Icon: enlarge
Image of an electron microscope of the Sars-CoV-2 coronavirus: Infectious mortality can be significantly influenced
Photo: narvikk / Getty Images / iStockphoto
"New WHO study astonishes even experts - the virus is really that deadly": Headlines like this one in the past few days have given the impression that the World Health Organization (WHO) now considers Covid-19 to be less dangerous.
But that is a mistake.
What is the study about?
The reason for the current discussion are the first research results of the renowned epidemiologist John Ioannidis from Stanford University, which have appeared on the website of the "Bulletin of the World Health Organization" - a scientific magazine published by the WHO.
The publication contains the note that the first online version of the study has been reviewed, but does not yet contain all of the final corrections.
For the analysis, Ioannidis had analyzed 61 studies from around the world as well as national data on the mortality rate from Covid-19.
According to this, the average mortality rate from all of the studies evaluated was around 0.23 percent.
This means that on average 23 out of 10,000 people who were infected with the coronavirus died.
"Inferred infection mortality rates tended to be lower than estimates made earlier in the pandemic," concluded Ioannidis.
First data from China had estimated the infection mortality at 3.4 percent.
Mathematical models have now assumed around one percent.
The only apparent contradiction
Some media then referred to the case mortality reported by the Robert Koch Institute (RKI), which, according to the daily management report on Monday, is around 2.7 percent - almost ten times as high as the value of Ioannidis.
This can give the impression of a scholarly dispute, which, however, does not exist at all.
Because the contradiction to the official RKI statistics is only apparent.
Two parameters can be used to measure the fatal potential of a disease:
The
infection mortality
, short IFR: It indicates what percentage of the deceased is considering all infections.
This value is also examined in the current study.
The
case fatality
shortly CFR: It indicates what percentage of the dead at all - and that one word makes all the difference -
known
infections.
The RKI refers to this value in the daily management report.
Since an infection with Sars-CoV-2 can proceed without symptoms and is by no means extensively tested in all countries, many infections worldwide are likely to go undetected.
As a result, case mortality is by definition higher than infection mortality.
Then why is it used at all?
It is currently difficult to say how high the number of unreported cases is.
According to the WHO, twenty times more people could have been infected worldwide than has been proven.
However, this number is also at best a rough estimate.
Antibodies against the virus in human blood could provide information about the number of unreported cases.
If they are found, the person has very likely gone through an infection - possibly without noticing it themselves.
In Germany, an antibody study by the RKI and the German Institute for Economic Research (DIW Berlin) started in October, in which 34,000 people are taking part.
It should clarify how many people in Germany have actually been infected with the virus.
Antibody studies of this kind are not perfect either, because apparently not everyone who can be shown to have been infected with the coronavirus also produces measurable antibodies.
But they are currently the best chance to shed light on the real figure.
"A death more or less that is recorded is really important"
However, antibody studies published to date are often based on low case numbers.
This can be seen in the so-called Heinsberg study, which has also been included in the current meta-analysis and puts the infection mortality at around 0.37 percent.
However, the extrapolation is only based on a handful of deaths in connection with a corona infection.
"A death more or less that is recorded is really important," said Gérard Krause from the Helmholtz Center for Infection Research (HZI) in Braunschweig after the Heinsberg study was published a good six months ago.
He therefore recommended investigating the cause of death more closely.
It is possible that the mortality rate from infections was underestimated because hardly any old people's homes were affected.
Almost a month ago, the epidemiologist Rod Jackson also emphasized in a guest article for the "NZ Herald" that antibody studies are only meaningful if the sample is representative of the population.
Evaluations from New Zealand and Iceland are not very helpful because only a few Covid 19 deaths have been reported there.
"Studies with fewer than several hundred Covid-19 deaths are hardly worth looking into," said Jackson.
In its current meta-analysis, Ioannidis considered studies with over 500 samples; not all of them - including the Heinsberg study - were representative of the entire population.
So has he worked scientifically inappropriately?
No fixed reading
The goal of Ioannidis was not to determine a globally valid infection mortality.
That is not even possible.
Infectious mortality is not a fixed measurement that, once calculated for a disease, remains valid around the world.
It depends on many factors.
In fact, according to initial data from China, at the beginning of the pandemic, mortality was initially estimated to be higher - as Ioannidis rightly notes.
Little was known about Covid-19 at the time.
It is now clear that a large proportion of those infected have only mild symptoms or no symptoms.
In addition, they can probably pass the virus on before they even notice the infection.
Ioannidis also emphasizes how much mortality varies - depending on whether older people in particular become infected, how good the medical care is, how high the proportion of people with previous illnesses in the population is.
In the studies he evaluated, the infection mortality fluctuated between zero and 1.63 percent.
The results of the representative antibody study planned in Germany are not yet available.
It is therefore difficult to estimate how high the current infection rate is in this country.
Just recently, a research team led by Andrew Levin from Dartmouth College published the results of another meta-analysis, according to which the infectious mortality rate for Covid-19 in the USA could be around 0.8 percent.
Virologist Christian Drosten also referred to this study in the NDR podcast.
Because the population in this country is a little older than in the USA, Drosten considers an infection mortality rate of around one percent for Germany or a little more plausible.
"I'm not a demographer, I can only estimate that," said the virologist, however.
Even if the infectious mortality in Germany is currently difficult to quantify precisely, it is clear how strongly it can be influenced.
This is also shown by the analysis from the USA.
If older people are protected against infection with Covid-19, the infection mortality can be more than halved - from 0.8 to 0.3 percent.
Icon: The mirror