Content uploaded by Cadegiani Flávio
Author content
All content in this area was uploaded by Cadegiani Flávio on Jul 01, 2021
Content may be subject to copyright.
Proxalutamide (GT0918) Improves Lung Injury in Hospitalized COVID-19
Patients – an Analysis of the Radiological Findings of the Proxa-Rescue AndroCoV
Trial
Flávio Adsuara Cadegiani, MD, PhD1,2*, Daniel do Nascimento Fonseca, MD3, Michael
do Nascimento Correia, MD4, Renan Nascimento Barros, MD5, Dirce Costa Onety,
MD3, Karla Cristina Petruccelli Israel, MD6,7, Emilyn Oliveira Guerreiro3, José Erique
Miranda Medeiros3, Raquel Neves Nicolau3, Luiza Fernanda Mendonça Nicolau3,
Rafael Xavier Cunha3, Maria Fernanda Rodrigues Barroco3, Patrícia Souza da Silva3,
Raysa Wanzeller de Souza Paulain, MD5, Claudia Elizabeth Thompson, PhD8, Ricardo
Ariel Zimerman, MD9, Carlos Gustavo Wambier, MD, PhD10, Andy Goren, MD2
1Corpometria Institute, Brasilia, Brazil.
2Applied Biology, Inc. Irvine, CA, USA.
3Samel & Oscar Nicolau Hospitals, Manaus, Brazil
4Hospital Regional José Mendes, Itacoatiara, Amazonas, Brazil
5Hospital Municipal Jofre Cohen, Parintins, Amazonas, Brazil
6Centro de Doenças Renais do Amazonas, Manaus, Brazil
7Programa de Pós-Graduação em Medicina Tropical – FMT/UEA, Manaus, Brazil
8Department of Pharmacosciences, Universidade Federal de Ciências da Saúde de Porto
Alegre
9Hospital da Brigada Militar, Porto Alegre, Brazil
10Department of Dermatology, Alpert Medical School of Brown University, Providence,
RI, USA.
*Corresponding author:
Flavio Adsuara Cadegiani, MD, PhD
Applied Biology, Inc.
17780 Fitch
Irvine, CA 92614
f.cadegiani@gmail.com
Abstract Word Count: 407
Text-Only Word Count: 2914
Total Number of References: 26
Tables (#): 2
Figures (#): 2
Abstract
Introduction: Antiandrogen are good candidates against the severe acute respiratory
syndrome coronavirus 2 (SARS-CoV-2) disease (COVID-19) due to the inhibition of its
entry into host cells by the suppression of TMPRSS2, an enzyme that primes the SARS-
CoV-2 spike (S) protein and is key for its cell entry. Proxalutamide is a second-
generation nonsteroidal anti-androgen (NSAA) with strong activities on androgen
receptor (AR) antagonism, suppression of AR nuclear expression, and downregulation
of the membrane-attached angiotensin converting enzyme-2 (ACE2). The efficacy of
proxalutamide was previously demonstrated for early COVID-19 patients, and has now
demonstrated efficacy to reduce deaths in hospitalized COVID-19 patient in a double-
blind, placebo-controlled randomized clinical trial (RCT). Whether radiological changes
would follow the improvement in clinical outcomes with proxalutamide is not
established. The present post-hoc analysis aims to evaluate whether proxalutamide
improves lung injury observed through chest computed tomography (CT) scans, in
addition to the clinical improvement, thus providing further objective evidence of the
drug response in COVID-19.
Methods: This is a post-hoc analysis of the radiological findings of a double-blinded,
placebo-controlled, prospective, two-arm RCT (The Proxa-Rescue AndroCoV Trial)
with all enrolled patients from the three participating institutions of the city of Manaus,
Amazonas, Brazil, that had at least two chest CT scans during hospitalization. The
quantification of lung parenchyma involvement was performed by independent board-
certified radiologists with expertise in analysis of COVID-19 images, that were blind to
the assigned intervention in the RCT. A first chest CT scan was performed upon
randomization and a second CT scan was performed approximately five days later,
whenever patient transportation was feasible.
Results: Of the 395 patients initially evaluated, 72 and 179 patients from the
proxalutamide and placebo arms, respectively, were included (n=251). Baseline and
clinical characteristics, interval between first and second chest CT scans, and percentage
of lung parenchyma affected in the baseline chest CT scan were similar between groups.
In the second chest CT scan, the percentage of lungs affected (Median – IQR) was
35.0% (25.0-57.5%) in the proxalutamide group versus 67.5% (50.0-80.0%) in the
placebo group (p < 0.001). The absolute and relative change between the second and
first chest CT scans (Median – IQR) were -15.0 percent points (p.p.) (-30.0 – 0.0p.p.)
and -25.0% (-50.0 – 0.0%) in the proxalutamide group, respectively, and +15.0p.p. (0.0
- +30.0p.p.) and +32.7% (0.0 - +80.0%) in the placebo group, respectively (p < 0.001
for both absolute and relative changes).
Conclusion: Proxalutamide improves lung opacities in hospitalized COVID-19 patients
when compared to placebo. (NCT04728802)
Keywords: COVID-19, SARS-CoV-2, proxalutamide, antiandrogen, non-steroidal
antiandrogen (NSAA), lung injury.
