key: cord-0913352-qyayhmqk
authors: Kilcoyne, A.; Jordan, E.; Thomas, K.; Pepper, A. N.; Zhou, A.; Chappell, D.; Amarapala, M.; Theriault, R.-K.; Thompson, M.
title: Clinical and economic benefits of lenzilumab plus standard of care compared with standard of care alone for the treatment of hospitalized patients with Coronavirus Disease 19 (COVID-19) from the perspective of National Health Service England
date: 2022-02-15
journal: nan
DOI: 10.1101/2022.02.11.22270859
sha: 5a90f845589c3205bedd1ce972d776b36097e485
doc_id: 913352
cord_uid: qyayhmqk

Purpose: Estimate the clinical and economic benefits of lenzilumab plus standard of care (SOC) compared with SOC alone in the treatment of hospitalized COVID-19 patients from the National Health Service (NHS) England perspective. Methods: A cost calculator was developed to estimate the clinical benefits and costs of adding lenzilumab to SOC in newly hospitalized COVID-19 patients over 28 days. The LIVE-AIR trial results informed the clinical inputs: failure to achieve survival without ventilation (SWOV), mortality, time to recovery, intensive care unit (ICU) admission, and invasive mechanical ventilation (IMV) use. Base case costs included drug acquisition and administration for lenzilumab and remdesivir and hospital resource costs based on level of care required. Clinical and economic benefits per weekly cohort of newly hospitalized patients were also estimated. Results: In all populations examined, specified clinical outcomes were improved with lenzilumab plus SOC over SOC treatment alone. In a base case population aged <85 years with C-reactive protein (CRP) <150 mg/L, with or without remdesivir, adding lenzilumab to SOC was estimated to result in per patient cost savings of 1,162 British Pounds (GBP). In a weekly cohort of 4,754 newly hospitalized patients, addition of lenzilumab to SOC could result in 599 IMV uses avoided, 352 additional lives saved, and over 5.5 million GBP in cost savings. Scenario results for per-patient cost savings included: 1) aged <85 years, CRP <150 mg/L, and receiving remdesivir (3,127 GBP); 2) Black patients with CRP <150 mg/L (9,977 GBP); and 3) Black patients from the full population (2,369 GBP). Conversely, in the full mITT population, results estimated additional cost of 4,005 GBP per patient. Conclusion: Findings support clinical benefits for SWOV, mortality, time to recovery, time in ICU, time on IMV, and ventilator use, and an economic benefit from the NHS England perspective when adding lenzilumab to SOC for hospitalized COVID-19 patients.

4 patients with COVID-19, especially those requiring IMV, have more detrimental health outcomes 79 and a higher economic burden, safe and effective treatment options that provide good value for 80 money represent a significant unmet need. 17-19 81 An increased risk of IMV and death in COVID-19 patients has been linked to a process 82 known as immune-mediated hyperinflammation. 20 This process is largely mediated by 83 granulocyte-macrophage colony-stimulating factor (GM-CSF), a protein that is produced locally in 84 inflamed tissue in response to 22 Activated GM-CSF functions to activate 85 and mobilize myeloid cells, resulting in the excessive production of downstream pro-inflammatory 86 cytokines, including interleukin-6 (IL-6). 21,23,24 C-reactive protein (CRP) is an acute phase reactant 87 produced by the liver largely in response to IL-6, and CRP levels serve as a reliable surrogate for 88 IL-6 bioactivity. 25, 26 Lung tissue injury, disease severity, and ICU admission of patients with 89 COVID-19 have been directly correlated with both GM-CSF levels [27] [28] [29] [30] and CRP levels at the time 90 of hospital admission. [31] [32] [33] [34] [35] [36] [37] Further, neutralization of GM-CSF is expected to prevent or reduce the 91 severity and sequelae of immune-mediated hyperinflammation. 24 As such, GM-CSF has been 92 identified as an important therapeutic target of Lenzilumab is a novel Humaneered ® anti-human GM-CSF monoclonal antibody that 94 directly binds GM-CSF, thereby preventing its downstream signaling. 23 This mechanism of action 95 of lenzilumab renders it a promising new treatment option as its efficacy is likely unaffected by 96 new COVID-19 variants. 38 LIVE-AIR (NCT04351152) was a Phase 3, prospective, randomized, 97 multicentered, double-blind, placebo-controlled clinical trial that evaluated the efficacy and safety 98 of lenzilumab or placebo (in addition to institutional standard of care [SOC] in both treatment 99 groups) in hospitalized participants with COVID-19 pneumonia. 23 Patients must have been 100 hospitalized with an oxygen saturation (SpO2) ≤94% on room air or required supplemental 101 oxygen but had not progressed to IMV. 23 Results from the LIVE-AIR trial showed that lenzilumab 102 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted February 15, 2022. ; https://doi.org/10.1101 https://doi.org/10. /2022 5 plus SOC significantly improved the likelihood of achieving survival without ventilation (SWOV) a 103 (sometimes referred to as ventilator-free survival) by day 28 compared with placebo plus SOC. 23 104 Additional analyses revealed that lenzilumab plus SOC had greater efficacy in patients aged <85 105 years with CRP levels <150 mg/L, overall and exclusively in patients receiving remdesivir, 106 compared with placebo plus SOC, as measured by SWOV. 39, 40 As previously mentioned, CRP 107 levels are elevated in the process of immune-mediated hyperinflammation and have been linked 108 to poor clinical outcomes among patients with severe 34, 35, 40 Therefore, evidence 109 from the LIVE-AIR trial suggest that lenzilumab may have greater efficacy when administered as 110 an early intervention in hospitalized patients with Results from an exploratory analysis of the LIVE-AIR trial also suggested that Black and 112 African American patients with baseline level CRP <150 mg/L may exhibit the greatest response 113 to lenzilumab, with a nearly 9-fold increase in the likelihood of SWOV. 41 This is notable as the 114

