Nebulised dornase alfa reduces inflammation and improves clinical outcomes in severe COVID-19: a randomised clinical trial

Joint Public Review:

In this study by Porter et al reports on outcomes from a small, open-label, pilot randomized clinical trial comparing dornase-alfa to the best available care in patients hospitalized with COVID-19 pneumonia. As the number of randomized participants is small, investigators describe also a contemporary cohort of controls and the study concludes about decrease of inflammation (reflected by CRP levels) after 7 days of treatment but no other statistically significant clinical benefit.

Suggestions to the authors:

• Please re-analyze findings by omitting from all Tables and Figures all data of comparators who were not randomized (BAC). I understand the difficulties of running this trial but the results of excess reduction of mortality do not allow the publication of a trial where comparators do not come from the randomized patient population.
• The presentation remains confusing and the manuscript should be critically revised for clarity. There is a repetition of methods (e.g. lines 176-187 repeat 160-175) and redundant results (e.g. Figure S2, Table 3). At Table 4: the authors should select one method of illustration for lab results, either Table or figure, without repetitions
• Regarding inclusion criteria, it is unclear whether high radiological suspicion is sufficient for inclusion or whether PCR based confirmation is required in all instances (differences in wording between lines 153 and 191), and under which oxygen requirements (lines 155 and 192)
• Table 1 should be merged with Table S2 and a better description of cohort baseline severity (P/F, SOFA, APACHE, organ support, number of patients in each point of the WHO severity score) and treatments should be made available

Author Response

The following is the authors’ response to the original reviews.

Reviewer 1:

Public review:

In this study, Porter et al report on outcomes from a small, open-label, pilot randomized clinical trial comparing dornase-alfa to the best available care in patients hospitalized with COVID-19 pneumonia. As the number of randomized participants is small, investigators describe also a contemporary cohort of controls and the study concludes about a decrease of inflammation (reflected by CRP levels) aJer 7 days of treatment but no other statistically significant clinical benefit.

Suggestions to the authors:

• The RCT does not follow CONSORT statement and reporting guidelines

We thank you for this suggestion and have now amended the order and content of the manuscript to follow the CONSORT statement as closely as possible.

• The authors have chosen a primary outcome that cannot be at least considered as clinically relevant or interesting. AJer 3 years of the pandemic with so much research, why investigate if a drug reduces CRP levels as we already have marketed drugs that provide beneficial clinical outcomes such as dexamethasone, anakinra, tocilizumab and baricitinib.

We thank the reviewer for bringing up this central topic. The answer to this question has both a historical and practical component. This trial was initiated in June of 2020 and was completed in June of 2021. At that time there were no known treatments for the severe immune pathology of COVID19 pneumonia. In June 2020, dexamethasone data came out and we incorporated dexamethasone into the study design. It took much longer for all other anti-inflammatories to be tested. Hence, our decision to trial an approved endonuclease was based purely on basic science work on the pathogenic role of cell-free chromatin and NETs in murine sepsis and flu models and the ability of DNase I to clear them and reduce pathology in these animal models. In addition, evidence for the presence of cell-free chromatin components in COVID-19 patient plasma had already been communicated in a pre-print. Finally, several studies had reported the anti-inflammatory effects of dornase treatment in CF patients. Hence there was a strong case for a cheap, safe, pulmonary noninvasive treatment that could be self-administered outside the clinical se]ng.

The Identification of novel/repurposed treatments effective for COVID-19 were hampered by patient recruitment to competing studies during a pandemic. This resulted in small studies with inconclusive or contrary findings. In general, effective treatments were only picked up in very large RCTs. For example, demonstrating dexamethasone as effective in COVID-19 required recruitment of 6,425 patients into the RECOVERY study. Multiple trials with anti-IL-6 gave conflicting evidence until RECOVERY recruited 4116 adults with COVID-19 (n=2022, tocilizumab and 2094, control) similar for Baracitinib (4,148 randomised to treatment and 4,008 to standard care). Anakinra is approved for patients with elevated suPAR, based on data from one randomized clinical trial of 594 patients, of whom 405 had active treatment (PMID: 34625750). However, a systematic review analysing over 1,627 patients (anakinra 888, control 739) with COVID-19 showed no benefit (PMID: 36841793). Regarding the choice of the primary endpoint, there is a wealth of clinical evidence to support the relevance of CRP as a prognostic marker for COVID-19 pneumonia patients and it is a standard diagnostic and prognostic clinical parameter in infectious disease wards. This choice in March 2020 was supported by evidence of the prognostic value of IL-6; CRP is a surrogate of IL-6. We also provide our own data from a large study of severe COVID-19 pneumonia in figure 1, showing how well CRP correlates with survival.