Introduction
Antiandrogen drugs are good candidates against the severe acute respiratory syndrome
coronavirus 2 (SARS-CoV-2) disease (COVID-19) due to the inhibition of its entry into
cells.1 The blockade of viral cell entry would occur indirectly by the mitigation of the
structural modification of the SARS-CoV-2 spike (S) protein that is required for
infection. This inhibition occurs though the reduction of the expression of the
endogenous transmembrane protease, serine 2 (TMPRSS2). The enzyme is responsible
for priming the viral S protein, allowing its proper coupling to ACE-2 receptor and
consequent cell entry.2 The suppression of the TMPRSS2 expression occurs through the
inhibition of the TMPRSS2 promoter, that includes a 15 base pair androgen response
element.3 Since androgens are the only known endogenous regulators of TMPRSS2,
antiandrogens play a key role in the inhibition of TMPRSS2 expression. Pre-clinical
studies have shown that nonsteroidal antiandrogens down regulate TMPRSS24 and
inhibit viral replication in human cell culture.5,6 However, other mechanisms to explain
the potential interplay between antiandrogens and SARS-CoV-2, such as
downregulation of ACE-2 receptors, may play an additional role, and should be further
elucidated.
Proxalutamide is a second-generation nonsteroidal anti-androgen (NSAA) that is more
potent than other NSAAs, such as enzalutamide or bicalutamide.7 In addition to the
competitive antagonism in the androgen receptor (AR), NSAAs also prevent androgen
receptor nuclear translocation and binding to DNA.8
The efficacy of proxalutamide was previously demonstrated for SARS-CoV-2 positive
men in an outpatient setting.9,10 The Proxa-Rescue AndroCoV Trial has now
demonstrated the efficacy of proxalutamide for hospitalized COVID-19 men and
women patients regarding clinical recovery speed (128% increase in recovery speed)
and reduction of mortality rate (77.7% reduction in the 28-day mortality rate), though a
double-blinded, placebo-controlled, multicenter randomized clinical trial, when
compared to patients under usual care.11
Radiological improvement of COVID-19 tends to occur later in the recovery process,
with easing of lungs appearance on chest computed tomography (CT) scan usually
observed only after seven to 14 days 12 . Due to the dramatic improvement observed in
part of the patients of the proxalutamide arm, we hypothesized that an early
improvement in the radiological aspect would also occur, which would reinforce the
efficacy of proxalutamide for hospitalized COVID-19 patients.
The objective of the present analysis is to compare the radiological findings of
hospitalized COVID-19 patients included in the RCT between the proxalutamide and
placebo arms, to reinforce the efficacy of proxalutamide observed clinically through
increase in recovery speed and reduction of mortality rate.
Methods
Trial Design, Setting and Locations
Study design, criteria for eligibility, randomization, procedures, outcomes are described
elsewhere.11
This is a post-hoc analysis of the radiological findings of a double-blinded, placebo-
controlled, prospective, two-arm randomized clinical trial (RCT), that encompassed the
three institutions of the city of Manaus, Amazonas, Brazil, that participated in the study.
The other five centers included in the RCT were not included due to the lack of
available CTs for regular evaluation of COVID-19. The study was conducted between
February 1 and April 15, 2021, including enrollment and follow-up.
The RCT was approved by Brazilian National Ethics Committee of the Ministry of
Health, under the approval number 4.513.425 of the process number (CAAE)
41909121.0.0000.5553 (original name of the Ethics Committee: Comitê de Ética em
Pesquisa (CEP) do the Comitê Nacional de Ética em Pesquisa (CONEP) do Ministério
da Saúde - CEP/CONEP/MS). All data used for the present post-hoc analysis was
entirely covered by the approval obtained with the Brazilian National Ethics Committee
of the Ministry of Health (MS) (approval number 4.513.425). The RCT was registered
in clinicaltrials.gov (NCT04728802).
Eligibility criteria
In short, for inclusion, men and women above 18 years old hospitalized due to COVID-
19 confirmed with a positive real-time reverse transcription polymerase chain reaction
(rtPCR) test for SARS-CoV-2 (Cobas SARS-CoV-2 rtPCR kit test protocol, Roche,
USA) were considered.
Exclusion criteria included mechanical ventilation at the time of randomization, known
congestive heart failure class III or IV (New York Heart Association),
immunosuppression, alanine transferase (ALT) above five times ULN (> 250 U/L),
creatinine above 2.5 mg/ml or a calculated eGFR below 30 ml/min, current use of
antiandrogen medications, planning to attempt to have kids within 90 days after the
intervention, and women that were pregnant or breastfeeding.
For the present post-hoc analysis of the radiological findings, all patients that
participated in the Proxa-Rescue AndroCoV Trial11 from the three hospitals located in
the city of Manaus, Amazonas, Brazil, were included. There were no selection criteria
among patients enrolled in the RCT from these hospitals for the initial assessment.
Patients with at least two chest CT scans during hospitalization were included in the
present analysis, since at least two scans were needed for comparison purposes. All
potential limitations of a subgroup post-hoc analysis of a RCT described by Pocock et
al were addressed.13
Procedures
Patients were randomized to receive either proxalutamide 300 mg/day plus usual care or
a placebo plus usual care for 14 days in a 1:1 ratio. If patients were discharged before
14 days, they were instructed to continue treatment. Therapy compliance was monitored
daily for both inpatients and patients that were discharged until day 14, and then in days
21 and 28 if discharged before, or daily if still hospitalized.
The COVID-19 8-point ordinal scale was used as the parameter for monitoring. The
ordinary clinical scale is defined as: 8. Death; 7. Hospitalized, on invasive mechanical
ventilation; 6. Hospitalized, on non-invasive ventilation or high flow oxygen devices; 5.