Office of National Statistics (ONS) reported that the COVID-19 death rate was higher for Black 115 African (2.16 greater for males and 1.62 greater for females) and Black Caribbean (1.69 greater 116 for males and 1.35 greater for females) populations in the UK compared to their White British 117 counterparts, during the second wave of the COVID-19 pandemic (September 12, 2020 to March 118 31, 2021 . 42 The hyper-vulnerability in this population can be largely attributed to low vaccination 119 rates, as a recent UK-wide survey identified Black or Black British respondents as having the 120 highest rate of vaccine hesitancy (71.8%). 43 Further, the Black population, on average, has a 121 higher incidence and prevalence of chronic illness (eg, diabetes, hypertension, obesity) and lower 122 socioeconomic status, both of which result in higher risk of infection, hospitalization, and 123 death. 43,44 124 a SWOV is a robust composite endpoint used in many of the recent COVID-19 studies that is less prone to favor treatments with discordant effects on survival and days free of ventilation while avoiding the need for sample sizes approaching those of mortality trials to enable timely availability of study results.

. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted February 15, 2022. ; https://doi.org/10.1101 https://doi.org/10. /2022 6 Although clinical evidence supports that lenzilumab effectively improves the likelihood of 125 SWOV in hospitalized patients with COVID-19, 23,40 its overall clinical and economic value from 126 the National Health Service (NHS) England perspective has not been previously characterized. 127

The objective of this study was to report the clinical benefits of the LIVE-AIR trial in a meaningful 128 way to the NHS England perspective, and to estimate the per-patient and population-level costs 129 using the clinical trial data to demonstrate the economic benefit of lenzilumab plus SOC 130 compared with SOC alone in the treatment of hospitalized COVID-19 patients. This analysis was 131 conducted ex ante (ie, before regulatory approval of lenzilumab). 132

Calculator Structure 134 We previously developed a Microsoft ® Excel-based cost calculator to estimate the clinical benefits 135 and per-patient costs of lenzilumab plus SOC versus SOC alone for newly hospitalized patients 136 with COVID-19 pneumonia from a United States (US) hospital perspective. 45 This calculator was 137 adapted to represent a UK healthcare payer perspective, specifically that of NHS England. A 138 28-day model time horizon for the index COVID-19 hospitalization was selected to align with data 139 censoring for the LIVE-AIR trial (28 days following enrollment). 23 A patient-level analysis was 140 conducted to estimate the clinical benefits and per-patient costs for an average of newly 141 hospitalized patients with COVID-19 pneumonia in England. A cohort-level analysis was also 142 conducted considering all newly hospitalized patients in a 1-week period. The calculator structure 143 is presented in Figure 1 . 144

