days) compared to the Control-Group (median, 6.3 days) (HR = 2.8; 95% CI 1.1, 7.1; p = 0.0339) (Fig. 4; Supplementary Table S6). This effect was more prominent in the PP-population (HR = 6.3; 95% CI 1.8, 21.8; p = 0.0035); (Supplementary Table S7).

Figure 4
figure 4

Oxygen support duration. Kaplan–Meier Estimates of time to duration of oxygen support among the intent-to-treat (ITT) population. Duration of oxygen support includes the need for oxygen during treatment and at home. The Kaplan–Meier curve was truncated at Day 35 to reflect clinically meaningful assessment of oxygen support. P-value = 0.0339 based on Cox proportional hazard model. NO nitric oxide, SST standard supportive treatment.

Full size image

Hospital stay duration

Median time of 3 days to hospital discharge was similar between the two study groups in both ITT (Supplementary Table S6) and PP (Supplementary Table S7) populations (HR = 1.7; 95% CI 0.7, 4.2; p = 0.2394). Although not significant, beginning on the fourth day of hospitalisation the Kaplan–Meier curves of the two groups separated favouring benefit of iNO (Fig. 5).

Figure 5
figure 5

Hospital stay duration. Kaplan–Meier estimates of time to duration of hospital stay among the intent-to-treat (ITT) population. P-value = 0.2394 based on Cox proportional hazard model. NO nitric oxide, SST standard supportive treatment.

Full size image

Other data collected

In order to further explore high-dose iNO clinical effect, additional parameters were collected during the study. These included signs of respiratory distress such as tachycardia, retractions, tachypnoea, wheezing, and other vital signs such as fever, Blood Pressure (BP) and Respiratory Rate (RR). Sparse data was observed in these outcomes due to an incomplete data collection caused by the difficulties posed by the pandemic-environment.

Discussion

In this open-label randomised study, addition of high-dose intermittent iNO to the SST is indicated to be safe and beneficial for spontaneous breathing hospitalised adults diagnosed with VP.

Safety findings showed high dose iNO treatment administered repeatedly for short intervals was well-tolerated with no treatment-related AEs or SAEs as assessed by the investigators. Overall, AEs rate was high as expected in this type of population33,34, however, rate was comparable between the groups and mostly related to underlying medical conditions. The two reported SAEs during the study were determined to be unrelated to the study treatment and both reported to be related to pre-existing conditions.

Safe treatment with iNO requires close monitoring of potential side-effects and risks. A recognised risk associated with iNO treatment is the inhalation of NO2, which is a by-product of the synthesis between NO and environmental O2; Hence, NO2 concentration predictably increases with the elevation of either NO or O2. Since NO2 can cause tissue injury, the LungFit® PRO device has a NO2 filter to scavenge the gas, and an alarm keeps the level of NO2 in the breathing circuit under the safe threshold of < 5 ppm (CDC threshold35,36. In this study, NO2 levels remained below this safety threshold in all inhalations; 4.4-ppm was the highest level recorded. Monitoring MetHb blood levels is another prerequisite for safe iNO inhalation. NO oxidises haemoglobin converting ferrous (Fe2+) to the ferric (Fe3+) form and, with that, reducing the carrying capacity for O2, which can exacerbate the hypoxia caused by pneumonia. In the present study, MetHb levels were continuously monitored during NO delivery and remained well below 10%, the safety threshold level below which no clinical consequences are expected, and no treatment was required37. MetHb findings are aligned with other published studies of intermittent high-dose iNO treatment for various aetiologies of respiratory morbidities9,11,21,26,27,28. Notably, MetHb measurement with the Masimo co-oximeter device is stated to be unaffected by variation in skin pigmentation. Additional safety measures included continuous monitoring of NO and O2 concentration as well as heart rate and BP, with no related findings.

Efficacy findings indicated that subjects who received iNO benefited from significantly earlier removal of O2 support and had fivefold higher chance to reach stable SpO2 ≥ 93% during hospital stay compared to the Control-Group, as measured by the HR. The duration of hospital stay was not significantly different between groups, but demonstrated a trend in earlier discharge in the treatment group.