Hospitalized, requiring supplemental oxygen; 4. Hospitalized, not requiring
supplemental oxygen- requiring ongoing medical care (COVID-19 related or
otherwise); 3. Hospitalized, not requiring supplemental oxygen - no longer requires
ongoing medical care; 2. Not hospitalized, limitation on activities; and 1. Not
hospitalized, no limitations on activities.11
Baseline characteristics, previous medical history, comorbidities and concomitant
medications were recorded. Usual care for hospitalized COVID-19 patients as per the
hospitals protocol included enoxaparin, colchicine, methylprednisolone or
dexamethasone, and antibiotic therapy as required. The usual care was not changed for
the RCT.
Before the onset of the RCT, a random sequence using 4, 6 and 8 block sizes and a list
length for 662 treatments was created thought a randomization software. 14 The
randomization sequence and allocation concealment were performed remotely and was
not stratified by institution. Pre-packing of tablets of either active or placebo group was
manufactured to have identical physical characteristics, and was manufactured and
transported by Kintor Pharmaceuticals Ltd. Suzhou, China.
Protocol for the exploratory analysis of the chest CT scans
As per the protocol of the three hospitals located in the city of Manaus, Amazonas,
Brazil, patients hospitalized due to COVID-19 had chest CT-scans approximately every
five days, or whenever it was feasible to transport patient to the CT-scan room. A first
chest CT scan was performed upon randomization and a second CT scan was performed
approximately five days later, whenever patient health condition permitted the
transportation. Patients in ICU or clinically unstable were not eligible for the five-day
interval CT-scan follow-up.
The analysis of the chest CT scan was performed by board-certified radiologists with
previous clinical expertise in COVID-19, that quantified the percentage of lung
parenchyma involved in COVID-19, based on the classifications proposed by Xie et al,
Zhao et al, Pan et al, Li et al, Chung et al, and Yuan et al,15-20 following the
standardization proposed by Martinez Chamorro et al.12 The three hospitals unified and
standardized the methods for the quantification of lung affected in order to avoid inter-
operator differences. All chest CT scans were performed in CT SOMATOM model with
64-slice data acquisition (Simens Healthineers, Siemens, Germany).
Bilateral reticular patterns, peripheral bilateral ground-glass opacities, and patchy or
confluent multifocal consolidation were considered as findings consistent with COVID-
19 pneumonia. Central consolidation and unilateral ground-glass opacities were
considered as indeterminate for COVID-19. Pneumothorax, pneumomediastinum,
pleural effusion, lobar consolidation, military patterns or cavitation were not considered
as part of COVID-19 pneumonia, although a series of case reports have described the
first two characteristics. Long term fibrotic changes, such as honeycombing or traction
bronchiectasis, could be consequences of COVID-19, but were not considered as part of
the quantification of lungs affected. In short, only CO-RADS 6 were included in this
analysis. All patients had diagnosis of COVID-19 through positive rtPCR-SARS-CoV-2
test.21
The analyses were performed in a complete independent manner. Radiologists were not
informed whether patients were or were not participating in the RCT, as well as in
which arm they were designated.
For quantification purposes, whenever an interval of percentage was provided instead of
an exact percentage, this was replaced by an exact value, as following: <5% = 2%;
<10% = 5%: <25% = 10%; 10-25% = 20%; <30% = 15%; <50% = 30%; 25-50% =
40%; >30% = 50%: 50-60% = 60%; >50% = 70%; 50-75% = 65%; >75% = 90%; >80%
= 90%; and >90% = 95%.
For the present exploratory analysis, all COVID-19 hospitalized patients enrolled in the
RCT from the three hospitals of the city of Manaus, Amazonas, were initially
considered. Among these patients, all those with at least two chest CT scans during
hospitalization were included for the analysis.
Endpoints
The differences in the quantification of lung parenchyma affected by COVID-19 chest,
seen through chest CT scan results, between the baseline (first) and second exam, in
terms of: 1. absolute changes (in points percent – p.p.; eg.. If the first CT scan showed
50% of lungs affected and the second CT scan showed 25% of lungs affected, a 25p.p.
reduction – -25p.p – was observed); and 2. relative changes (percentage of change.; eg..
If the first CT scan showed 50% of lungs affected and the second CT scan showed 25%
of lungs affected, a 50%. reduction – -50% – was observed, compared to the first CT
scan), were compared between proxalutamide and placebo arms.
Baseline characteristics and clinical outcomes are described to evaluate whether the
dimension of the drug efficacy of the RCT is represented in this post-hoc analysis.
Statistical Analysis
An original intention-to-treat (ITT) protocol (unmodified) was used for data analysis.
Analysis was not stratified by sex since both men and women presented similar clinical
responses to proxalutamide in the RCT compared to placebo.
Cox proportional hazards model was used to calculate hazard ratio (HR) for all-cause
14-day and 28-day mortality and their 95% confidence interval (CI), to measure the
effects of proxalutamide versus placebo. Non-parametric Kruskal-Wallis Test was
employed to measure the effects of proxalutamide versus placebo for radiological
endpoints, and disclosed as p-values. All statistical tests were performed using IBM-
SPSS statistics version 25.0 software (IBM, USA).
Results
The flowchart depicting the subject selection in the proposed protocol is shown in
Figure 1. A total of 395 patients were initially evaluated, including 97 patients in the
active arm and 298 patients in the placebo arm, which corresponds to all patients
enrolled in the three hospitals from Manaus, Amazonas, Brazil. Imbalances between
sites in terms of active and placebo proportions are explained elsewhere.11 Of these, 72
patients from the proxalutamide arm and 179 patients from the placebo arm had at least
two chest CT scans, and were included in the present analysis.
Figure 1. Protocol flowchart.