Patient eligibility for the calculator was based on inclusion criteria for the LIVE-AIR trial. In brief, 146 eligible patients were newly hospitalized with COVID-19 pneumonia, with SpO2 ≤94% on room air 147 and/or requiring supplemental oxygen, but not on IMV. 23 Further analyses of LIVE-AIR trial data 148 suggested that patients aged <85 years with CRP <150 mg/L particularly benefited from 149 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted February 15, 2022. ; https://doi.org/10.1101 https://doi.org/10. /2022 lenzilumab, and this population was selected for the base case. 39, 40 Since the ongoing and fully 150 enrolled ACTIV-5/BET-B trial, which aims to further elucidate the efficacy of lenzilumab in 151 COVID-19 in patients aged <85 years with CRP <150 mg/L and receiving remdesivir, this 152 population was explored in a scenario analysis (scenario #1). 46,47 This population receiving 153 remdesivir was not selected for the base case given the lower use of remdesivir in the UK 154 (16.7% 48 to 48.4% 49 ) compared with that observed in the LIVE-AIR trial (72.4%) conducted 155 primarily in the US. 23 Scenario analyses were also conducted in the full LIVE-AIR modified 156 intent-to-treat (mITT) population b (scenario #2), in Black patients with CRP <150 mg/L (with or 157 without remdesivir) (scenario #3), and in Black patients from the full mITT population (scenario 158 #4). Data for the mITT population of the LIVE-AIR trial, the pre-specified primary analysis 159 population, were used for all analyses. 50 160

Weekly cohort sizes for the base case and scenario analyses were calculated using recent case 162 count reports from UK government sources, the LIVE-AIR trial, and a physician survey. The 163 physician survey was conducted between August 13 to 25, 2021 with 82 medically qualified and 164 practicing pulmonary medicine physicians, infectious disease specialists, and intensive care 165 specialists in the UK. 49 All physicians were consultants and senior non-consultant hospital doctors 166 and all had treated patients with COVID-19. 49 In a 1-week period from January 17 to January 23, 167 2022, there were an estimated 11,343 newly hospitalized COVID-19 patients in England. 13 Based 168 on the physician survey, 60.1% of hospitalized patients have SpO2 less than 94%, 49 and would be 169 eligible for treatment in the base case and scenario analyses. To calculate all subgroups with 170 CRP <150 mg/L (base case, scenario #1 and #3), it was assumed that 77.9% of the population 171 b The mITT population was the analysis set used for the primary analysis of efficacy, defined as all randomized subjects who received at least one dose of study drug under the documented supervision of the principal investigator or sub-investigator and excluding sites that experienced documented limitations to access of basic supportive care for COVID-19.

. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted February 15, 2022. ; https://doi.org/10.1101/2022.02.11.22270859 doi: medRxiv preprint 8 with SpO2 <94% had CRP <150 mg/L, as observed in the LIVE-AIR trial population. 40 For 172 subgroups restricted to age <85 years (base case, scenario #1), it was assumed that 89.5% of 173 the hospitalized population was <85 years according to NHS data for hospital admissions in 174 2020-2021. 51 The subgroup receiving remdesivir (scenario #1) was calculated using data from the 175 physician survey for the percentage of patients with CRP <150 mg/L who are treated with 176 remdesivir, which was estimated at 37.7%. 49 Finally, for the scenarios in the Black population 177 (scenario #3 and #4) it was assumed that 3. 

Using the treatment efficacy data from the LIVE-AIR trial, clinical outcomes were then reported in 190 a manner relevant to healthcare payers. This included number needed to treat (NNT) for one 191 c Time to recovery was a pre-specified secondary outcome of the LIVE-AIR trial and was defined as the first day on which a patient was discharged or ready for discharge by satisfying one of the following 3 categories from the 8-point ordinal scale: hospitalized, not requiring supplemental oxygen, no longer requiring ongoing medical care; not hospitalized, limitation on activities and/or requiring home oxygen; not hospitalized, no limitations on activities.

. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted February 15, 2022. ; https://doi.org/10.1101/2022.02.11.22270859 doi: medRxiv preprint patient to achieve SWOV, reduction in IMV use, NNT for one life saved, and the number of bed 192 days, ICU days, and IMV days saved. 193 Cost Calculator <150 mg/L), remdesivir use was estimated at 37.7% for patients who received SOC alone, based 213 on the physician survey for treatment of patients with CRP <150 mg/L. 49 As scenario #1 214 considered patients aged <85 years with CRP <150 mg/L and receiving remdesivir, remdesivir 215 use was set at 100% for both the lenzilumab plus SOC and SOC alone arms. For scenarios #2 216 (full mITT population) and #4 (Black patients from the full mITT population), 48.4% of patients in 217 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted February 15, 2022. ; https://doi.org/10.1101/2022.02.11.22270859 doi: medRxiv preprint the SOC alone arm were assumed to receive remdesivir, based on the physician survey for 218 treatment of patients with SpO2 ≤94%, regardless of CRP level. 49 Additional details on the 219 assumptions and calculations for remdesivir use can be found in the Supplementary Appendix. 220

The drug acquisition cost for lenzilumab was £7,300 per patient for the entire treatment 221 course, based on the official UK list price; 53 this was calculated as a cost of £406 per 100 mg vial 222 and a dosing regimen of three 600 mg doses, each administered by a 1-hour intravenous infusion 223 8 hours apart, for a total of 1800 mg per patient. 23 The total drug acquisition cost for remdesivir 224 was £2,040; this was calculated using a cost of £340 per 100 mg vial 54 and assuming a course of 225 6 vials based on the recommended 200 mg loading dose on day 1 and 100 mg maintenance 226 doses on days 2-5. 55 227

Lenzilumab administration costs were assumed to be £33.92 per intravenous infusion, for a total 229 administration cost of £101.76. This cost was calculated based on the pharmacy labor cost 230 associated with the aseptic preparation and administration of a monoclonal antibody. 56,57 Aseptic 231 technique was included for lenzilumab as its preparation requires the compounding of six 100 mg 232 vials and is therefore considered a medium-risk for contamination. 58 It was assumed that 233 remdesivir administration costs were £14.75 per infusion, for a total administration cost of £73.75. 234 This cost was calculated based on the pharmacy labor cost associated with administration of a 235 monoclonal antibody, with no aseptic preparation. 57 It was assumed that additional time for 236 aseptic compounding was not needed for remdesivir as its preparation includes combination of no 237 more than two drug vials and thus is considered a low-risk for contamination. 58 238

Nursing administration costs for intravenous infusions, as well as drug and administration 239 costs for other SOC drugs, including corticosteroids, were assumed to be captured in the hospital 240 resource costs. Costs associated with other SOC drugs were not included as separate inputs in 241 the calculator because it was assumed that lenzilumab would not impact the utilization or cost of 242 background therapies other than remdesivir; this assumption was supported by balanced 243 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted February 15, 2022. 

In alignment with the US hospital perspective model, 45 patients were divided into four levels of 247 care required during their hospital stay: no ICU and no IMV, ICU but no IMV, IMV but no ICU, and 248 both ICU and IMV. Daily hospital resource cost inputs were obtained from a study by Czernichow 249 and colleagues (2020), which reported mean cost per bed type per day in the UK. 16 A cost 250 corresponding to the "IMV but no ICU" level of care was not provided in the study, so it was 251 assumed to be equivalent to the difference between the "both ICU and IMV" and "ICU but no IMV" 252 groups, added to the base hospitalization cost (ie, "no ICU or IMV"). Czernichow and colleagues 253 reported all costs in their study in EUR, using an exchange rate of 1.11 EUR to 1 GBP; 16 these 254 costs were reconverted using that exchange rate and then inflated from 2020 to 2021 values to 255 provide the following hospital costs per day: £876 (no ICU, no IMV), £1,978 (ICU, but no IMV), 256 £2,043 (IMV, but no ICU), and £3,145 (both ICU and IMV). 52 It was assumed that these daily 257 costs incorporated all costs associated with SOC in the UK, including any corticosteroid use. 258

Inputs for hospital resource use were informed by data from the LIVE-AIR trial 50 and included 259 patient distributions and time to recovery for each level of care, as previously reported by 260 Kilcoyne and colleagues (2022) . 45 Costs and resource use for each level of care are provided in 261 the Supplementary Appendix. 262

