How many people have died from COVID-19?

Cause of death is never as simple as it sounds and SARS-CoV-2 makes it even less straightforward.

---

The COVID-19 pandemic has so far claimed more than 1.3 million lives, and those are just the deaths we know about.

Like the proverbial iceberg, the true scale of pandemic-related mortality is still hidden below the surface.

It’s the elderly woman who died of a heart attack because she was too afraid of COVID-19 to go to hospital. It’s the young father who died of an acute asthma exacerbation because there were no ambulances free to attend him. It’s the chronic kidney failure patient whose regular doctor was self-isolating with COVID-19. It’s the septicaemia missed because hospital staff were frantically running from one code blue to the next, as more and more patients were brought into the ED gasping for breath. It’s the woman murdered by her abusive partner because lockdowns took away her only chance to escape.

These deaths may not have occurred were it not for the physiological, societal, economic and institutional havoc that SARS-CoV-2 has wrought on humanity.

Equally, the pandemic may have actually saved lives: the teenager who didn’t take her car too fast around a bend after a few beers, the elderly man with COPD who didn’t pick up a respiratory infection from his grandchild, the construction worker who didn’t die from a site accident.

There is a time to be born, and a time to die. For potentially hundreds of thousands of people around the world, 2020 was not supposed to be their time to die. Now, their misfortunes are coming to light through the study of excess deaths.

Excess deaths – or excess mortality – describes the difference between how many deaths occurred over a specific period of time and the number of deaths that were expected to occur over that period based on historical trends.

Major events such as natural disasters, influenza epidemics, even recessions, are associated with excess deaths. But when observed deaths are compared with expected deaths around the world for 2020 and the COVID-19 pandemic, the difference is shocking.

A study published in Nature Medicine^[1] found that from mid-February to May 2020 across 19 European countries as well as Australia and New Zealand, an estimated 206,000 more people died than would have been expected had not SARS-CoV-2 reared its spiky head.

There was an overall 18% increase in mortality that shouldn’t have happened.
To put that into perspective, excess mortality in New York City during the first wave of the COVID-19 pandemic was higher even than what was seen during the 1918 H1N1 influenza pandemic.^[2]

Some countries have fared worse than others. Observed deaths in England and Wales during the pandemic’s first wave were 37% greater than expected, representing more than 100 additional deaths per 100,000 people. In Spain, deaths were 38% greater than expected. In the United States, where COVID-19 numbers are spiralling seemingly out of control, deaths were 44% greater than expected in April. ^[3] But around half the nations in the study, including Australia, had excess deaths below 5%.

The obvious conclusion is that these excess deaths were the result of SARS-CoV-2 infection, particularly in nations such as Spain and Italy, which experienced high infection rates and casualties in the first wave.

But it’s not that simple, says epidemiologist Gideon Meyerowitz-Katz from the University of Wollongong. “If there is a very large epidemic at the time, you can probably attribute most of those deaths to COVID in some sense,” he says. “But the challenge then comes, do you attribute the deaths directly to COVID – as in people getting sick and dying of COVID – or do you attribute the deaths to the burden on the healthcare system generally?”

According to the Nature Medicine study, the number of excess deaths from all causes across the 21 countries was actually 23% greater than the number of deaths that were recorded as being caused by COVID-19. Again, there was considerable variance: in Spain, all-cause excess deaths were 69% greater than those assigned to COVID-19, and in Italy they were 46% higher. Another study found that in the United States, only two-thirds of excess deaths were coded as COVID-19 deaths.^[4]

Even in Australia, data from the Australian Bureau of Statistics for the first seven months of 2020 reveals a small but noticeable spike in excess mortality in Australia around the last week of March – right around the peak of the first wave of COVID-19 infection. That week ending on 31 March, there were 2824 deaths recorded in Australia, 298 more than the average for that same week over the previous five years, and 188 more than the highest number of deaths recorded for that week during the previous five years. Yet that week only 11 people officially died of COVID-19.

So if people weren’t dying of COVID-19, what were they dying of?

One possibility is that they were in fact COVID-19 deaths, but just not recognised as such.

“Internationally, we’re seeing huge body of information being built up around comorbidities or pre-existing conditions that act as risk factors for COVID-19, and also the range of conditions that can actually be consequences of COVID-19,” says James Eynstone-Hinkins, director of mortality statistics at the Australian Bureau of Statistics. “Pneumonia and respiratory arrest are quite common but even in Australian data we are seeing numbers of acute cardiac complications, kidney complications, et cetera.”