Supplemental O2, airway augmentation, cough control, and fluid replacement are among the standard of care approaches for spontaneous breathing hospitalised patients with VP38,39; While Remdesivir demonstrated in vitro and in vivo activity against SARS-CoV-240, no clinical benefits were observed in patients admitted to hospital for COVID-1941. Therefore, supportive or therapeutic options for pneumonia caused by viruses are lacking. With consideration of the impact of the COVID-19 pandemic, the clinical value of iNO as an adjunct therapy for managing VP subjects’ oxygenation status has gained attention. The underlying mechanism of iNO activity by which oxygenation improvement occurs requires further investigation42,43. NO has an array of potential beneficial effects that may explain the efficacy trends in the present study28. NO increases oxygenation through selective pulmonary vasodilation1, induces bronchodilatation, and has an anti-inflammatory effect. At high doses, NO exhibits antimicrobial properties through the formation of highly reactive molecules named reactive nitrogen and oxygen species (RNS and ROS); these in turn may interact with multiple microbial targets leading to overall microbial damage44. In vitro, NO has been shown to inhibit viral replication in multiple viruses including SARS-CoV-245. This effect has been shown to intensify with higher NO concentrations30,31. This specific effect was also studied and demonstrated in vivo46; however, the optimal treatment regimen to prove effectiveness was not found47.

Leveraging the specific antimicrobial effect, a few clinical studies were performed with high iNO doses. Consistent with our results, these found that high-dose iNO treatment (160-ppm) was safe, reduced tachypnoea, and improved oxygenation of hypoxaemic spontaneously breathing high-risk patients with COVID-19 pneumonia22 and further confirmed by a follow-up controlled study20,21.

This study has several limitations to consider due to being conducted during unexpected pandemic conditions. All but one subject of the ITT-population had COVID-19-induced pneumonia, unlike the planned 2:1 ratio of COVID-19 to non-COVID-19 cases. This divergence from the enrolment plan is attributed to both the ominous presentation of COVID-19 and in parallel the marked decrease in non-COVID-19 viral infections during the COVID-19 pandemic observed globally47,49. Though the focus on COVID-19 pneumonia cases may be considered a limitation, the putative mechanisms by which NO inhalation improves common VP symptoms are by increasing oxygenation through pulmonary vasodilation1, and preventing cytokine storm50,51. NO has also been demonstrated to have a viricidal effect on several viruses that are common causes of VP including influenza, RSV, and coronavirus12,13,14. Our study did not address these specific mechanisms of action, but it stands to reason that they are shared by VP of diverse aetiology.

The small sample size posed an additional limitation in the study, which was terminated early due to multiple pandemic-feasibility issues. Among the conditions that affected the overall subject recruitment rate were a constant department rotation, challenging COVID subject recruitment, and a limited population diversity based on an evident worldwide reduction in other respiratory viruses during that period, minimising subject-population of VP due to other viruses52. With this, the enrolment rate decreased and the study ended after only 40 subjects were enrolled, of which 5 subjects were withdrawn without completing the first treatment. Despite the relatively small sample size, some of the efficacy measures achieved statistical significance consistent with improvement in the clinical outcome of VP. Even though the sample size is considered low, the findings align with previous studies demonstrating the safety and efficacy of treating COVID-19 VP subjects with high-dose iNO20,21,22.

As the study was conducted in a demanding COVID-19 pandemic environment, both the design and procedure of the study were affected. It became apparent during the conduct of the trial that many planned assessments that were considered routine, such as respiratory rate measurement and SpO2 measurement in room air, could not be completed due to the COVID-19 environment. This resulted in reduction of protocol requirements for these assessments.

Finally, the lack of blinding should be taken into consideration when interpreting the results. As a pilot trial, this study was not powered to detect differences between study groups and test iNO-therapeutic effects in terms of mortality. Moreover, the findings of this study served as a data-generating trial to inform extensive double-blind clinical trials with a larger sample size.

Pandemic study management exhibited major challenges with constantly changing population characteristics at study entry and hospital patient management practices. This contributed to the limited ability to recruit and to well define the population at study entry.

