Humanigen’s lenzilumab and Kiniksa Pharmaceuticals’ mavrilimumab are best placed to provide the most benefit in Covid-19 patients who need supplementary oxygen, experts said. However, clinical improvements in advanced intubated patients are a potentially dangerous bet, they added.
Choosing trial endpoints based on mitigating the need for respiratory support is significant, even though the initial research was mainly driven by deriving a mortality benefit, experts said. However, experts had mixed views on the ideal period to measure the primary endpoint.
While mavrilimumab is being studied in both ventilated and nonventilated Covid-19 patient groups, lenzilumab is being tested in nonventilated Covid-19 patients alone in a Humanigen-led Phase III study. Lenzilumab is also being tested in the adaptive NIH-led ACTIV-5 study, which includes patients on mechanical ventilation or extracorporeal membrane oxygenation (ECMO).
Lenzilumab and mavrilimumab have the most potential to dampen the cytokine release storm (CRS) observed in Covid-19 at a stage when patients may need supplementary oxygen. In this stage, patients present with a combination of clinical features, like fever and elevated biomarkers. Once the disease has advanced far enough for intubation to be necessary, the impact of such anti-granulocyte-macrophage colony stimulating factor (anti-GM-CSF) antibodies remains unclear, based on early clinical data with both antibodies and other immunomodulators. As such, experts expect efficacy and initial use to be first seen in nonventilated patients, on top of drugs considered to be standard of care (SOC) in that population: Gilead Sciences’ Veklury (remdesivir) and dexamethasone. Moreover, an anti-GM-CSF tactic provides a more targeted effect than the broadly immunosuppressive dexamethasone.
Data from the Phase II portion of the Phase II/III mavrilimumab trial is expected in 1H, while Phase III enrolment has begun, as per the company’s 4Q20 results. Humanigen finished accruing 520 patients in its Phase III trial and expects topline data in March, as per a 29 January press release.
Humanigen will file for an emergency use authorization (EUA), which would allow for an expedited process compared to a BLA, once the data is available, said CEO Cameron Durrant. Since the ACTIV-5 trial is still enrolling, the EUA will not include data from the trial, but there is potential to combine both datasets once the trial is completed, he added.
In addition to Humanigen and Kiniksa, several other companies have competing anti-GM-CSF antibodies, including GlaxoSmithKline , which announced Phase II data with otilimab on 25 February. The data showed the drug did not improve the proportion of patients who were alive and free of respiratory failure for 28 days. Kiniksa did not respond to a request for comment.
Timing is key
Patients who require oxygen support, but who are not yet mechanically ventilated, may have the most potential to benefit, said Dr Kristin Hudock, assistant professor, Pulmonary and Critical Medicine, University of Cincinnati, Ohio. Lenzilumab’s proposed mechanism is to block the effects of T cells, which are induced to produce GM-CSF from activating macrophages with proinflammatory properties, said Dr Andrew Badley, professor, Department of Molecular Medicine, Mayo Clinic, Rochester, Minnesota. Thus, intervening before the inflammation is too severe is appropriate and more likely to make a difference, he added.
Even the otilimab trial included both intubated and nonintubated patients, so the timing of the antibody’s use is key, said Dr Deepa Gotur, associate professor of Clinical Medicine, Academic Institute at Houston Methodist, Texas. GlaxoSmithKline plans to expand a subset of patients who are ages 70 years or older, where an efficacy benefit was seen.
At the start of the pandemic, therapies had to prove a mortality benefit to be considered clinically significant, but now other endpoints are more acceptable, experts said. The lenzilumab study is using ventilator-free survival over a 28-day timeframe as a primary endpoint. Similarly, the mavirilumab study uses the proportion of patients who are alive and without respiratory failure—defined as the need for high-flow oxygen, non-invasive ventilation, invasive mechanical ventilation or ECMO—at 15 days as an endpoint for the same population. Preventing the need for mechanical ventilation and reducing length of hospital stay are valuable goals, said Dr Antonio Abbate, professor of Medicine, Department of Internal Medicine, Virginia Commonwealth University, Richmond. Even parameters looking at the elimination of the need for dialysis or renal replacement therapy are clinically significant because they provide value from a patient and staff and hospital resource perspective, said Hudock.
In a non-randomised cohort study, all 13 patients in the mavrilimumab group and 17 of 26 patients in the control group showed clinical improvement, (p=0·030), as well as an earlier mean time to improvement, with the antibody at eight days compared to 19 days. (De Luca, et al., The Lancet Rheumatology, 2(8), pp. E465-E473, 1 August 2020).
