Coronavirus: The Science

By Miles Henderson

Here is an in-depth breakdown of relevant research on COVID-19, and a straightforward explanation of how it applies to the situation in The Netherlands.

1. The Virus

2. Clinical Characteristics of COVID-19

3. Transmission Dynamics

4. Coronavirus in The Netherlands

Links to papers and data are in purple.


1. The Virus

Coronaviruses are a family of viruses that can infect avian and mammalian hosts: birds, bats, mice, dogs, camels, and humans. Each coronavirus is different. Each coronavirus is usually specific to one species, and two notable coronaviruses have infected humans in the past. The first being Severe Acute Respiratory Syndrome Coronavirus (SARS-CoV) in 2003, and MERS-CoV in 2012.

The terminology used to refer to the most recent outbreak of coronavirus is often used interchangeably and can be confusing.

The virus is officially referred to as SARS-CoV-2: Severe Acute Respiratory Syndrome Coronavirus 2. It is also referred to as 2019-nCoV: 2019 Novel Coronavirus. The disease the virus causes is COVID-19: Coronavirus Disease 2019.

Origin

SARS-CoV-2 did not emerge from the consumption of bat soup.

By looking at genetic data, it’s possible to compare SARS-CoV-2 genes to other known coronaviruses. Using this phylogenetic method, researchers have shown that SARS-CoV-2 likely originated from a bat coronavirus. Both SARS-CoV and MERS-CoV are also thought to have originated from bats.

A phylogenetic tree of SARS-CoV-2, similar to an ancestry tree, showing its high sequence similarity to a Bat Coronavirus. Notice that the first SARS-CoV as well as MERS-CoV share a common ancestor with a Bat Coronavirus.

So, how did the zoonotic transmission (viral infection that jumps from one species to another) of SARS-CoV-2 occur? It’s not entirely clear, however, it almost certainly did not occur from eating bats. Before the 2012 MERS-CoV jumped to humans, it was determined to have made a cross-species jump from bats to camels, which then infected human hosts in open markets.

Zoonoses have often been attributed to the crowded environment of open markets, where different animal species and humans interact in close proximity. In the case of SARS-CoV-2, it was quickly discovered that many of the first infected patients were exposed in a local seafood market in Wuhan.

Transmission

SARS-CoV-2 is an entirely new human virus, so there is no existing immunity in humans. It is primarily transmitted via respiratory droplets and direct contact: coughing, sneezing, hand-shakes. It is now known that the virus has extremely high transmissibility: it spreads extremely quickly without preventative measures.

The virus primarily infects the respiratory tract (lungs and airways) — but it can also be detected in the gastrointestinal tract, saliva, and urine. It is not yet known if faeces, saliva, and urine are also routes for transmission.

New research by the American National Institute of Health indicates that SARS-CoV-2 can remain viable (infectious/active) for different lengths of time, depending on the surface material it is found on: 4 hours on copper, 24 hours on cardboard, and 2-3 days on plastic and stainless steel. So far it is unclear to what extent that fomites (objects/materials that spread infection) contribute significantly to transmission of the virus.

It is possible that infected persons are infectious before symptoms appear. The latent period may be shorter than the incubation period.

The dynamics of transmission, including an explanation of the basic reproduction number and SIR epidemic models are discussed in part 3.

In The Human Body

What is known about SARS-CoV-2 in the human body? As of now there is robust data on the full genome of the virus. This means that the surface proteins of the virus are known, and the host cell receptor they bind has been determined to be ACE 2. To enter a human host cell, the virus binds this receptor. In a diagram, the viral surface proteins are called the “spike” proteins and look like a series of protruding crowns — which is where coronaviruses get their namesake.

Being able to access the full genome of the virus gives researchers important information for vaccine development, treatment, and diagnosis. Understanding the proteins that make up the SARS-CoV-2 virion (viral particle) will also allow researchers to design more effective vaccines and synthesise pharmaceuticals that target different stages of the virus “life cycle”.

Taken from https://doi.org/10.26434/chemrxiv.11728983.v1

Treatment and Vaccines

As of now, the various treatment options are all short-term and have limited application outside of severe and critical cases.

Some of the treatments that have been proposed are:

Chloroquine Phosphate — A well-established anti-malarial drug, recent papers suggest that it can also be effective against SARS-CoV-2 in cell cultures. The exact mechanism of its action is currently not well understood.