Hospital resource costs per patient were calculated as a weighted average based on the 263 four levels of care required. First, the average time to recovery for each of the four levels of care 264 was multiplied by the corresponding daily hospital resource cost to obtain the average total 265 hospital resource cost for each level of care. Next, the weighted average cost per patient was 266 calculated as the sum product of average total hospital resource costs for each level of care and 267 the corresponding proportion of patients requiring each level of care. These calculations were 268 performed independently for the lenzilumab plus SOC and SOC alone arms to obtain separate 269 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. and remdesivir use (SOC alone arm only). Daily hospital resource cost inputs were adjusted by 291 ±25% and each level of care was adjusted simultaneously. Given that in the cost calculator, the 292 total hospital resource cost at each level of care is the product of the respective daily hospital 293 resource cost and the time to recovery, the adjustment of daily hospital resource costs by ±25% 294 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted February 15, 2022. ; https://doi.org/10.1101/2022.02.11.22270859 doi: medRxiv preprint 13 has the same impact as adjusting time to recovery by ±25%; therefore, additional sensitivity 295 analyses for time to recovery were not conducted. For patient distributions, the proportion of 296 patients requiring both ICU and IMV was adjusted ±25% first and then the three other levels of 297 care were reweighted, maintaining the proportionality between the levels of care prior to 298 adjustment. The decision to use the ICU and IMV group for adjustment in the sensitivity analysis 299 was based on the assumption that this group was the largest driver of costs in the model, as a 300 result of the increased cost of care per day and the longer length of stay. The input for lenzilumab 301 drug cost was adjusted by +10% and -25%. The variability in the upper estimate (+10% price 302 increase) was selected to be smaller than the lower estimate (-25% price reduction) as the drug 303 cost used in the base case analysis of the cost calculator represents the official UK list price of 304 lenzilumab and, therefore, an increase of more than 10% of the list price in the UK is not 305 anticipated. Lastly, for remdesivir use in the SOC arm, 16.7% was used as the low estimate 306 based on a study by Arch and colleagues 48 and 72.8% was used as the high estimate based on 307 the placebo plus SOC arm in the LIVE-AIR trial. 23 308

In the base case and all scenario analyses, all specified clinical outcomes were improved with 311 lenzilumab plus SOC over SOC alone (Table 1 ). In the base case, the NNT findings showed that 312 every eighth patient treated with lenzilumab plus SOC rather than SOC alone avoided IMV or 313 death. Across the scenario analyses this NNT ranged from a low of 4 in the Black with 314 CRP <150 mg/L subgroup up to 15 in the full LIVE-AIR mITT population. Lenzilumab also 315 provided a benefit in terms of mortality, with a base case NNT of 14 to save one life and this 316 ranged from 8 to 23 across the scenario analyses. The addition of lenzilumab to SOC compared 317 with SOC alone also reduced ventilator use by 12.6% (ranging from 6.5% to 26.8%) and 3.33 IMV 318 days were saved in the base case analysis (ranging from 1.85 up to 5.50 days). Finally, the time 319 to recovery was improved with lenzilumab plus SOC, saving 2.40 bed days (ranging from 0.99 up 320 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted February 15, 2022. In addition to improved clinical outcomes, an estimated cost savings of £1,162 per patient 323 resulted from adding lenzilumab to SOC in the base case analysis ( Table 2) . Scenario analysis 324 #1 (aged <85 years with CRP <150 mg/L, with remdesivir) also produced an estimated cost 325 savings of £3,127 per patient. In scenario analysis #2 (full LIVE-AIR mITT population), addition of 326 lenzilumab to SOC was estimated to result in an additional cost of £4,005 per patient over SOC 327 alone. Scenario analysis #3 (Black patients with CRP <150 mg/L, with or without remdesivir) and 328 scenario analysis #4 (Black patients from the full mITT population) estimated cost savings of 329 £9,977 and £2,369 per patient, respectively. 330

The estimated clinical benefits and cost impact of adding lenzilumab to SOC at the cohort 331 level are shown in Table 3 . In the base case analysis, it was estimated that for a weekly cohort of 332 4,754 newly hospitalized patients, the addition of lenzilumab to SOC would result in 609 333 additional patients achieving SWOV, 599 IMV uses avoided, and 11,400 bed days, 12,978 ICU 334 days, 15,831 IMV days, and 352 additional lives saved. Additionally, in this weekly cohort, 335 lenzilumab plus SOC produced an estimated cost savings of £5,524,952 compared to SOC 336 alone. At the cohort level the scenario analyses also resulted in estimated cost savings ranging 337 from £606,442 (scenario #4) to £5,604,188 (scenario #1), with the exception of scenario #2 (full 338 LIVE-AIR mITT population) which produced an estimated £27,304,581 in additional costs. 339

Overall, the economic impact of the sensitivity analyses was variable relative to the base case 341 results, with adjustments to the inputs for hospital resources costs having the greatest impact on 342 the estimated results. The findings of the sensitivity analyses are summarized in Table 4 . 343

