Scientists in the UK have discovered a cheap and available treatment for COVID-19 patients.
They have described the use of a common steroid drug as a "major breakthrough".
The drug, dexamethasone, reduced death rates by a third for patients on ventilators, and by a fifth for patients needing oxygen.
The results have been published from the Recovery trial which is assessing a number of different possible COVID-19 treatments.
It is thought the drug could have saved between 4,000 to 5,000 lives if it had been used earlier in the pandemic.
Martin Landray, an Oxford University professor who is co-leading the trial, said: "This is a result that shows that if patients who have COVID-19 and are on ventilators or are on oxygen are given dexamethasone, it will save lives, and it will do so at a remarkably low cost."
"It's been around for probably 60 years.
"It costs in the order of £5... for a complete course of treatment in the NHS, and substantially less - probably less than one dollar - in other parts of the world, for example in India."
His co-lead investigator, Peter Horby, said dexamethasone - a generic steroid widely used in other diseases to reduce inflammation - is "the only drug that's so far shown to reduce mortality - and it reduces it significantly".
"It is a major breakthrough," he said.
There are currently no approved treatments or vaccines for COVID-19.
More than 11,500 patients from 175 NHS hospitals have been enrolled on the Recovery trial since it was set up in March to test a range of potential coronavirus treatments.
In the dexamethasone study, 2,104 patients received 6mg of dexamethasone once a day by mouth or intravenous injection for 10 days.
Their outcomes were compared with a control group of 4,321 patients.
Over a 28-day period, the mortality rate among patients requiring ventilation was 41%, and for those needing oxygen it was 25%.
Among those not requiring respiratory intervention the figure was 13%.
While the study revealed the steroid reduced deaths in ventilated patients and in people needing oxygen, there was no change in deaths among patients who did not require respiratory support.