Nucleoside Analogs — Like us, SARS-CoV-2 needs building blocks to replicate its genetic code. Unlike us, SARS-CoV-2 is an RNA virus ( A, C, G, U) and not DNA based. Nucleoside analogs like Remdesivir mimic these building bocks but block the complete formation of viral RNA. Remdesivir has been shown to be effective in cell cultures, and has also been used in the treatment of Ebola.

Protease Inhibitors — Viral proteases make sure viral proteins are formed correctly. Strictly speaking, they “break-down” longer viral proteins into their final form. They also interfere with the host cell’s defence response. Effective protease inhibitors for the 2003 SARS Coronavirus have been developed in the past. Since the viral proteases of SARS-CoV-2 are 96% identical to those of the 2003 virus, it’s possible that old drugs can be repurposed.

Blood Donation (Convalescent Sera) This 20th century approach is quickly deployable and has a significant amount of data on its efficacy in situations without a vaccine. The technique uses blood from recovered patients to obtain Coronavirus antibodies produced by the patient’s immune system. These antibodies can then be injected into an actively infected patient.

Dozens of biopharmaceutical companies and academic institutions are currently developing or testing vaccine candidates. Preliminary research was focused on identifying potential epitopes (parts of the virus that can be recognise by our immune system). Now that the race to an effective vaccine has begun, it will take many months before a vaccine is made available. This is due to the need for exhaustive clinical testing to make sure that any vaccine is effective and, more importantly, safe.

ACE 2

To be explained


2. Clinical Characteristics

Symptoms

The most common symptoms are now well known by the public. Data from the Chinese Center of Disease Control and Prevention (CDCC): Fever (88.7% of infected), dry cough (67.8% of infected), and fatigue (38.1%). Rare symptoms are diarrhoea (3.8% of infected) and nausea/vomiting (5% of infected). A runny nose is unlikely to be a symptom.

However, the CDCC found that approximately 1% of those infected are asymptomatic: lacking common symptoms.

Another important point to note is that fever only presents in 43.8% of patients when they are admitted to hospital, this rises to 88.7% during extended hospitalisation. Therefore, an absence of fever does not rule out infection early on.

The median time between exposure and onset of symptoms is around 4 days. However, the incubation period can be between 2 to 14 days.

People at High Risk

Data from CDCC

DemographicCase-fatality Rate (%)
All2.3
Aged 80+14.8
Aged 70-798.0
Aged 60-693.6
Cardiovascular Disease10.5
Hypertension6.0
Diabetes7.3
Chronic Respiratory Disease6.3
Male2.8
Female1.7

Although it is now well known that the elderly are more susceptible to COVID-19, other demographics can be at higher risk as well.

COVID-19 is primarily a disease of the lungs and airways. People with existing respiratory disease are known to be at higher risk. However, those with existing conditions such as cardiovascular disease, crude hypertension (high blood pressure), and diabetes have also been shown to have a higher case-fatality rate.

People between 0 to 49 years old without any existing conditions have a very low case-fatality rate, up to a maximum of 0.4%. If you are young and healthy, you have a very low risk.

Unrepeated preliminary research indicates that cigarette smoking puts people at higher risk, and may account for the higher case-fatality rate of men (who are more likely to smoke).

Case Severity of COVID-19

A lot of concern has been placed on the number of infected people that will die from SARS-CoV-2. The overall case-fatality rate for COVID-19 was estimated to be 2.3%. However, this number could have been skewed higher as many infected were not diagnosed. Using data from the quarantined Diamond Princess cruise ship, where the exact number of infected is known, researchers have determined that the case-fatality rate is likely to be lower at 0.6%. This does leave quite a high range, but case-fatality is partly dependent on the environment — i.e., healthcare system or country. The true value is estimated to be somewhere between 0.25% and 3%. Although this number may seem small, the high transmissibility of SARS-CoV-2 means the this small percentage translates to a very high number in the world population.

A greater concern now is the proportion of infected that experience severe and critical symptoms. These patients must be hospitalised. This requires hospital beds, health care personnel, and respiratory support.

For severe symptoms, which include difficult breathing, the case percentage is estimated to be 13.8%. For critical symptoms, which includes respiratory failure and septic shock, the case percentage is estimated to be 4.7%. Half of patients with critical symptoms will succumb to the disease.

The consequences of this in The Netherlands will be further explained in parts 3 and 4.


3. Transmission Dynamics

What does “flatten the curve” mean?

It is expected that the majority of susceptible people in the world will be infected regardless of measures against transmission. These measures could be quarantines, city-wide lockdowns, or social distancing. Either way, people are going to get infected. The primary concern now is reducing the rate of new infections in order to keep daily infections within the capacity of our healthcare infrastructure.

