A review of seven clinical trials finds corticosteroids improve the chance of survival for COVID-19 patients.
pooled data from more than 1,700 patients from around the world, including a trial run in New Zealand. It is authored
by the World Health Organization's REACT working group, which includes a New Zealand scientist.The SMC asked experts to comment on the study.Dr Michael Maze, Respiratory and Infectious Disease Physician, University of Otago, comments:
"This publication provides valuable information for doctors treating COVID-19 patients. The study combines the
information from almost all the relevant clinical trials that have been investigating whether corticosteroid drugs help
people who are critically unwell with COVID-19 survive. They found that across the studies corticosteroids reduced the
chance of death by around one third.
"This is important, as it is a readily available treatment that can be used in New Zealand from today. We need to
remember that it is not a cure, and that even with corticosteroids, around 1 in every three patients enrolled still
died. We also need to remember that the study is only of patients who are critically unwell with COVID-19 – the patients
who are needing help with breathing on a ventilator, or who are otherwise extremely unwell.
"The existing information from studies including patients who are sick enough to be in hospital on oxygen, but not on a
ventilator, have shown a smaller benefit, and we don’t have data to suggest that corticosteroids help patients with
COVID-19 who are well enough to be at home."
No conflict of interest declared.
Christopher Gale, Clinical Lecturer, Department of Psychological Medicine, University of Otago Medical School, Dunedin,
"This paper will be read with interest by intensive care physicians and general physicians, and it could mean that a
corticosteroid becomes part of the management protocol for acute respiratory distress associated with COVID-19. However,
the results are skewed by one large study, and it would be hard to confidently choose one corticosteroid over another,
or low dose over high dose, because, as the authors note, the comparisons were not able to be precisely estimated.
"Data from seven trials was pooled with a total of 1703 critically ill patients from 13 countries including New Zealand.
One trial was excluded because it was published after the cut off date for the protocol, and one other trial was still
recruiting patients at the time of meta analysis. One study from Denmark contributed 1000 patients to the population.
"The meta analysis suggests that the use of corticosteroid did make a significant reduction in the 28 day mortality
rate. This seemed to be more apparent if one considered only those who had mechanical ventilation: 30% of those patients
receiving steroids died compared with 40% receiving usual care in the whole group. That is a difference of 10% which
implies you would need to treat 10 people to save one life. Among those who had mechanical ventilation 30% of those
receiving corticosteroids died compared with 38% for placebo. That is a small but significant absolute difference of 8%
which implies you would need to treat 12.5 people to save an extra life.
"The patients in this group are extremely sick. Corticosteroids make a small difference, but in people with respiratory
distress small differences matter. The number of adverse effects was lower among those treated with corticosteroids.
"Three of the studies used high doses of corticosteroids, four used lower doses. All studies allowed usual care for
patients in addition: Four studies recorded the use of antivirals, four different studies recorded the number of
participants with hydroxychloroquine, four papers recorded the use of Azathioprine, and one paper the use of
convalescent plasma. This variation in medications across studies, with people this sick, is not unusual, but is a
limitation on the analysis.
"There were no significant differences noted in the response rate by age, gender, or rapidity or number of days the
patient had symptoms for."
Disclaimer: I have worked on Cochrane protocols and meta analysis and on reviews of meta analysis for BMJ Clinical
Evidence and BMJ Best Practice. I am not an intensivist: I am reviewing the paper from an Evidence based Medicine
Professor Kurt Krause, Infectious Diseases Physician; Professor of Biochemistry, University of Otago, comments:
"This meta-analysis centres on the effect of corticosteroid administration to critically ill patients with COVID-19. To
complete this analysis they considered 7 randomised controlled trials involving 1703 patients. 1025 of these patients
received either placebo or usual treatment while 678 patients were randomised to receive either dexamethasone,
hydrocortisone or methylprednisolone. The primary outcome reported was all cause mortality at 28 days following
"The main overall finding was a significant reduction in mortality in severely ill patients receiving steroids (Odds
ratio 0.66). A positive benefit was found in patients who received dexamethasone and hydrocortisone but not in those who
received methylprednisolone, although patient numbers for the latter group were very small. There was no measurable
benefit found in administering corticosteroids to patients who were on medication to support their blood pressure. This
group of patients would be among the most critically ill, and high mortality in this group would be expected.
"Overall this study in mostly in line with data from the RECOVERY trial
released earlier this year showing a significant benefit from the use of dexamethasone in critically ill COVID-19
patients. This is not surprising especially given that the RECOVERY dataset was included in this meta-analysis.
Continued use of corticosteroids in critically ill patients with COVID-19 would be expected to continue following this
study. The results for patients requiring vasoactive medications will need to be followed up on. The results with
methylprednisolone involve a small number of patients and are not statistically significant."
No conflict of interest declared.