In summary, our data suggest that Dornase yields an anti-inflammatory effect that is comparable or potentially superior to cytokine-blocking monotherapies at a fraction of the cost and potentially without the additional adverse effects such as the increase for co-infections.

We now provide additional justification on these points in the introduction on pg.4 as follows:

“The trial was ini.ated in June 2020 and was completed in September of 2021. At the start of the trial only dexamethasone had been proven to benefit hospitalized COVID-19 pneumonia pa.ents and was thus included in both arms of the trial. To increase the chance of reaching significance under challenging constraints in pa.ent access, we opted to increase our sample size by using a combina.on of randomized individuals and available CRP data from matched contemporary controls (CC) hospitalized at UCL but not recruited to a trial. These approaches demonstrated that when combined with dexamethasone, nebulized DNase treatment was an effec.ve an.-inflammatory treatment in randomized individuals with or without the implementa.on of CC data.”

We also added the following explanation in the discussion on pg. 16:

“Our study design offered a solution to the early screening of compounds for inclusion in larger platform trials. The study took advantage of frequent repeated measures of quantifiable CRP in each patient, to allow a smaller sample size to determine efficacy/futility than if powered on clinical outcomes. We applied a CRP-based approach that was similar to the CATALYST and ATTRACT studies. CATALYST showed in much smaller groups (usual care, 54, namilumab, 57 and infliximab, 35) that namilumab that is an antibody that blocks the cytokine GM-CSF reduced CRP even in participants treated with dexamethasone whereas infliximab that targets TNF-α had no significant effect on CRP. This led to a suggestion that namilumab should be considered as an agent to be prioritised for further investigation in the RECOVERY trial. A direct comparison of our results with CATALYST is difficult due to the different nature of the modelling employed in the two studies. However, in general Dornase alfa exhibited comparable significance in the reduction in CRP compared to standard of care as described for namilumab at a fraction of the cost. Furthermore, endonuclease therapies may prove superior to cytokine blocking monotherapies, as they are unlikely to increase the risk for microbial co-infections that have been reported for antibody therapies that neutralize cytokines that are critical for immune defence such as IL-1β, IL-6 or GM-CSF. “

• Please provide in Methods the timeframe for the investigation of the primary endpoint

This information is provided in the analysis on pg. 8:

“The primary outcome was the least square (LS) mean CRP up to 7 days or at hospital discharge whichever was sooner.”

• Why day 35 was chosen for the read-out of the endpointt?

We now state on pg. 8 that “Day 35 was chosen as being likely to include most early mortality due to COVID-19 being 4 weeks after completion of a week of treatment. ( i.e. d7 of treatment +28 (4 x 7 days))”

• The authors performed an RCT but in parallel chose to compare also controls. They should explain their rationale as this is not usual. I am not very enthusiastic to see mixed results like Figures 2c and 2d.

We initially aimed at a fully randomized trial. However, the swiJ implementation of trial prioritization strategies towards large and pre-established trial plamorms in the UK made the recruitment COVID19 patients to small studies extremely challenging. Thus, we struggled to gain access to patients. Our power calculations suggested that a mixed trial with randomized and contemporary controls was the best way forward under these restrictions in patient access that could provide sufficient power.

That being said, we also provide the primary endpoint (CRP) results in Fig. 3B as well as the results for the length of hospitalization (Fig. S3D) for the randomized subjects only.

• Analysis is performed in mITT; this is a major limitation. The authors should provide at least ITT results. And they should describe in the main manuscript why they chose mITT analysis.

We apologize if this point was confusing. We performed the analysis on the ITT as defined in our SAP: “The primary analysis population will be all evaluable patients randomised to BAC + dornase alfa or BAC only who have at least one post-baseline CRP measurement, as well as matched historical comparators.”

We understand that the reason this might be mistaken as an mITT is because the N in the ITT (39) doesn’t match the number randomised and because we had stated on pg. 8 that “ Efficacy assessments of primary and secondary outcomes in the modified inten.on-to-treat popula.on were performed.”