Baseline characteristics, including age, proportion between males and females, and
presence of comorbidities, and additional parameters such as median time since
hospitalization, distribution of the score in the COVID-19 ordinary scale, and use of
concomitant medications were similar between proxalutamide and placebo group (Table
1).
Of the patients included in the present exploratory analysis, the 14-day mortality was
3.1% in the proxalutamide group and 38.9% in the placebo group, with a mortality risk
ratio (RR) of 0.08 (0.03-0.14). The 28-day mortality was 6.2% in the proxalutamide
group and 48.7% in the placebo group, with a mortality RR of 0.13 (0.06-0.28). The
Figure 1. Procotol flowchart –
Analysis of the chest CT scans in the Proxa-Rescue AndroCoV Trial.
CT = computed tomography
Assessed for eligibility (n=697)
Excluded (n=52)
¨ Not meeting inclusion criteria (n=12)
¨ Declined to participate (n=40)
At least 2 CT scans (n=72)
¨ 3 Hospitals – City of Manaus, Amazonas (n=97)
Allocated to proxalutamide
¨ Overall (n=317)
¨ 3 Hospitals – City of Manaus, Amazonas (n=97)
¨ 3 Hospitals – City of Manaus, Amazonas (n=298)
Allocated to placebo
¨ Overall (n=328)
¨ 3 Hospitals – City of Manaus, Amazonas (n=298)
At least 2 CT scans (n=179)
Allocation
CT scans
Initial analysis
Randomized
¨ Overall (n=645)
¨ 3 Hospitals – City of Manaus, Amazonas (n=395)
Enrollment
Analysed (n=72)
Analysed (n=179)
Analysis
Initial analysis!
median hospitalization length stay was 8.0 days in the proxalutamide group and 12.0
days in the placebo group (p < 0.0001) (Table 1).
No drug-related severe adverse effects (SAEs) were reported (Table 1).
Table 1. Baseline clinical characteristics, outcomes, and adverse effects.
INTENTION-TO-TREAT IN THE
THREE INTITUTIONS LOCATED IN THE CITY OF MANAUS, AMAZONAS, BRAZIL
Characteristic
Overall
N= 395
Proxalutamide
N=97
Placebo
N=298
p
Age
Median – years (IQR)
47.5 (38-59)
46.5 (39-59)
47.5 (37-60)
n/s
> 55 yr – no. (%)
254 (36.5%)
34 (35.1%)
110 (36.9%)
n/s
Sex – no. (%)
Female
168 (42.6%)
37 (38.2%)
131 (44.0%)
.527
Male
227 (57.4%)
60 (61.8%)
167 (56.0%)
BMI > 30 – no. (%)
35 (8.9%)
s12 (12.4%)
23 (7.7%)
n/s
Hypertension – no. (%)
99 (25.1%)
30 (30.9%)
69 (23.2%)
n/s
Type 2 diabetes mellitus – no. (%)
35 (8.9%)
13(13.4%)
32 (10.7%)
n/s
COPD – no. (%)
12 (3.3%)
5 (5.2%)
7 (2.3%)
n/s
Chronic kidney disease – no. (%)
0 (0.0%)
0 (0.0%)
0 (0.0%)
n/s
Median time from hospitalization to
randomization (IQR) – days
2.0 (1.0-4.0)
2.0 (1.0-3.0)
2.0 (1.0-4.0)
n/s
Score on the COVID-19 ordinal scale– no. (%)
3. Hospitalized, not requiring
supplemental oxygen - no longer requires
ongoing medical care
2 (0.5%)
0 (0.0%)
2 (0.7%)
n/s
4. Hospitalized, not requiring
supplemental oxygen, requiring ongoing
medical care (COVID-19 related or
otherwise)
13 (3.3%)
3 (3.1%)
10 (3.4%)
n/s
5. Hospitalized, requiring supplemental
oxygen
117 (29.6%)
22 (22.7%)
95 (31.9%)
n/s
6. Hospitalized, receiving non-invasive
ventilation or high flow oxygen devices
263 (66.6%)
72 (74.2%)
191 (64.1%)
n/s
Concomitant medications – no. (%)
Ceftriaxone
638 (98.9%)
95 (97.9%)
295 (99.0%)
n/s
Colchicine
407 (63.1%)
97 (100%)
298 (100%)
n/s
Enoxaparin
645 (100%)
97 (100%)
298 (100%)
n/s
Macrolides (azithromycin,
clarithromycin)
631 (97.8%)
89 (91.7%)
296 (99.3%)
n/s
Glucocorticosteroids (dexamethasone,
methylprednisolone)
645 (100%)
97 (100%)
298 (100%)
n/s
Omeprazole
645 (100%)
97 (100%)
298 (100%)
n/s
Score on the COVID-19 ordinal scale at Day 14–
Median (IQR)
5.0 (1-7)
1.0 (1-2)
6.5 (2-8)
< 0.001
Score on the COVID-19 ordinal scale at Day 28–
Median (IQR)
4.0 (1-8)
1.0 (1-1)
5.5 (2-8)
< 0.001
Mortality at 14 days
119 (30.1%)
3 (3.1%)
116 (38.9%)
< 0.001
Mortality at 28 days
151 (38.2%)
6 (6.2%)
145 (48.7%)
< 0.001
Median hospitalization days (IQR)
11.0 (7.0-17.0)
8.0 (6.0-12.0)
12.0 (8.0-18.0)
< 0.001
Median hospitalization days after randomization
(IQR)
8.0 (5.5-12.0)
6.0 (4.0-8.0)
9.0 (6.0-14.0)
< 0.001
Grade 5 – n (%)
Death, Day 14
119 (30.1%)
3 (3.1%)
116 (38.9%)
< 0.001
CI = Confidence interval; IQR = Interquartile range; BMI = body mass index; COPD: chronic obstructive pulmonary disorder,
COVID-19: coronavirus disease 2019.
