[Annals of Internal Medicine, 18 July 2023]

Regarding their trial, Loeb and colleagues (1) state, “the overall estimates rule out a doubling in hazard”; however, no such conclusion is possible. The trial used an extraordinary threshold (a hazard ratio of 2, or a 100% relative increase in risk) for noninferiority and was underpowered to find smaller but still important risks. (We estimate that a 4-fold increase in sample size would have been needed to identify a 50% increase in relative hazard.) Power aside, design flaws biased the study toward the null result that was obtained.

The intervention under study was incorrect use of N95 respirators—intermittently rather than continuously. SARS-CoV-2 is an airborne pathogen (2). Infection occurs via inhalation of shared air, and infective aerosols accumulate over time in closed indoor settings. As such, only continuous use of N95 respirators protects health care workers against respiratory infection; intermittent use of medical masks and respirators is equally ineffective (3). Unplanned crossover (those randomly assigned to medical masks could reassign themselves to the N95 group on the basis of unrecorded risk assessment) and contamination due to failure to use a cluster design further biased study results toward the null (4).

Notwithstanding lack of power and multiple biases—and although only 21 infections developed among 301 participants recruited in Canada and Israel through May 2021—analysis according to the registered protocol reveals a doubling of risk for infection for medical masks (relative risk, 2.05 [95% CI, 0.85 to 4.95]; P = 0.10) in these participants. The study had come close to showing inferiority after recruiting only a fraction of its prespecified sample size. Around this time, the authors recalculated their required sample size (in July 2021) as 1010 participants and began recruiting participants in Pakistan at a site not mentioned in the trial's registered protocol. Six months later, recruitment in Pakistan was discontinued and was begun in Egypt (also not registered in the protocol). Final results were heavily influenced by the inclusion of the sites in Egypt, with more than 70% of infections originating there. As Altman and associates note, “when authors substitute other outcomes after the trial has started there must be concern that such changes were done with knowledge of the data. That casts doubt on the reliability and integrity of the results” (5).

Lastly, the performance of this trial lacked equipoise in the face of clear engineering evidence of the superiority of respirators for airborne pathogens. The fact that this trial was done in a flawed manner that could not provide valid results means that participants were endangered for no reason.

[The Lancet Respiratory Medicine, December 2022]

Background: The large number of patients worldwide infected with the SARS-CoV-2 virus has overwhelmed health-care systems globally. The Anti-Coronavirus Therapies (ACT) outpatient trial aimed to evaluate anti-inflammatory therapy with colchicine and antithrombotic therapy with aspirin for prevention of disease progression in community patients with COVID-19.

Methods: The ACT outpatient, open-label, 2 × 2 factorial, randomised, controlled trial, was done at 48 clinical sites in 11 countries. Patients in the community aged 30 years and older with symptomatic, laboratory confirmed COVID-19 who were within 7 days of diagnosis and at high risk of disease progression were randomly assigned (1:1) to receive colchicine 0·6 mg twice daily for 3 days and then 0·6 mg once daily for 25 days versus usual care, and in a second (1:1) randomisation to receive aspirin 100 mg once daily for 28 days versus usual care. Investigators and patients were not masked to treatment allocation. The primary outcome was assessed at 45 days in the intention-to-treat population; for the colchicine randomisation it was hospitalisation or death, and for the aspirin randomisation it was major thrombosis, hospitalisation, or death. The ACT outpatient trial is registered at ClinicalTrials.gov, NCT04324463 and is ongoing.

Findings: Between Aug 27, 2020, and Feb 10, 2022, 3917 patients were randomly assigned to colchicine or control and to aspirin or control; after excluding 36 patients due to administrative reasons 3881 individuals were included in the analysis (n=1939 colchicine vs n=1942 control; n=1945 aspirin vs 1936 control). Follow-up was more than 99% complete. Overall event rates were 5 (0·1%) of 3881 for major thrombosis, 123 (3·2%) of 3881 for hospitalisation, and 23 (0·6%) of 3881 for death; 66 (3·4%) of 1939 patients allocated to colchicine and 65 (3·3%) of 1942 patients allocated to control experienced hospitalisation or death (hazard ratio [HR] 1·02, 95% CI 0·72–1·43, p=0·93); and 59 (3·0%) of 1945 of patients allocated to aspirin and 73 (3·8%) of 1936 patients allocated to control experienced major thrombosis, hospitalisation, or death (HR 0·80, 95% CI 0·57–1·13, p=0·21). Results for the primary outcome were consistent in all prespecified subgroups, including according to baseline vaccination status, timing of randomisation in relation to onset of symptoms (post-hoc analysis), and timing of enrolment according to the phase of the pandemic (post-hoc analysis). There were more serious adverse events with colchicine than with control (34 patients [1·8%] of 1939 vs 27 [1·4%] of 1942) but none in either group that led to discontinuation of study interventions. There was no increase in serious adverse events with aspirin versus control (31 [1·6%] vs 31 [1·6%]) and none that led to discontinuation of study interventions.

Interpretation: The results provide no support for the use of colchicine or aspirin to prevent disease progression or death in outpatients with COVID-19.