Varying the hospital resource costs resulted in an additional cost of £1,161 per patient 344 with lower (-25%) resource use costs and cost savings of £3,485 per patient with higher (+25%) 345 resource use costs, a decrease and increase in cost savings of 200%, relative to the base case. 346

Varying the patient distribution in the IMV and ICU level of care for the lenzilumab plus SOC arm 347 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted February 15, 2022. ; https://doi.org/10.1101/2022.02.11.22270859 doi: medRxiv preprint resulted in a 125% change in cost savings relative to the base case; reducing the patients in the 348 IMV and ICU level of care by 25% increased the cost savings to £2,610 per patient, while 349 increasing the patients in the IMV and ICU level of care by 25% resulted in an additional cost of 350 £286 per patient. Altering the lenzilumab drug costs produced cost savings of £2,987 per patient 351 with lower drug costs (-25%) and additional costs of £432 per patient with higher drug costs 352 (+10%), an 157% increase and 63% decrease in cost savings, respectively, relative to the base 353 case, respectively. Finally, a higher use of remdesivir in the SOC arm resulted in cost savings of 354 £1,904 per patient, an increase of 64% relative to the base. Conversely, a lower use of remdesivir 355 in the SOC arm decreased the cost savings to £718 per patient, a reduction of 38%. 356

The present analysis evaluated the clinical and economic benefits of adding lenzilumab to SOC 358 for the treatment of COVID-19 pneumonia from the NHS England perspective. The overall 359 objective was to provide evidence to healthcare payers that may support the use of lenzilumab as 360 a treatment option for this condition that urgently requires new efficacious therapies. 361

In the base case and all scenario analyses, treatment with lenzilumab plus SOC 362 improved all specified clinical outcomes over SOC alone and resulted in a greater number of 363 patients achieving SWOV and avoiding IMV use, as well as additional lives, bed days, ICU days, 364 and IMV days saved. In the scenario analyses, the relative clinical benefits were smaller in the 365 broader full mITT population scenario but greater in the narrower subgroup populations (ie, aged 366 <85 years with CRP <150 mg/L receiving remdesivir and Black with CRP <150 mg/L, with or 367 without remdesivir) compared to the base case. 368

Of the clinical outcomes assessed, SWOV is an important outcome for patients with 369 COVID-19 from the NHS England perspective, as the average length of a patient's hospital stay 370 is greatly reduced by avoiding IMV. 59 Notably, the average time to recovery for patients with IMV 371 was more than double that of patients without IMV for all five analysis populations explored from 372 the LIVE-AIR trial. 50 In the base case analysis, addition of lenzilumab to SOC resulted in a 12.6% 373 absolute reduction in the probability of requiring IMV compared with SOC alone among patients 374 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted February 15, 2022. ; https://doi.org/10.1101 https://doi.org/10. /2022 aged <85 years with CRP <150 mg/L. As observed in the LIVE-AIR trial, the estimates for failure 375 to achieve SWOV were similar between patients with CRP <150 mg/L regardless of age and 376 patients aged <85 years with CRP <150 mg/L (hazard ratio [95% confidence interval] of 2.54 377 [1.46-4.41 ] and 3.04 [1.68-5.51], respectively). 39, 40 These results suggest that the clinical benefits 378 observed in the present base case analysis would also likely be generalizable to a broader 379 population of patients with CRP <150 mg/L, irrespective of age. 380

In the base case analysis, lenzilumab was associated with per-patient cost savings from 381 the NHS England perspective. Although there were increases in drug acquisition and 382 administration costs associated with the addition of lenzilumab to SOC compared with SOC 383 alone, these additional costs were offset by the reduced costs for hospital resource use among 384 patients treated with lenzilumab. This resulted in a net per-patient cost savings of £1,162, and an 385 estimated total savings of £5,524,952 for a weekly cohort of 4,754 newly hospitalized patients 386 with CRP <150 mg/L and aged <85 years. Results from the sensitivity analyses suggested that 387 the variable with the greatest impact on the estimated cost savings of adding lenzilumab to SOC 388 versus SOC alone was hospital resource costs (±25%), resulting in a change of 200% from the 389 base case analysis. By comparison, varying patient distributions to different levels of care 390 (±25%), lenzilumab drug costs (+10% and -25%) and using different assumptions for remdesivir 391 use in the SOC alone arm (high estimate of 72.8%, low estimate of 16.7%) had less of an impact 392 on the cost savings associated with adding lenzilumab to SOC in the base case analysis. 393