The number of patients with at least two chest CT scans during hospitalization due to
COVID-19 was 72 (74.2%) in the proxalutamide group and 179 (60.1%) in the placebo
group (p < 0.001). The median interval between two chest CT scans was 5.0 days for
both groups (p = n/s).
Radiological findings are described in Table 2. Figure 2 illustrates the percentage of
lung parenchyma affected in the baseline and on-treatment chest CT scans (A) and the
variation of percentage of lungs affected between baseline and on-treatment chest CT
scans (B). The percentage of lung parenchyma involvement due to COVID-19 in the
baseline chest CT scan (Median – Interquartile range (IQR)) was 60.0% in both
proxalutamide and placebo groups, when all patients with a baseline chest CT scan were
considered. When only those with at least two chest CT scans were considered, the
percentage of lung affected (Median – IQR) was 60.0% (45.0-70.0%) in the
proxalutamide group and 50.0% (30.0-70.0%) in the placebo group. In both cases,
baseline chest CT scan was statistically similar between groups.
The percentage of lungs with COVID-19 opacities in the second chest CT scan (Median
– IQR) was 35.0% (25.0-57.5%) in the proxalutamide group and 67.5% (50.0-80.0%) in
the placebo group (p < 0.001).
Death, Day 28
151 (38.2%)
6 (6.2%)
145 (48.7%)
< 0.001
Grades 4 or 3 – n (%)
Mechanical ventilation, Day 14
36 (9.1%)
3 (3.1%)
33 (11.1%)
< 0.001
Mechanical ventilation, Day 28
3 (0.8%)
1 (1.0%)
2 (0.7%)
n/s
Renal failure (creatinine increase > 100%)
22 (5.6%)
3 (3.1%)
19 (6.4%)
0.07
Liver damage (ALT > 250 U/L or >100%
increase)
20 (5.1%)
4 (4.1%)
16 (5.4%)
n/s
Grades 2 or 1 – n (%)
Diarrhea
32 (8.1%)
22 (22.7%)
10 (3.4%)
0.005
Abdominal pain
3 (0.8%)
2 (2.1%)
1 (0.3%)
n/s
Irritability
4 (1.0%)
4 (4.1%)
0 (0.0%)
n/s
Spontaneous erection
4 (1.0%)
4 (4.1%)
0 (0.0%)
n/s
Vomiting, dyspepsia, or palpitations
0 (0.0%)
0 (0.0%)
0 (0.0%)
-
The absolute change between the second and first chest CT scans (Median – IQR) was -
15.0 percent points (p.p.) (-30.0 – 0.0p.p.) in the proxalutamide group, and +15.0p.p.
(0.0 - +30.0p.p.) in the placebo group (p < 0.001).
The relative change in terms of percentage between the second and first chest CT scans
(Median – IQR) was -25.0% (-50.0 – 0.0%) in the proxalutamide group and +32.7%
(0.0 - +80.0%) in the placebo group (p < 0.001).
Table 2. Radiological outcomes.
CT = computed tomograph; IQR = interquartile range
* Included for analysis; **In the day of randomization;
Parameter
Overall
N=395
Proxalutamide
N=97
(n=72*)
Placebo
N=298
(n=179*)
p
Number of subjects with at least 1 Chest
CT scan – n (%)
357 (90.4%)
89 (91.8%)
268 (89.9%)
n/s
Number of subjects with at least 2 Chest
CT scans – n (%)
251 (63.5%)
72 (74.2%)
179 (60.1%)
< 0.001
Number of subjects with unknown
percentage of affected lungs in first chest
CT scan – n (%)
4 (1.0%)
0 (0.0%)
4 (1.3%)
n/s
Number of subjects with unknown
percentage of affected lungs in second
chest CT scan – n (%)
4 (1.0%)
1 (1.0%)
3 (1.0%)
n/s
% of affected lungs in the first chest CT-
scan, including those without a second
exam - Median (IQR)*
60.0
(40.0-70.0)
60.0
(40.0-70.0)
60.0
(40.0-70.0)
n/s
% of affected lungs, excluding those
without a second exam - Median (IQR)**
55.0
(40.0-70.0)
60.0
(45.0-70.0)
50.0
(30.0-70.0)
n/s
Interval (days) between first and second
chest CT scan - Median (IQR)
5.0
(4.0-7.0)
5.0
(4.0-6.0)
5.0
(4.0-7.0)
n/s
% of affected lungs in the second chest
CT-scan - Median (IQR)
55.0
(40.0-70.0)
35.0
(25.0-57.5)
67.5
(50.0-80.0)
< 0.001
Absolute change (%) between first and
second chest CT scan - Median (IQR)
0.0
(-10.0 - +20.0)
-15.0
(-30.0 – 0.0)
+15.0
(0.0 - +30.0)
< 0.001
Relative change in terms of percentage
between first and second chest CT scan -
Median (IQR)
0.0%
(-30.0 - +20.0%)
-25.0%
(-50.0 - 0.0%)
+32.7%
(0.0 - +80.0%)
< 0.001
Figure 2. CT scans.
Discussion
The present analysis reinforces the efficacy of proxalutamide for hospitalized patients
with COVID-19. The radiological improvement provides an independent and objective
evaluation of the efficacy of the drug and predicts better clinical outcomes.12,21 Unlike
reports of clinical improvement, radiological findings analyzed by independent
radiologists are not influenced by placebo effect, even though objective parameters such
as the 8-point WHO COVID ordinary scale is would also hardly be influenced by this
effect, except scores 1 to 4. Hence, an open label study would probably be enough to
imply causality, without major interferences of the results from lack of blinding.