Conclusions

Intermittent inhalation of high-dose iNO (150-ppm) for up to 7 consecutive days is a safe adjunct therapy for VP subjects. The treatment with high dose iNO may mitigate VP severity and enhance VP subjects’ recovery by reducing time to stable room air SpO2 ≥ 93%, and duration of supportive O2. This study may support further investigation of iNO therapy for spontaneously breathing VP patients as it demonstrates preliminary clinical benefits risk ratio in this population.

Methods

The CONSORT reporting guidelines53 were followed in the present study.

Study population

This was a prospective, open-label, randomised, multi-centre study to assess the safety and efficacy of 150-ppm intermittent iNO for the treatment of hospitalised subjects diagnosed with acute VP. The study was conducted during COVID-19 pandemic thus most patients were diagnosed with COVID-19 related VP. The study performed in 3 medical centres in Israel (Supplementary Table S1), was approved by their Institutional Ethics Committees and by the Israeli Ministry of Health (clinicaltrials.gov registration number NCT04606407; 28/10/2020). The study was conducted in accordance with the declaration of Helsinki and Good Clinical Practice guidelines as well as monitored by Data and Safety Monitoring Board, which assessed at intervals the progress of the trial, the safety data, and the critical efficacy endpoints, provided feedback to the sponsor on safety signals, with the capacity to recommend whether to continue, modify, or stop the trial. Written informed consent was obtained from all participants before screening in the emergency room or upon admission to the hospital.

Inclusion criteria were clinical diagnosis of VP confirmed by x-ray (including positive COVID-19 nasal swab); age 18–80 years; provision of informed consent. Exclusion criteria included pneumonia with ≥ 2 of the following 3: WBC > 15,000 lobar pneumonia or pleural effusion; which could indicate a predominant bacterial infection, though these could not be rolled out completely, the need of high flow nasal cannula; continuous positive airway pressure; immunodeficiency; congestive or unstable heart disease; left ventricular dysfunction (LVEF < 40%); severe pulmonary hypertension and/or unstable systemic hypertension; history of daily continuous oxygen supplementation; Subjects on prolonged (> 2 weeks) systemic steroids, excluding dexamethasone or prednisone, within 30 days prior to enrolment (Supplementary Table S2).

Study design

Subjects who were determined to be eligible for the study were randomised by a precomputed randomisation list based on blocks of 2 (stratified by site) with a 1:1 ratio (Supplementary Methods S1) to either the Treatment-Group receiving intermittent inhalations of 150-ppm iNO plus SST, or to the Control-Group receiving SST. The Treatment-Group subjects received intermittent inhalations of 150 (± 10%) ppm iNO for 40 min, 4-times per day (24-h) every 4.5-h (± 30 min) up to 7 days, which is not considered standard treatment. The inhalation schedule was halted at least 7 h during the night to ensure subjects’ well-being and adequate sleep.

During hospitalisation, the following assessments were performed—vital signs twice a day (i.e., heart rate, BP, respiratory rate (RR), body temperature, %SpO2 in room air); physical examination once a day (Supplementary Table S3); blood and urinalysis parameters at screening and at the end of treatment course. The physical examination included monitoring of signs indicative for disease severity such as fever, tachycardia, retractions, tachypnea, wheezing, and CRP as indication of inflammation. The optimal SpO2 in adults with COVID-19 is unknown. However, according to the National Institutes of Health’s COVID-19 Treatment Guidelines (NIH, USA), a target between 92 and 96% is considered reasonable based on evidence that an SpO2 < 92% or > 96% may be harmful for patients without COVID-1954. Further evidence suggests saturation < 92% is associated with major AEs among patients with pneumonia55.

Prior to-, during, and at the end of each NO administration %MetHb, %SpO2 in room air, and heart rate were monitored via a non-invasive method utilising a commercially available pulse co-oximeter (Masimo, Irvine CA, model Rad-87). NO, NO2, and O2 levels were continuously monitored during inhalations as well by the Lungfit® PRO device. RR was measured prior to and at the end of each treatment and BP was taken at the end of each inhalation. Subjects were followed from enrolment day for 180 days including post-hospitalisation (Fig. 1).