In addition to the clinical benefits observed in the various populations examined, the 394 results show that adding lenzilumab to SOC for a narrower group of patients with baseline 395 CRP <150 mg/L, aged <85 years, and receiving remdesivir may also provide economic benefits 396 for UK hospitals, with a net per-patient cost savings of £3,127. However, in a broader population 397 of patients with COVID-19 pneumonia without consideration of CRP criteria or age, adding 398 lenzilumab to SOC had additional per-patient costs of £4,005, suggesting that there is greater 399 economic value in providing lenzilumab to a more targeted population. 400

Evidence suggests that CRP levels at the time of admission have a positive correlation 401 with disease severity and are associated with adverse clinical outcomes in patients hospitalized 402 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted February 15, 2022. ; https://doi.org/10.1101 https://doi.org/10. /2022 with [34] [35] [36] [37] 40 Furthermore, one of the strongest independent predictors of critical 403 illness is CRP levels >100 mg/L at the time of admission 34,35 , with patients having CRP 404 levels >150 mg/L considered to be high risk for escalation of respiratory support (ie, need for non-405 invasive ventilation or intubation) or death. 60 Given that in patients aged <85 years with CRP 406 levels <150 mg/L, irrespective of remdesivir use, lenzilumab produced favorable clinical 407 outcomes and cost savings, it appears that using lenzilumab as an early intervention would be 408 most effective and provide the best economic value. As testing for CRP levels is accessible and 409 inexpensive in the UK hospital setting and has previously been recommended by the National 410

Institute for Health and Care Excellence (NICE) as an option to inform the diagnosis and 411 treatment of pneumonia in adults, 61 it may be a valuable and feasible approach to identify 412 hospitalized COVID-19 patients who may benefit most from lenzilumab treatment. 413

Clinical trials frequently underrepresent the Black population, despite this population 414 being disproportionately affected by Consequently, there is limited evidence across 415 racial populations regarding differences in disease severity, outcomes, and treatments. 62 An 416 exploratory analysis of the LIVE-AIR trial revealed that Black and African American patients, 417 particularly those with a CRP level <150 mg/L, demonstrated the greatest response to treatment 418 with lenzilumab. 41 In the current analysis, the use of lenzilumab in Black patients with CRP <150 419 mg/L resulted in cost savings of £9,977 per patient, and a cost savings of £2,369 per patient 420 when all Black patients were assessed. Assuming the Black and African American subgroup of 421 LIVE-AIR are representative of potential outcomes in the Black African and Black Caribbean 422 persons in the UK, these findings may be significant since this demographic is disproportionally 423 affected by COVID-19, with an increased rate of death. 42 424

As with extrapolation of any clinical trial results, this study had several limitations. The 425 analysis used data from the LIVE-AIR Phase 3 clinical trial that may have been subject to 426 selection bias and that was predominantly conducted in US hospitals. Consequently, the findings 427 from the trial may not be fully generalizable to a UK population and/or UK hospital setting based 428 on potential differences in treatment patterns (eg, length of hospital stay) for hospitalized patients 429 with COVID-19 between the US and UK. It should also be noted that the LIVE-AIR trial was 430 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted February 15, 2022. ; https://doi.org/10.1101 https://doi.org/10. /2022 conducted earlier in the pandemic and thus, prior to vaccinations and the emergence of new 431 COVID-19 variants, both of which have been reported to affect hospital length of stay. 11,63,64 As a 432 result, time to recovery inputs derived from the clinical trial may differ from current real-world 433 values. However, it is anticipated that the ongoing ACTIV-5/BET-B trial, which has completed the 434 enrollment of over 400 patients in the primary analysis population, will provide data on the use of 435 lenzilumab for hospitalized patients infected with different COVID-19 variants and with differing 436 vaccination statuses. 46,47 Additionally, the Black subgroups from the LIVE-AIR trial that were 437 assessed in the per-patient calculator were limited by small sample sizes and will require further 438 validation using the results from the upcoming ACTIV-5/BET-B trial. Consideration should also be 439

given to the fact that the current analyses were conducted ex ante and therefore, were conducted is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted February 15, 2022. ; https://doi.org/10.1101 https://doi.org/10. /2022 19 patients with COVID-19 pneumonia, particularly early in the disease process to prevent ICU 459 admission, IMV, and/or death. Lenzilumab is a particularly promising therapeutic as it addresses 460 an unmet need for treatment options that prevent the progression to IMV and/or death in 461 hospitalized patients. Additionally, the mechanism of action of lenzilumab functions by binding 462 and neutralizing GM-CSF, which is produced in response to COVID-19 infection. Therefore . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted February 15, 2022. ; https://doi.org/10.1101 https://doi.org/10. /2022 