However, our exploratory analysis derived from a double-blind, placebo-controlled,
two-arm RCT. Importantly, radiologists that analyzed the images were blind to the
intervention.
We analyzed the results utilizing the most appropriate type of analysis, the unmodified
ITT, a more conservative analysis that tend to underestimate drug efficacy,22 in the
primary analysis of the RCT. Utilizing ITT population, we were able to find a 87%
reduction in the 28-day all-cause mortality in the city of Manaus. In an on-treatment
0
20
40
60
80
100
Proxalutamide
(n=72)
% Lungs affected (Chest CT Scan)
Placebo
(n=179)
Baseline
CT scan
On-treatment
CT-scan
(median = 5 days)
Median (Interquartile Range)
0
-20
-10
+10
+20
+30
Proxalutamide
(n=72)
% Change between baseline and on-treatment
(Chest CT Scan)
Placebo
(n=179)
Median (Interquartile Range)
-30
-40
-50
+50
+40
+60
+70
+80
A
B
Figure 2. Baseline and on-treatment chest computed
tomography (CT) scans in proxalutamide versus placebo.
(OT) analysis, the magnitude of the results tended to be even higher (92% reduction in
28-day mortality rate).11. The conservative nature of the ITT analysis and the higher
efficacy on treatment completers, compared to non-treatment completers, showing a
“dose response-like” behavior, are also suggestive of the efficacy of proxalutamide for
hospitalized COVID-19 patients, in addition to the primary findings.
We avoided selection of patients. Instead, we included all participating subjects from
the three hospitals. This regional subgroup analysis maintained a balance between
proxalutamide and placebo groups in terms of demographic and baseline characteristics.
Samples were precisely representative of the groups in the overall analysis. Typical
limitations of a post-hoc analysis of a RCT are mostly absent in the present case.13
The number of patients without at least two chest TC scans was significantly higher in
the placebo group than in the proxalutamide group, possibly because the number of
patients that needed ICU was significantly higher in the placebo group (Table 2). Due to
the absence of a mobile CT scan, the performance of a CT scan became unfeasible
when patients needed ICU, in particular when they were under mechanical ventilation
(Table 2).
The present analysis possibly underestimates the efficacy of proxalutamide for
hospitalized patients with COVID-19, for two reasons: 1. Patients that had better
responses to proxalutamide did not undergo a second CT scan, since they were
discharged before five days, when a second chest CT scan would be performed, as per
the hospitals protocol. Consequently, the best responders to to proxalutamide were
probably selectively removed from analysis: and 2. Patients in the placebo group that
had worse progression of the COVID-19 were not able to undergo a second CT scan
because most of them needed ICU before five days, which precluded them from a
second CT scan. In this way, patients that had better responses to usual care were
probably selectively were included in the analysis. In short, this is an analysis that
compared the group that responded relatively worse to proxalutamide, and therefore
remained in the hospital for a longer period of time, with the group that presented a
better COVID-19 disease course and did not require ICU. This hypothesis is reinforced
by the fact that the median percentage of lung affected was 60.0% when all patients
from the placebo group were included, and 50.0% when only those patients with at least
two chest CT scans were included, while this difference did not occur in the
proxalutamide group. This means that patients that were worse tended to be excluded
from the placebo group. A significant improvement even under this conservative bias
reinforces the potential efficacy of the drug.
The imbalance between actives and placebos in present analysis is a result of a
conservative bias of the RCT: more actives were randomly designated to institutions
with fewer resources in rural areas, while more placebos were randomly designated to
hospital with better infra-structure. Since in-hospital mortality of COVID-19 is highly
variable and largely depends on the hospital resources23, the use of more actives in
institutions with fewer resources avoided overestimation of the drug efficacy. The
analysis was conducted as Intention-To-Treat (ITT), i.e., considering patients that
dropped out the study, which is another conservative bias, since the efficacy of
proxalutamide in hospitalized patients largely depended on a regular and uninterrupted
14-day treatment regimen. In fact, early discontinuation of the drug is highly
discouraged.
The radiologists that analyzed the chest CTs were experienced with quantifying the
percentage of lungs compromised by COVID-19 as the institutions for which they
worked has managed more than 20,000 cases of COVID-19,24 among which the vast
majority underwent chest CT scans. In addition, the correlation between the
quantification of COVID-19 lungs opacities by a board-certified radiologist experienced
with COVID-19 chest CT-scans and artificial intelligence (AI) is strong in the majority
of the cases.25-27 All these aspects reduce the possibility of operational-bias of the study.
The finding of radiological improvement was unexpected since the interval between the
baseline and the second chest CT scan was relatively short (median of five days in both
groups). We would expect that radiological changes would occur in the long-, not short-
run, as per the capacity of the disease resolution, even under effective therapies.12,20 We
hypothesize that this could be particularly true in our patient population, virtually solely
infected by the Variant of Concern (VOC) P.1, arguably one of the most pathogenic
SARS-CoV-2 variants described to date28.
In addition to the strong antiandrogen activity and to the ACE-2 antagonism, further
analyses demonstrated that proxalutamide may present direct protective actions in the
lungs and vessels,7-8 as well as anti-inflammatory effects, such as mitigation of tumor
necrosis factor alpha (TNF-alpha) and nuclear factor kappa beta (NF-kB).8 This may
explain the dramatic clinical and radiological improvements observed with
proxalutamide.