There’s also the role that coagulopathies seem to play in some COVID-19 deaths: an autopsy study^[5] of 12 consecutive COVID-19 deaths in one German hospital found four had died from massive pulmonary embolism and another three cases were found to have fresh deep venous thrombosis. Eight cases had microvascular thromboemboli in their lungs. Other studies have pointed to a significantly higher risk of stroke among individuals with COVID-19 compared to those with influenza.^[6]

It raises the question of how many deaths from conditions such as stroke, pulmonary embolism, cardiovascular disease or kidney disease around that first peak might have actually been attributable to COVID-19.

Eynstone-Hinkins says the ABS can only work with what the attending doctor writes as the cause of death on the medical certificate.

“If, for instance, on a death certificate you have COVID-19 and pneumonia, and pneumonia is a consequence of COVID-19, then COVID-19 is the underlying cause of death,” Eynstone-Hinkins says. “In most cases where COVID-19 is in the sequence, it would be the thing that is the underlying cause.”

But that’s in Australia, which Meyerowitz-Katz says has been pretty good at picking up COVID-19 cases. “In New York City, where they were building emergency morgues, it’s very likely that a large number of people were not classified correctly.”

An article in JAMA ^[7] in June stressed the importance of proper certification of deaths from COVID-19, giving the US example of an elderly woman with dementia and stroke, whose death was originally certified as being from acute respiratory failure until post-mortem RT-PCR testing revealed she was SARS-CoV-2-positive.

But testing isn’t always possible. In India, where access to laboratory testing for SARS-CoV-2 is extremely limited – one report in the BMJ^[8] from July suggests there were just 8.55 tests per thousand people – there may not even be the capacity to test individuals for the virus.

Suspected COVID-19 cases in India may not have been reported as confirmed COVID-19 cases, which could also lead to undercounting of COVID-19 deaths.^[9]

A clue comes from figures on deaths from other conditions during the pandemic’s first wave. The ABS’s data from January to July 2020 shows above-average deaths from pneumonia and diabetes that both correspond to the peak of the first wave of COVID-19 in Australia. The numbers are small and in a nation the size of Australia, could simply represent random chance, but perhaps some of those pneumonia deaths were in fact COVID-19.

There is also growing evidence that individuals with diabetes – type 1 and type 2 – are at greater risk of more severe COVID-19, ¹⁰ which raises the question of whether some diabetes deaths were precipitated by exposure to SARS-CoV-2. The virus doesn’t even have to touch people directly to kill them. All it needs to do is to prevent them from getting necessary care for other health conditions.

A systematic review posted in the non-peer-reviewed preprint server MedRxiv ^[11] found that healthcare utilisation during the pandemic declined by around one-third globally. Healthcare visits were down by a median of 42%, admissions down by a median of 28%, diagnostic service use down by a median of 31%, and therapeutics down by a median of 30%.

One study across four New York hospitals^[12] found significant decreases in hospital admissions during the late March-early April peak of COVID-19 infections. And these weren’t for benign conditions: hospitalisation rates were down for conditions including septicaemia, heart failure, myocardial infarction, cerebral infarction, epilepsy and even appendicitis. These are not conditions that will go away if ignored.

Similarly, in New South Wales, the number of people presenting to the emergency department dropped by 40% by mid-April compared with mid-March, and urgent presentations were down by 14% compared with the same time in 2019.^[13] While numbers later picked up, they were still down by 9% in the last week of June compared with the same week in 2019.

Professor Christine Jenkins, head of the respiratory group at The George Institute for Global Health in Melbourne says the message that hospital services were straining under the load – particularly in the northern hemisphere – has had unintended and potentially deadly consequences.

“People being invited to stay away from the emergency department unless they’re critically ill, that probably encourages a kind of late-presentation phenomenon which for some things such as an evolving myocardial infarction or stroke, is catastrophic,” Professor Jenkins says. “The later you wait, the less chance you have of recovery.”

Even for those who do turn up, the dire state of the pandemic in countries such as the United States has meant that treatment may not be available.

Professor Krutika Kuppalli, Assistant Professor in the Division of Infectious Diseases at the Medical University of South Carolina, says an increasing number of hospitals in the US are now at full capacity.

“If you have a trauma, and you end up in the hospital, there’s no bed for you and there may not be an OR for you to go to if you need surgery,” Professor Kuppalli says.

The emergency departments are full, nurses who would normally be taking care of three patients are taking care of six. “Your medical care is going to be delayed, so all those things impact the care you’re going to get and unfortunately, the outcomes from everything are going to be affected.”

Many regions have also put elective surgery on hold, either to prevent an overloaded hospital network or in response to it. And that has a cost. “Those procedures are still important so people are not getting their surgeries, people are not getting their chemotherapy,” Professor Kuppalli says. “That leads to people being at risk of having complications from their diseases.”

The mortality picture during this pandemic isn’t entirely bleak. Public health measures such as stay-at-home orders, closing public venues, social distancing and hand hygiene have had benefits that extend well beyond reining in the spread of SARS-CoV-2.