However, experts had mixed views on the timeframes chosen to record the primary endpoint in the lenzilumab and maviriliumab trials. An improvement in ventilator-free survival or hospitalisation rate is still a strong indicator of therapeutic efficacy if it is over two weeks, said Badley. However, longer-term recovery would also be important, he added. However, this timeframe may be too short, and data over 28 days or even 90 days is needed, Abbate said. Even though 28-day data is reasonable, deaths can occur after that time, which needs to be recorded, Abbate explained. Longer-term data is needed because infections can happen after the acute illness, said Hudock.
In a separate cohort trial, clinical improvement—defined as improvement of at least 2 points on the 8-point ordinal clinical endpoints scale—was observed in 11 of 12 (91.7%) patients treated with lenzilumab and 22 of 27 (81.5%) untreated patients, with a shorter median time to improvement of five days (1 – 14 day range) for the lenzilumab arm versus 11 days (4–42 day range) in the untreated cohort (Temesgen, et al., Mayo Clinic Proceedings, 95(11), pp. P2382-2394, 1 November 2020). This data is too small to draw conclusions, said Dr Richard Zuckerman, director, Transplant and Immunocompromised Host Program, Dartmouth-Hitchcock Medical Centre, New Hampshire. Still, Zuckerman noted he is optimistic about lenzilumab’s clinical utility based on the Phase II data and the fact immune modulators are best given to patients before they progress to severe disease. Data from both publications were clinically significant, indicating an early recovery once patients received these antibodies, said Gotur.
Abbate and Hudock did not dismiss the possibility of intubated patients benefiting from anti-GM-CSF antibodies. However, at the point when mechanical ventilation is necessary, too many pathways are activated in advanced disease and the ability to make a difference in mortality with a single inhibitor may be limited, said Abbate. The Phase II/III endpoint for ventilated patients on maviriliumab is mortality.
Factors driving usage
It would be significant if the Phase III results match the interim lenzilumab results, said Badley, adding it would see wide uptake if approved. The lenzilumab study specifically allows SOC. According to interim Phase III results, which were released on 6 November 2020, 78% of trial participants were on Veklury or dexamethasone. The SOC in the mavrilimumab study is not defined on ClincalTrials.gov.
Despite standard treatments like Veklury and dexamethasone, the greatest efficacy gain may be seen in patients who require oxygen and are progressing rapidly, but are not yet intubated, as well as those who may have been intubated within the last 48 hours, said Abbate. Veklury and dexamethasone are now considered part of SOC, which means it would be difficult to study any drug without those as well, said Gotur. While dexamethasone is a broad immunosuppressant, anti-GM-CSF antibodies provide a more targeted approach, Gotur added.
Continued hyperinflammation indicated by persistent fever and failure to improve under oxygenation would also indicate the potential for treatment with anti-GM-CSF antibodies, said Abbate. Biomarkers like C-reactive protein and ferritin give a snapshot of CRS but do not specify when it started, so a combination of clinical status and a biomarker study is essential to understand CRS in Covid-19, said Gotur. Badley noted a combination of clinical indicators like fevers, reduced blood pressure, shortness of breath and the biomarkers as being indicative of when treatment should begin.
While there has been mixed data on other immunomodulators like anti-interleukin-6 (anti-IL-6) antibodies, including Roche’s Actemra (tocilizumab), one of the potential side effects of anti-IL-6 antibodies is neutropenia and potential suppression of T cell response, leading to opportunistic infections, said Abbate. This is also observed with their chronic use in conditions other than Covid-19. While anti-GM-CSF antibodies have not been studied as much, early data has indicated they are safe, he added. Mavrilimumab was found to be well tolerated with no infusion reactions, and no treatment-emergent adverse events were attributed to lenzilumab, as noted in previous data. There is a theoretical risk of secondary infections, but more data beyond what is currently available is needed to draw conclusions, said Hudock. Additionally, anti-GM-CSF antibodies are not associated with any liver or renal toxicity because they do not affect monocytes or neutrophils, which cause those toxicities, said Hudock. If both drug classes are authorised, their adverse event profile, cost and ease of use may drive therapy choice, said Abbate.
Manasi Vaidya and Sean Rai-Roche are Senior Reporter and Reporter for Clinical Trials Arena parent company GlobalData’s investigative journalism team. A version of this article originally appeared on the Insights module of GlobalData’s Pharmaceutical Intelligence Center. To access more articles like this, visit GlobalData.