However, we did randomise 41 participants, but:

One participant in the DA arm never received treatment. The individual withdrew consent and was replaced. We also have no CRP data for this participant in the database, so they were unevaluable, and we couldn’t include them in the baseline table even if we wanted to. In addition, 1 participant in BAC only had a baseline CRP measurement available. Hence not evaluable as we only have a baseline CRP measurement for this participant.

We have corrected the confusing statement on pg. 8 and added an additional explanation.

“Efficacy assessments of primary and secondary outcomes in the inten.on-to-treat (ITT) popula.on were performed on all randomised par.cipants who had received at least one dose of dornase alfa if randomized to treatment. For full details see Sta.s.cal Analysis Plan. The ITT was adjusted to mi.gate the following protocol viola.ons where one par.cipant in the BAC arm and one in the DA arm withdrew before they received treatment and provided only a baseline CRP measurement available. The par.cipant in the DA arm was replaced with an addi.onal recruited pa.ent. Exploratory endpoints were only available in randomised par.cipants and not in the CC. In this case, a post hoc within group analysis was conducted to compare baseline and post-baseline measurements.”

• It is also not usual to exclude patients from analysis because investigators just do not have serial measurements. This is lost to follow up and investigators should have pre-decided what to do with lost-to-follow-up.

Our protocol pre-specified that the primary analysis population should have at least one postbaseline CRP measurement (pg. 13 of protocol). The patient that was excluded was one that initially joined the trial but withdrew consent after the first treatment but before the first post-treatment blood sample could be drawn. Hence, the pre-treatment CRP of this patient alone provided no useful information.

• In Table 1 I would like to see all randomized patients (n=39), which is missing. There are also baseline characteristics that are missing, like which other treatments as BAT received by those patients except for dexamethasone.

Table 1 includes all 39 patients plus 60 CCs.
Table 2 shows additional treatments given for COVID-19 as part of BAC.

• In the first paragraph of clinical outcomes, the authors refer to a cohort that is not previously introduced in the manuscript. This is confusing. And I do not understand why this analysis is performed in the context of this RCT although I understand its pilot nature.

One of the main criticisms we have encountered in this study has been the choice of the primary endpoint. The best way respond to these questions was to provide data to support the prognostic relevance of CRP in COVID-19 pneumonia from a separate independent study where no other treatments such as dexamethasone, anakinra or anti-IL6 therapies were administered. We think this is very useful analysis and provides essential context for the trial and the choice of the primary endpoint, indicating that CRP has good enough resolution to predict clinical outcomes.

• Propensity-score selected contemporary controls may introduce bias in favor of the primary study analysis, since controls are already adjusted for age, sex and comorbidities.

The contemporary controls were selected to best match the characteristics of the randomized patients including that the first CRP measurement upon admission surpassed the trial threshold, so we do not see how this selection process introduces biases, as it was blinded with regards to the course of the CRP measurements. Given that this was a small trial, matching for baseline characteristics is necessary to minimize confounding effects.

• The authors do not clearly present numerically survivors and non-survivors at day 34, even though this is one of the main secondary outcomes.

We now provide the mortality numbers in the following paragraph on pg. 13.

“Over 35 days follow up, 1 person in the BAC + dornase-alfa group died, compared to 8 in the BAC group. The hazard ra.o observed in the Cox propor.onal hazards model (95% CI) was 0.47 (0.06, 3.86), which es.mates that throughout 35 days follow-up, there was a 53% reduced chance of death at any given .mepoint in the BAC + dornase-alfa group compared to the BAC group, though the confidence intervals are wide due to a small number of events. The p-value from a log-rank test was 0.460, which does not reach sta.s.cal significance at an alpha of 0.05.”

• It is unclear why another cohort (Berlin) was used to associate CRP with mortality. CRP association with mortality should (also) be performed within the current study.

As we explained above, the Berlin cohort CRP data serve to substantiate the relevance of CRP as a primary endpoint in a cohort that experienced sufficient mortality as this cohort did not receive any approved anti-inflammatory therapy. Mortality in our COVASE trial was minimal, since all patients were on dexamethasone and did not reach the highest severity grade, since we opted to treat patients before they deteriorated further. The overall mortality was 8% across all arms of our study, which does not provide enough events for mortality measurements. In contrast the Berlin cohort did not receive dexamethasone and all patients had reached a WHO severity grade 7 category with mortality at 30%.