Limitations of the present analysis include the fact that radiological findings are not the
primary outcomes, with a subgroup of patients enrolled in three of the eight institutions
that participated in the RCT. with the inherent limitations of a post-hoc analysis of a
subgroup of patients of the RCT, despite the full representation of the group in the
subgroup analysis and balanced characteristics between proxalutamide and placebo
groups. 11 The present findings should be strengthened by further external analysis, in
particular including an analysis using AI.
To our knowledge, this is the first RCT that demonstrated radiological improvement in
response to a drug intervention in COVID-19. These findings reinforce the efficacy of
proxalutamide in hospitalized COVID-19 patients.
In conclusion, proxalutamide plus usual hospitalized care demonstrated to improve
hospitalized COVID-19 patients radiologically, when compared to placebo plus usual
hospitalized care.
REFERENCES
1. Wambier CG, Goren A, Vaño-Galván S, et al. Androgen sensitivity gateway to
COVID-19 disease severity. Drug Dev Res [Internet] 2020;81(7):771–6.
Available from: http://www.ncbi.nlm.nih.gov/pubmed/32412125
2. Hoffmann M, Kleine-Weber H, Schroeder S, et al. SARS-CoV-2 Cell Entry
Depends on ACE2 and TMPRSS2 and Is Blocked by a Clinically Proven
Protease Inhibitor. Cell [Internet] 2020;181(2):271-280.e8. Available from:
https://linkinghub.elsevier.com/retrieve/pii/S0092867420302294
3. Lin B, Ferguson C, White JT, et al. Prostate-localized and androgen-regulated
expression of the membrane-bound serine protease TMPRSS2. Cancer Res
[Internet] 1999;59(17):4180–4. Available from:
http://www.ncbi.nlm.nih.gov/pubmed/10485450
4. Wu S, Miao L, Zhou Q, et al. Suppression of Androgen Receptor (AR)-
ACE2/TMPRSS2 Axis by AR Antagonists May Be Therapeutically Beneficial
for Male COVID-19 Patients. SSRN Electron J 2020.
5. Qiao Y, Wang XM, Mannan R, et al. Targeting transcriptional regulation of
SARS-CoV-2 entry factors ACE2 and TMPRSS2. Proc Natl Acad Sci U S A
2020;118(1).
6. Li F, Han M, Dai P, et al. Distinct mechanisms for TMPRSS2 expression
explain organ-specific inhibition of SARS-CoV-2 infection by enzalutamide.
Nat Commun [Internet] 2021;12(1):866. Available from:
http://www.nature.com/articles/s41467-021-21171-x
7. Qu F, Gu Y, Wang Q, et al. Metabolomic profiling to evaluate the efficacy of
proxalutamide, a novel androgen receptor antagonist, in prostate cancer cells.
Invest New Drugs [Internet] 2020;38(5):1292–302. Available from:
http://www.ncbi.nlm.nih.gov/pubmed/32008178
8. Tran C, Ouk S, Clegg NJ, et al. Development of a second-generation
antiandrogen for treatment of advanced prostate cancer. Science [Internet]
2009;324(5928):787–90. Available from:
http://www.ncbi.nlm.nih.gov/pubmed/1939544
9. Cadegiani FA, McCoy J, Gustavo Wambier C, et al. Proxalutamide Significantly
Accelerates Viral Clearance and Reduces Time to Clinical Remission in Patients
with Mild to Moderate COVID-19: Results from a Randomized, Double-
Blinded, Placebo-Controlled Trial. Cureus [Internet] 2021;Available from:
https://www.cureus.com/articles/52299-proxalutamide-significantly-accelerates-
viral-clearance-and-reduces-time-to-clinical-remission-in-patients-with-mild-to-
moderate-covid-19-results-from-a-randomized-double-blinded-placebo-
controlled-trial
10. McCoy J, Goren A, Cadegiani FA, Vaño-Galván S, Kovacevic M, Fonseca DN,
Dorner E, Onety DC, Zimerman RA, Wambier CG. Proxalutamide (GT0918)
Reduces the Rate of Hospitalization for COVID-19 Male Outpatients: A
Randomized Double-Blinded Placebo-Controlled Trial. Front. Med. | doi:
10.3389/fmed.2021.668698
11. Unpublished.
12. Martínez Chamorro E, Díez Tascón A, Ibáñez Sanz L, Ossaba Vélez S, Borruel
Nacenta S. Radiologic diagnosis of patients with COVID-19. Diagnóstico
radiológico del paciente con COVID-19. Radiologia (Engl Ed). 2021;63(1):56-
73. doi:10.1016/j.rx.2020.11.001.
13. Pocock SJ, Assmann SE, Enos LE, Kasten LE. Subgroup analysis, covariate
adjustment and baseline comparisons in clinical trial reporting: current practice
and problems. Stat Med, 2002, vol. 21 (pg. 2917-30).
14. Sealed Envelope Ltd. Create a blocked randomisation list [Internet]. Simple
Randomiser. 2021 [cited 2021 Apr 16];Available from:
https://www.sealedenvelope.com/simple-randomiser/v1/lists
15. Y. Wang, C. Dong, Y. Hu, C. Li, Q. Ren, X. Zhang, et al. Temporal changes of
CT findings in 90 patients with COVID-19 pneumonia: a longitudinal study.
