Less than a fifth of deaths registered in the week ending May 29 in England and Wales involved coronavirus – the lowest proportion since the week lockdown was imposed, figures show.

There were 9,824 deaths registered in the week ending May 29 – a fall from the previous week but still 1,653 deaths higher than what would usually be expected, the Office for National Statistics (ONS) said.

Of these, 1,822 involved Covid-19 – 18.5% of the total that week and the lowest number of weekly coronavirus deaths for eight weeks.

It is also the first time the proportion of weekly coronavirus deaths has fallen to under a fifth since the week lockdown was imposed, the week ending March 27, when the virus accounted for 5% of the deaths.

While numbers are falling, there have been tens of thousands of “excess” deaths compared to the average number of deaths over five years for the same period.

The total number of excess deaths has passed 63,500, with Tuesday’s figures showing 57,961 excess deaths in England and Wales between March 21 and May 29 2020.

Added together with the numbers of excess deaths for Scotland and Northern Ireland published last week, the total number of excess deaths in the UK across this period now stands at 63,596.

All figures are based on death registrations.

In the week ending May 29, there were 819 more deaths in care homes compared with the five-year average, and 30 fewer deaths in hospitals.

Tuesday’s release takes the overall coronavirus death toll for the UK to 51,766, based on death certificates where coronavirus was mentioned and deaths of confirmed cases in hospitals.

This is more than 10,000 above the latest tally of related deaths calculated by the Department of Health & Social Care – 40,597 people who have died after testing positive for Covid-19.

Nick Stripe, head of health analysis at the ONS, said some deaths involving coronavirus in care homes “will have brought forward deaths that might otherwise have happened relatively soon”.

He tweeted: “We might expect deaths not involving Covid in care homes to fall below 5-yr avgs (average) in the next few weeks”.

Carl Heneghan, professor of evidence-based medicine at the University of Oxford, said Covid-19 is behaving in a way similar to seasonal flu in its disproportionate impact on the elderly.

Society will have to have “rational debates” about when is the least risky age to be infected with coronavirus if a vaccine does not materialise, he said.

He told a Science Media Centre briefing on the ONS figures: “It matters when you get infections in relation to your risk.

“I think that’s incredibly important, so we will start to have some rational debates, because many infections, we know if you get them when you’re very young you actually do very well, for instance chickenpox, you get that when you’re young and you confer immunity.

“So, if you took away the concept that there were no vaccinations, let’s park that aside and this becomes a circulating endemic infection, we will start to have to have a debate around when is it better to get this infection?

“If you look at the data, it is when you’re under 45. As you get older then it’s worse for you.”

More than a quarter (25.6%) of deaths in the North East that week involved Covid-19 – the highest proportion across all of England’s regions.

While there was a decrease in the number of deaths from all causes in hospitals, care homes, private homes and hospices, the proportion of hospital deaths involving coronavirus increased.

It rose from just over half (51%) in the week ending May 22 to 55.1% in the week ending May 29.

The number of coronavirus deaths occurring in care homes fell from 42.1% to 38.7% over the same period.

A separate study published today by the University of Manchester suggests that more than a fifth (21%) of excess deaths that had taken place in England and Wales by early May were not linked to Covid-19.

The authors say these deaths were “driven by the inability or reluctance of people to access health services for other health needs”.

The study found that the highest rate of non-Covid-19 excess deaths during this period was in the West Midlands – 26 per 100,000 population – followed by Eastern England (21 per 100,000) and north-west England and London (both 20 per 100,000).

The authors used ONS data going back to January 2010 to better reflect long-term trends in seasonality and mortality.