Some may have heard that social distancing is necessary to “flatten the curve” and reduce the burden on our healthcare infrastructure. This is usually coupled with a quick diagram that shows a steep curve next to a “flatter” curve, where each curve represents a different scenario: if social distancing is either implemented or not. To understand what these curves really mean, it is helpful to see how epidemics are understood mathematically.

We’ll keep it straightforward and avoid the greek.

Basic Reproduction Number

The basic reproduction number (R0) gets mentioned a lot. Its working definition is: how many people will the first infected person infect?

If the R0 is 2 then that means the first infected person will infect two other people. A virus can only start an epidemic if this number is greater than 1, known as the epidemic threshold. If the R0 is less than one, that means the first person will not be able to infect enough people to sustain the infection — as a result the infection will fizzle out.

Taken from https://doi.org/10.1503/cmaj.090885

This number is very important for determining if an epidemic can occur, but it’s also crucial for determining how fast the virus might spread in a population. However, it is highly dependent on the context and environment. A virus may have different R0 in different settings — depending on healthcare, culture, or population dynamics. This is why the overall R0 is typically given as a range.

The R0 of SARS-CoV-2 during the early stages of the Wuhan outbreak was estimated to be around 2.68. If every infected person infects 2.68 people on average, that’s exponential growth. Other estimates of the R0 have broadened the range from around 2 to more than 6.

To put this into context, weighing the virus’s R0 against other infectious agents we can see that SARS-CoV-2 has a higher R0 than the 1918 Spanish Flu, but a lower R0 compared to Polio.

So, why is that this SARS-CoV-2 caused an epidemic but the fears about Ebola a few years ago never manifested into a serious epidemic outside Africa? Ignoring much of the detail and nuance: notice that Ebola’s overall R0 is hovering around 1-2. The R0 is very context dependent. If it dips below 1 in another population outside Africa then it can’t turn into an epidemic.

To make things more complicated. The reproduction number is not a constant. The basic reproduction number (R0) only tells us how many people the first infected person is expected to infect in a population. Later during the infection, the effective reproduction number (Rt) may change depending on if measures are implemented.

Taken from https://doi.org/10.1016/S1473-3099(20)30144-4

This image shows the estimated Rt throughout the outbreak in Wuhan. You can see that it starts at around 2.3 (basic reproduction number) at the very beginning of the outbreak and then evolves as the outbreak progresses. The red line marks the imposed lockdown of the city on 23 January, after which you can see that the Rt drops below 1 (the epidemic threshold) briefly.

SIR Epidemic Models

Without going into too much detail, a SIR epidemic model is one of the simplest ways to understand an epidemic. With a SIR model, it’s possible to estimate the R0 and predict the trajectory of the epidemic depending on the preventative measures taken.

To do this, the SIR model splits the human population into three smaller compartments of Susceptible (S), Infected (I), and Removed (R) people. The first two are fairly intuitive. “Removed” is comprised of people that are now immune after recovering from infection, or those that have succumbed to the infection.

People can only move in one direction from S to I to R. Think of it as people swimming in a big pool, then jumping into the hot tub, before plunging into a cold bath.

We can see here that at the start the population is completely susceptible (S), represented by the blue line. As more people become infected the number of people in the S pool move to the red infected (I) pool, until the S pool becomes completely exhausted. Those that recover or die then move into the green removed (R) pool.

What researchers take from the model are the rates in which people move from one pool to the next. The most of important of which is the infectivity parameter (transmission rate), this tells us how quickly people move from the susceptible (S) pool to the infected (I) hot tub.

Understanding how high the transmission rate is can tells us how hard we need to implement preventative measures. For example, an early estimate of these parameters using this type of model suggested that between 38% to 80% of transmission needs be prevented to bring the COVID-19 epidemic under control.

Notice that the red line looks a lot like the curve that needs to be “flattened”. That’s because it’s the very same one. The key is to make sure the daily number of infected cases, represented by the peak of the curve, does not exceed the capacity of the healthcare infrastructure.

The Curve

Now that we understand where the curve comes from. Here is that very same curve for the initial outbreak in Wuhan, China:

Notice how it grows exponentially at the beginning, with higher and higher cases each day. If no measures were implemented and the natural trajectory of the epidemic was allowed to progress, the curve would look something like the one in the SIR model shown above: much steeper with a higher peak number of infected patients per day.