“Since we’ve been doing all the physical distancing and hand hygiene and care over the way people gather together, there has been much less infection, and we know this is a true reflection of infection because so much testing has been happening,” Professor Jenkins says.

Influenza rates in particular have dropped through the floor: from January to June 2020, there were just 36 deaths from influenza compared with 430 over the same period in 2019, according to an ABC report.^[14]

Professor Jenkins says deaths from chronic respiratory conditions such as COPD are also way down in New South Wales compared with the five-year average. It’s a collateral benefit as measures designed to reduce the spread of one respiratory virus – SARS-CoV-2 – have also reduced the spread of a host of others.

“We know that about 60% of COPD exacerbations are initiated by viral infections and it’s almost certainly higher than that because we don’t test patients or by the time we test, the virus has disappeared,” she says.

Fewer respiratory infections mean fewer COPD exacerbations and fewer deaths, and this phenomenon is being observed around the world.

Professor Jenkins has long campaigned to study the effect of exactly these sorts of measures on COPD exacerbation rates, but the COVID-19 pandemic has provided the opportunity to see these interventions in action on a global scale.

And it’s not just COPD. Professor Jenkins suggests older people, and those with chronic health conditions which might put them at risk of severe COVID-19 outcomes, have especially responded to the public health messaging around COVID-19, and kept themselves isolated, away from SARS-CoV-2 and away from other pathogens at the same time.

Lockdowns have also benefited younger people in reducing the risk of accidents, injuries and assaults. The trauma centre at Sydney’s Westmead Hospital saw a 23%-34% decrease in monthly average trauma admissions during March and April 2020 compared with the same period in 2016 to 2019. ^[15] In particular, the average number of road traffic collisions dropped 40%-52% and falls dropped 13%-29%.

“Because of lockdowns, there are fewer accident deaths, which is actually fantastic for younger age groups who tend to be bear the brunt of those deaths,” Meyerowitz-Katz says.

The COVID-19 pandemic, how nations have responded to it, and what effect those responses have had will provide epidemiologists, statisticians and public health researchers with material to study for generations. What will they learn from these patterns of excess deaths around the world?

Professor Majid Ezzati, chair in global environmental health in the School of Public Health at Imperial College London and co-author of the global excess deaths study, says there have been two main takeaways from their research.

The first is that the timing of lockdown measures was critical: places that locked down sooner relative to where infection rates were at had fewer deaths.

The second more subtle factor is what Professor Ezzati describes as the “preparedness and resilience” of a nation’s health system. This wasn’t necessarily hospital capacity, but more the ability to identify and care for less severe cases in the community, without needing to bring them into the hospital system, and the ability to identify early those who do need more intense care and treat them accordingly.

“[It’s] the capacity of the health system to triage and care for people at the right spot to minimize harm,” Professor Ezzati says. He suggests that community-based care has particularly fallen victim to austerity measures in places such as the United Kingdom, so people haven’t been able to get the necessary care from their GP or their usual carer because the entire system has become overburdened.

Professor Ezzati believes that studies such as his can offer valuable insights into what community care pathways should look like, and how responsive they can be under extreme conditions, such as a pandemic, heatwave or earthquake: “health system preparedness and resilience that can manage the unusual as well as the usual.”

References:

1. https://www.cdc.gov/mmwr/volumes/69/wr/mm6942e2.htm?s_cid=mm6942e2_x
2. https://www.acpjournals.org/doi/10.7326/M20-2003
3. https://jamanetwork.com/journals/jamaneurology/fullarticle/2768098
4. https://www.cdc.gov/mmwr/volumes/69/wr/mm6942e2.htm?s_cid=mm6942e2_x
5. https://www.acpjournals.org/doi/10.7326/M20-2003
6. https://jamanetwork.com/journals/jamaneurology/fullarticle/2768098
7. https://jamanetwork.com/journals/jama/fullarticle/2767262
8. https://www.bmj.com/content/370/bmj.m2859
9. https://www.bmj.com/content/370/bmj.m2859
10. https://www.thelancet.com/journals/landia/article/PIIS2213-8587(20)30315-6/fulltext
11. https://www.medrxiv.org/content/10.1101/2020.10.26.20219352v1.full.pdf
12. https://jamanetwork.com/journals/jamainternalmedicine/fullarticle/2772351
13. https://www.bhi.nsw.gov.au/__data/assets/pdf_file/0008/603566/BHI_HQ_APR-JUN_2020_COVID-19_Supplement.pdf
14. https://www.abc.net.au/news/2020-07-23/coronavirus-restrictions-cause-flu-cases-to-drop-australia/12480190
15. https://onlinelibrary.wiley.com/doi/full/10.1111/ans.16307