Patients Using C-Reactive Protein as a Biomarker Improves Efficacy: Results From the 618 Phase 3 'LIVE-AIR' Trial. medRxiv. 2022. doi: 619

Lenzilumab Treatment May Provide Enhanced Likelihood of Survival 621

Without Ventilation in Hospitalized Black and African-American COVID-19 Patients

TreatmentMay-Provide-Enhanced-Likelihood-of-Survival-Without-Ventilation-in-624

Updating ethnic contrasts in deaths involving the 627 coronavirus (COVID-19)

es/24january2020to31march2021

Lessons about COVID-19 vaccine hesitancy among minority 632 ethnic people in the UK

Mitigating ethnic disparities 634 in covid-19 and beyond

Clinical and economic benefits of lenzilumab plus 636 standard of care compared with standard of care alone for the treatment of hospitalized 637 patients with COVID-19 in the United States from the hospital perspective

A UK Time and Motion Study in Comparison with Intravenous 675

Formulation for the Treatment of Patients with HER2-Positive Early Breast Cancer

Ravulizumab for treating atypical 678 haemolytic uraemic syndrome (aHUS) [ID1530] committee papers

USP 797: A breakdown of low, medium and high compounding risks

Joint analysis of duration of ventilation, length of 685 intensive care, and mortality of COVID-19 patients: a multistate approach

COVID-19-associated hyperinflammation and 688 escalation of patient care: a retrospective longitudinal cohort study

Cost-Effectiveness Analysis 691 of the Use of Point-of-Care C-Reactive Protein Testing to Reduce Antibiotic Prescribing 692 in Primary Care

Racial Disproportionality in Covid Clinical Trials

Hospital length of stay for COVID-19 patients

Data-driven methods for forward planning

International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity

Hospitlisation risk of Omicron 698 cases in England

701 65. Humanigen. FDA has declined Humanigen's Emergency Use Authorization (EUA)

Request for Lenzilumab in Hospitalized COVID-19 Patients

707 66. Government of the United Kingdom. Daily Update: UK Summary

709 67. National Institute for Health and Care Excellence

COVID-19 illness in native and immunosuppressed states: A 714 clinical-therapeutic staging proposal

Dexamethasone in Hospitalized Patients with Covid-19

Authorization for Actemra (Tocilizumab) for Treatment of COVID-19

International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity

The authors acknowledge Cameron Durant (Humanigen Inc.) and Avery Hughes (EVERSANA) 499 for their specific contributions to this project. 500

All authors made substantial contributions to conception and design, acquisition of data, or 502 analysis and interpretation of data; took part in drafting the article or revising it critically for 503 important intellectual content; agreed to submit to the current journal; gave final approval of the 504 version to be published; and agree to be accountable for all aspects of the work 505 Funding 506 This work was supported by Humanigen Inc. 507 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted February 15, 2022. . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted February 15, 2022. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted February 15, 2022. ; https://doi.org/10.1101 https://doi.org/10. /2022 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted February 15, 2022. ; https://doi.org/10.1101 https://doi.org/10. /2022 Notes: 776 a Estimated cohort sizes were calculated using the total number of new COVID-19 hospital admissions in England over a one-week period (11,343; 777 from January 17 to January 23, 2022). 13 Based on the physician survey, 60.1% of hospitalized patients have SpO2 ≤ 94% 49 and it was assumed 778 that 77.9% of this population has CRP <150 mg/L, as observed in the LIVE-AIR trial. 23 For base case and scenario #1, it was assumed that 89.5% 779 of the hospitalized population was <85 years 51 and the subgroup receiving remdesivir was calculated using data from the physician survey which 780 estimated 37.7%. 49 Lastly, for scenarios #3 and #4, it was assumed that 3.8% of the patient were Black Abbreviations: COVID-19, Coronavirus Disease 19; CRP, C-reactive protein; mITT, modified intent-to-treat; SpO2, oxygen saturation. 786