Radiology, 296 (2020), pp. E55-E64
16. N. Zhang, X. Xu, L.-Y. Zhou, G. Chen, Y. Li, H. Yin, et al. Clinical
characteristics and chest CT imaging features of critically ill COVID-19
patients. Eur Radiol, 30 (2020), pp. 6151-6160
17. W. Zhao, Z. Zhong, X. Xie, Q. Yu, J. Liu. Relation between chest CT findings
and clinical conditions of coronavirus disease (COVID-19) pneumonia: a
multicenter study. AJR Am J Roentgenol, 214 (2020), pp. 1072-1077.
18. C. Hani, N.H. Trieu, I. Saab, S. Dangeard, S. Bennani, G. Chassagnon, et al.
COVID-19 pneumonia: a review of typical CT findings and differential
diagnosis. Diagn Interv Imaging, 101 (2020), pp. 263-268.
19. J. Wang, Z. Xu, J. Wang, R. Feng, Y. An, W. Ao, et al. CT characteristics of
patients infected with 2019 novel coronavirus: association with clinical type.
Clin Radiol, 75 (2020), pp. 408-414.
20. M. Yuan, W. Yin, Z. Tao, W. Tan, Y. Hu. Association of radiologic findings
with mortality of patients infected with 2019 novel coronavirus in Wuhan,
China. PLoS One, 15 (2020), pp. e0230548.
21. M. Prokop, W. van Everdingen, T. van Rees Vellinga, J. Quarles van Ufford, L.
Stöger, L. Beenen, et al. CO-RADS–a categorical CT assessment scheme for
patients with suspected COVID-19: definition and evaluation. Radiology, 296
(2020), pp. E97-E104.
22. D'Agostino RB Sr, Massaro JM, Sullivan LM. Non-inferiority trials: design
concepts and issues - the encounters of academic consultants in statistics. Stat Med.
2003 Jan 30; 22(2):169-86.
23. Ranzani OT, Bastos LSL, Gelli JGM, Marchesi JF, Baião F, Hamacher S, Bozza
FA. Characterisation of the first 250,000 hospital admissions for COVID-19 in
Brazil: a retrospective analysis of nationwide data. Lancet Respir Med. 2021
Apr;9(4):407-418. doi: 10.1016/S2213-2600(20)30560-9. Epub 2021 Jan 15. PMID:
33460571; PMCID: PMC7834889.
24. Internet website: www.samel.com.br (Last access: June 20th, 2021).
25. Mader C, Bernatz S, Michalik S, Koch V, Martin SS, Mahmoudi S, Basten L,
Grünewald LD, Bucher A, Albrecht MH, Vogl TJ, Booz C. Quantification of
COVID-19 Opacities on Chest CT - Evaluation of a Fully Automatic AI-
approach to Noninvasively Differentiate Critical Versus Noncritical Patients.
Acad Radiol. 2021 Mar 6:S1076-6332(21)00099-4. doi:
10.1016/j.acra.2021.03.001. Epub ahead of print. PMID: 33741210; PMCID:
PMC7936551.
26. Shan F, Gao Y, Wang J, Shi W, Shi N, Han M, Xue Z, Shen D, Shi Y.
Abnormal lung quantification in chest CT images of COVID-19 patients with
deep learning and its application to severity prediction. Med Phys. 2021
Apr;48(4):1633-1645. doi: 10.1002/mp.14609. Epub 2021 Mar 9. PMID:
33225476; PMCID: PMC7753662.
27. Qiblawey Y, Tahir A, Chowdhury MEH, Khandakar A, Kiranyaz S, Rahman T,
Ibtehaz N, Mahmud S, Maadeed SA, Musharavati F, Ayari MA. Detection and
Severity Classification of COVID-19 in CT Images Using Deep Learning.
Diagnostics (Basel). 2021 May 17;11(5):893. doi:
10.3390/diagnostics11050893. PMID: 34067937; PMCID: PMC8155971.
28. Faria NR, Mellan TA, Whittaker C, Claro IM, Candido DDS, Mishra S, Crispim
MAE, Sales FC, Hawryluk I, McCrone JT, Hulswit RJG, Franco LAM,
Ramundo MS, de Jesus JG, Andrade PS, Coletti TM, Ferreira GM, Silva CAM,
Manuli ER, Pereira RHM, Peixoto PS, Kraemer MU, Gaburo N Jr, Camilo
CDC, Hoeltgebaum H, Souza WM, Rocha EC, de Souza LM, de Pinho MC,
Araujo LJT, Malta FSV, de Lima AB, Silva JDP, Zauli DAG, de S Ferreira AC,
Schnekenberg RP, Laydon DJ, Walker PGT, Schlüter HM, Dos Santos ALP,
Vidal MS, Del Caro VS, Filho RMF, Dos Santos HM, Aguiar RS, Modena JLP,
Nelson B, Hay JA, Monod M, Miscouridou X, Coupland H, Sonabend R,
Vollmer M, Gandy A, Suchard MA, Bowden TA, Pond SLK, Wu CH, Ratmann
O, Ferguson NM, Dye C, Loman NJ, Lemey P, Rambaut A, Fraiji NA, Carvalho
MDPSS, Pybus OG, Flaxman S, Bhatt S, Sabino EC. Genomics and
epidemiology of a novel SARS-CoV-2 lineage in Manaus, Brazil. medRxiv
[Preprint]. 2021 Mar 3:2021.02.26.21252554. doi:
10.1101/2021.02.26.21252554. Update in: Science. 2021 Apr 14;: PMID:
33688664; PMCID: PMC7941639.
FIGURE LEGEND.
Figure 1. Procotol flowchart – Analysis of the chest CT scans in the Proxa-Rescue
AndroCoV Trial.
Figure 2. Baseline and on-treatment chest computed tomography (CT) scans in
proxalutamide versus placebo.