NO and SST treatments

The NO gas flow was generated, delivered, and monitored by the LungFit® PRO device (Beyond Air Inc., Garden City, NY) at a flow rate of 15 L/min through a breathing circuit and CPAP facemask to spontaneously breathing subject. The novel LungFit® PRO is a portable device that generates NO from O2 and N2 in ambient air, at the point-of-care. Then, the NO gas is passed through a NO2-filter that removes NO2, a noxious byproduct of the generation of NO that increases with increased concentrations of O2. The replaceable filter electronically communicates with the LungFit® PRO and has the additional functions of setting the dose (150-ppm) to be generated by the device, tracking the NO delivered, and preventing a filter from being reused. Hence, one filter per treatment is required. The device continuously monitors NO, NO2, and O2 levels and alerts when out-of-range during administration of NO. The device largely shares the technologies of a device to administer continuous NO therapy for neonates with persistent pulmonary hypertension recently granted FDA approval for infants19 while the use for pulmonary hypertension in adults is currently off-label.

During inhalation, supplemental O2 can be added to the gas flow outlet of the device by delivering O2 from an external O2 source via a flowmeter and designated O2 tubing to the O2 inlet on the back panel of the device. The supplemental O2/air flows through the device via a separate gas pathway become part of the diluting gas flow, and hence contribute to generation of NO2. Safety parameters were continuously monitored during each iNO inhalation as follows: MetHb levels (< 3% at pre-inhalation, < 10% during inhalation), mean peripheral SpO2 (SpO2, safety threshold ≥ 89%) and heart rate percutaneously by pulse co-oximeter, NO concentration (150-ppm ± 10%), NO2 concentration (< 5 ppm) and O2 concentration sampled from the breathing circuit just prior to delivery to the patients and analysed by the LungFit® PRO device via a sample inlet port on the front panel of the device.

SST included: O2 supplementation, antivirals including remdesivir, inhaled bronchodilators, corticosteroids such as dexamethasone, antipyretics, antibiotics, and other concomitant medications, administered as prescribed by the treating physician according to routine facility practices and per physician’s discretion.

Analysis

Safety and efficacy analyses were conducted on the ITT population, which included all subjects who received at least one treatment and on the PP population which included ITT-population without major violation.

The primary safety endpoint was assessed by incidence of SAEs. The secondary efficacy endpoint was evaluated by measuring the number of subjects: (i) deteriorating to ICU, (ii) reaching stable room oxygen saturation (SpO2) ≥ 93% or baseline saturation, whichever is lower as well as the length of time it took to reach it during hospitalisation. The determination of stable saturation was based on three conditions to be achieved during hospitalisation: (1) the subject reached SpO2 ≥ 93%, (2) no lower values were recorded afterward, and (3) the subject was not supplemented with oxygen again. Subjects who were discharged without reaching stable saturation were censored.

Other data that were collected included: O2 support and hospital stay durations, MetHb, NO2 and SpO2 levels.

Following the methods outlined in the statistical analysis plan, the first step in the analysis focused on identification of which baseline characteristics correlate with subsequent time to reach room air SpO2 ≥ 93%. This examination identified one covariate (i.e., age). For time-to-event analysis, Kaplan–Meier curves were used to graphically describe the endpoints, and Cox-proportional hazards regression models calculated HR with its 95% CI with adjustment for gender and age group (≤ 45 years, > 45 to ≤ 54 years, and > 54 years). Changes from baseline in continuous outcomes were analysed by fitting fixed effects linear models or non-parametric models to the data. The Cox proportional hazard model was utilized for time-to-event outcomes with terms for treatment and age as a covariate, and analysis of variance (ANOVA) was used for change from baseline in continuous outcomes with term for treatment. Robustness analysis of the efficacy endpoints included an analysis on the PP-population. Descriptive analyses were used for the safety endpoints. This was a pilot study and was not powered to demonstrate statistical significance of either safety or efficacy outcomes. Statistical analyses were performed by SAS version 9.4 (SAS Institute, Cary, North Carolina; URL: https://www.sas.com/).