[The Lancet Respiratory Medicine, December 2022]

Background: The large number of patients worldwide infected with the SARS-CoV-2 virus has overwhelmed health-care systems globally. The Anti-Coronavirus Therapies (ACT) outpatient trial aimed to evaluate anti-inflammatory therapy with colchicine and antithrombotic therapy with aspirin for prevention of disease progression in community patients with COVID-19.

Methods: The ACT outpatient, open-label, 2 × 2 factorial, randomised, controlled trial, was done at 48 clinical sites in 11 countries. Patients in the community aged 30 years and older with symptomatic, laboratory confirmed COVID-19 who were within 7 days of diagnosis and at high risk of disease progression were randomly assigned (1:1) to receive colchicine 0·6 mg twice daily for 3 days and then 0·6 mg once daily for 25 days versus usual care, and in a second (1:1) randomisation to receive aspirin 100 mg once daily for 28 days versus usual care. Investigators and patients were not masked to treatment allocation. The primary outcome was assessed at 45 days in the intention-to-treat population; for the colchicine randomisation it was hospitalisation or death, and for the aspirin randomisation it was major thrombosis, hospitalisation, or death. The ACT outpatient trial is registered at ClinicalTrials.gov, NCT04324463 and is ongoing.

Findings: Between Aug 27, 2020, and Feb 10, 2022, 3917 patients were randomly assigned to colchicine or control and to aspirin or control; after excluding 36 patients due to administrative reasons 3881 individuals were included in the analysis (n=1939 colchicine vs n=1942 control; n=1945 aspirin vs 1936 control). Follow-up was more than 99% complete. Overall event rates were 5 (0·1%) of 3881 for major thrombosis, 123 (3·2%) of 3881 for hospitalisation, and 23 (0·6%) of 3881 for death; 66 (3·4%) of 1939 patients allocated to colchicine and 65 (3·3%) of 1942 patients allocated to control experienced hospitalisation or death (hazard ratio [HR] 1·02, 95% CI 0·72–1·43, p=0·93); and 59 (3·0%) of 1945 of patients allocated to aspirin and 73 (3·8%) of 1936 patients allocated to control experienced major thrombosis, hospitalisation, or death (HR 0·80, 95% CI 0·57–1·13, p=0·21). Results for the primary outcome were consistent in all prespecified subgroups, including according to baseline vaccination status, timing of randomisation in relation to onset of symptoms (post-hoc analysis), and timing of enrolment according to the phase of the pandemic (post-hoc analysis). There were more serious adverse events with colchicine than with control (34 patients [1·8%] of 1939 vs 27 [1·4%] of 1942) but none in either group that led to discontinuation of study interventions. There was no increase in serious adverse events with aspirin versus control (31 [1·6%] vs 31 [1·6%]) and none that led to discontinuation of study interventions.

Interpretation: The results provide no support for the use of colchicine or aspirin to prevent disease progression or death in outpatients with COVID-19.

[PLoS ONE, 9 September 2022]

Importance: The measurement of laboratory biomarkers plays a critical role in managing patients with COVID-19. However, to date most systematic reviews examining the association between laboratory biomarkers and mortality in hospitalized patients early in the pandemic focused on small sets of biomarkers, did not account for multiple studies including patients within the same institutions during overlapping timeframes, and did not include a significant number of studies conducted in countries other than China.

Objective: To provide a comprehensive summary and an evidence map examining the relationship between a wide range of laboratory biomarkers and mortality among patients hospitalized with COVID-19 during the early phase of the pandemic in multiple countries.

Evidence review: MEDLINE, EMBASE, and Web of Science were searched from Dec 2019 to March 9, 2021. A total of 14,049 studies were identified and screened independently by two raters; data was extracted by a single rater and verified by a second. Quality was assessed using the Joanna Briggs Institute (JBI) Case Series Critical Appraisal tool. To allow comparison across biomarkers, standardized mean differences (SMD) were used to quantify the relationship between laboratory biomarkers and hospital mortality. Meta-regression was conducted to account for clustering within institutions and countries.

Results: Our systematic review included 94 case-series studies from 30 countries. Across all biomarkers, the largest and most precise SMDs were observed for cardiac (troponin (1.03 (95% CI 0.86 to 1.21)), and BNP/NT-proBNP (0.93 (0.52 to 1.34)), inflammatory (IL-6 (0.97 (0.67 to 1.28) and Neutrophil-to-lymphocyte ratio (0.94 (0.59 to 1.29)), and renal biomarkers (blood urea nitrogen (1.01 (0.79 to 1.23)) and estimated glomerular filtration rate (-0.96 (-1.42 to -0.50)). There was heterogeneity for most biomarkers across countries with studies conducted in China generally having larger effect sizes.

Conclusions and relevance: The results of this study provide an early pandemic summary of the relationship between biomarkers and mortality in hospitalized patients. We found our estimated ESs were generally attenuated compared to previous systematic reviews which predominantly included studies conducted in China. Despite using sophisticated methodology to examine studies across countries, heterogeneity in reporting of case-series studies early in the pandemic limits clinical interpretability.