My other concerns are:

• This report is about an RCT and the authors should follow the CONSORT reporting guidelines. Please amend the manuscript and Figure 1b accordingly and provide a CONSORT checklist.

We now provide a CONSORT checklist and have amended the CONSORT diagram accordingly.

• Please provide in brief the exclusion criteria in the main manuscript

We have now included the exclusion criteria in the manuscript on pg. 6.

“1.1.1 Exclusion criteria

1. Females who are pregnant, planning pregnancy or breasmeeding

2. Concurrent and/or recent involvement in other research or use of another experimental inves.ga.onal medicinal product that is likely to interfere with the study medica.on within (specify .me period e.g. last 3 months) of study enrolment 3. Serious condi.on mee.ng one of the following:

a. Respiratory distress with respiratory rate >=40 breaths/min

b. oxygen satura.on<=93% on high-flow oxygen

1. Require mechanical invasive or non-invasive ven.la.on at screening

2. Concurrent severe respiratory disease such as asthma, COPD and/or ILD

3. Any major disorder that in the opinion of the Inves.gator would interfere with the evalua.on of the results or cons.tute a health risk for the trial par.cipant

4. Terminal disease and life expectancy <12 months without COVID-19

5. Known allergies to dornase alfa and excipients

6. Par.cipants who are unable to inhale or exhale orally throughout the en.re nebulisa.on period So briefly Pa.ents were excluded if they were:

7. pregnant, planning pregnancy or breasmeeding 2.     Serious condi.on mee.ng one of the following:  
   

a. Respiratory distress with respiratory rate >=40 breaths/min

b. oxygen satura.on<=93% on high-flow oxygen

1. Require ven.la.on at screening

2. Concurrent severe respiratory disease such as asthma, COPD and/or ILD

3. Terminal disease and life expectancy <12 months without COVID-19

4. Known allergies to dornase alfa and excipients

5. Par.cipants who are unable to inhale or exhale orally throughout the en.re nebulisa.on period”

• "The final trial visit occurred at day 35." "Analysis included mortality at day 35". I am not sure I understand why. In clinicaltrials.gov all endpoints are meant to be studies at day 7 except for mortality rate day 28. Why day 35 was chosen? Please be consistent.

Thank you for identifying this inconsistency. We have amended the record on clinicaltrials.gov to read ‘’the time to event data was censored at 28 days post last dose (up to d35) for the randomised participants and at the date of the last electronic record for the CC.”

• Please provide in Methods the timeframe for the investigation of the primary endpoint

• The authors performed an RCT but in parallel chose to compare also controls. They should explain their rationale as this is not usual. I am not very enthusiastic to see mixed results like Figures 2c and 2d.

• Analysis is performed in mITT; this is a major limitation. The authors should provide at least ITT results. And they should describe in the main manuscript why they chose mITT analysis.

• It is also not usual to exclude patients from analysis because investigators just do not have serial measurements. This is lost to follow up and investigators should have pre-decided what to do with lost-to-follow-up.

• Figure 1b as in CONSORT statement, please provide reasons why screened patients were not enrolled.

• In Table 1 I would like to see all randomized patients (n=39), which is missing. There are also baseline characteristics that are missing, like which other treatment as BAT received those patients except for dexamethasone.

• In the first paragraph of clinical outcomes, the authors refer to a cohort that is not previously introduced in the manuscript. This is confusing. And I do not understand why this analysis is performed in the context of this RCT although I understand its pilot nature.

• In Figure 2 the authors draw results about ITT although in methods describe that they performed an mITT analysis. Please be consistent.

Please see answers provided to these queries above.

Reviewer #2 (Recommendations For The Authors):

1. Suppl Figure 2B would be more informative if presented as a Table with N of patients with per day sampling

We now provide the primary end point daily sampling table in Table 3.

1. The numbers at risk should figure under the KM curves

The numbers at risk for figures 1E, 2C, 2D have been added as graphs either in the main figures or in the supplement.

1. HD in Supplementary figure 3 should be explained

We apologize for this omission. We now provide a description for the healthy donor samples that we used in the cell-free DNA measurements in figure S3B on pg. 14:

“Compared to the plasma of anonymized healthy donors volunteers at the Francis Crick ins.tute (HD), plasma cf-DNA levels were elevated in both BAC and DA-treated COVASE par.cipants.

1. Presentation is inappropriate for Table S4

We thank the reviewer for pointing this issue. We have now formaxed Table S4 to be consistent with all other tables.