Fortunately, on 23 January a complete lockdown was imposed on the city. This had the immediate effect of reducing the transmission rate and the number of new cases began to flatline. If the city had not implemented measures it’s possible that the high number of daily infected patients would have overwhelmed the additional health care infrastructure built.

Taken from https://doi.org/10.1016/S0140-6736(20)30260-9

This second figure elegantly demonstrates how the curve is “flattened” by reducing the rate of transmission. Reducing transmissibility by 25% would make sure the daily number of infected patients does not spike to unmanageable levels.

Time Lags

With all this data, it may seem like the situation is under control. If you’re reading this from a country with only a few hundred cases you might think that there is enough time to implement stricter measures.

Unfortunately, this is entirely dependent on the infrastructure available and the effectiveness of local testing in your area. To explain why, here is the second part of the graph shown above:

The blue bars show the date of symptom onset, whereas the orange bars show the date of diagnosis. The numbers that we are seeing in the news are a reflection of the orange bars — confirmed cases based on positive tests for SARS-CoV-2. The reason for the time lag between the two is due to the several day delay between first symptoms and a diagnosis.

If your country is currently at 1000 diagnosed cases, it is very likely that the actual number of infected people is much higher (in this graph it’s around 3000).


4. Coronavirus in The Netherlands

In Numbers

The Netherlands has seen an exponential growth in infected patients since its first diagnosed case on 27 February.

As of 16 March, the distribution of case severity appears to mirror that of the estimates in China — with around 3.2% (4.7% in China) of infected patients experiencing critical symptoms. Lacking data on the number of patients experiencing severe symptoms, it’s assumed that this will also mirror the distribution in China and be around 13.8%.

The graph above indicates the total diagnosed cases in The Netherlands according to data from the Rijksinstituut voor Volksgezondheid en Milieu (RIVM). Taking into account time lags in diagnosis, as discussed above, it is likely that the real number of infected is much higher.

Making a lot of simplifying assumptions, I’ve generated a curve of the projected number of real total cases in The Netherlands. Based on the data from the Wuhan outbreak, there was an approximately 7 day delay between onset of symptoms and diagnosis. By shifting the current data we have back 7 days, it’s possible to fit an exponential curve to essentially “nowcast” the number of real infections up to 19 March (shown in orange).

No countrywide measures were implemented until 13 March, when Prime Minister Mark Rutte first announced new restrictions. Further restrictions were applied on 15 March. So, it’s likely that the number of infections began to diverge from the predicted curve on these dates, which are marked with red vertical lines. Taking this into account, the true number of cases could be anywhere between 4,000 and 9,000 cases.

That’s a big range, and it comes with a lot of caveats. This simple projection assumes that there is an approximate 7 day delay in diagnosis, like in the city of Wuhan. It also assumes that the number of diagnosed cases today matches perfectly the number of people experiencing first symptoms 7 days ago. It also assumes that the rate of transmission is constant up until 13 March or 15 March.

Nevertheless, this is an interesting exercise in understanding the difficulty of evaluating the true number of cases in a country.

Critical Care Beds

As discussed above, a great concern is if the healthcare infrastructure is sufficient to support the number of severely and critically ill patients infected with COVID-19. A new paper by Harvard epidemiologist Marc Lipsitch warns that there may be shortages of intensive care hospital beds for hospitalised patients.

At the peak of the epidemic, Wuhan saw a hospitalisation rate of around 24 per 10,000 adults. For severely ill patients, this number was as high as 12 per 10,000 adults. For critically ill patients, this number was around 2.8 per 10,000 adults.

The following data from the Organisation for Economic Co-operation and Development (OECD) and shows the numbers of hospital beds in The Netherlands in 2017:

Bed TypeBeds Available per 10,000 peopleTotal Beds
Total Hospital Beds33.256,886
Curative (Acute) Care Beds29.250,047
Critical Care Beds0.87~1,500

If the peak of the COVID-19 epidemic in The Netherlands sees similar numbers to Wuhan (12 severely ill per 10,000 adults), then The Netherlands may not have enough critical care beds to accomodate newly hospitalised patients.

This does not take into account the number of additional ventilators and respiratory support equipment that may be needed.

On the other hand, this assumes that The Netherlands would have similar transmission dynamics to the city of Wuhan. This is not likely to be the case. Whereas Wuhan is a contiguous urban environment of 11 million people, The Netherlands has around 17 million residents spread out over multiple cities and rural environments.

Ultimately, the consequences of a late response to the transmission of SARS-CoV-2 will be seen in the coming weeks. The only way to minimise the burden on healthcare infrastructure at this point is to drastically reduce the rate of transmission.


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s