Asymptomatic spread of Covid-19

There are many aspects of Covid-19 that remain poorly understood, and many plausible theories about Covid-19 that still require good proof.  The fact that Covid-19 is spread in large part by people with little to no symptoms, however, is not one of those things: this we know with reasonable certainty.  However, I find that some people still think that this is a questionable, or even a disproved theory.  I think this is usually simple ignorance, and that these people are getting their information from outlets that only report on studies when they think they can poke holes in them or when they think they challenge the standard narrative.  These outlets present a very one-sided, narrative driven view of the current state of knowledge on Covid-19.

Therefore I wanted to present a summary of the actual current state of knowledge, as I understand it.  This is going to be merely a summary, and I am only going to cite enough studies to illustrate the main lines of thought that support the theory that asymptomatic spread is a very important percentage of the total spread of the disease.

Asymptomatic vs. Presymptomatic

I should start by clearing up--again--a key terminology problem when talking about asymptomatic spread of Covid-19.  Namely, the difference between the terms "asymptomatic" and "presymptomatic".  The terms are not used consistently in the literature.  Some scientists reserve the word "asymptomatic" for only those people who catch a disease but never develop any symptoms of that disease, from start to finish.  Only the spread of the disease done by these specific people is then termed "asymptomatic spread".  If there is any spread of the disease done by people before they develop symptoms, but those people do develop symptoms eventually, that is called "presymptomatic spread."  However, not all scientists are so careful with the terms, and a lot of scientists talk about "asymptomatic spread", meaning any spread caused by people who don't know they are ill because they don't have any symptoms yet.  This would be the combination of any "true" asymptomatic spread (spread by people who never develop symptoms at any time) and the presymptomatic spread (spread by people who don't have symptoms at the time they are spreading).  And some scientists throw in here also spread caused by "pauci-symptomatic" people: people who have a few symptoms but think it's no big deal and that they can safely go about their normal business regardless.

In the case of Covid-19, when scientists talk about asymptomatic spread being a large component of infections, they are certainly talking in the second way, lumping together "true" asymptomatic with presymptomatic.  At this point, it is known that people who never develop symptoms can spread Covid, but I think it is generally accepted that this is less common and that the bulk of what is called "asymptomatic spread" is really "presymptomatic spread" or "pauci-symptomatic spread".

Regardless, the fact that truly asymptomatic spread is known it happen is one reason to continue using the term "asymptomatic spread" in the broader way, rather than calling it "presymptomatic spread".  "Presymptomatic" would be technically incorrect because this type of spread includes some "true" asymptomatic in it, while "asymptomatic spread" can be considered precisely correct since it is identifying spread which occurs in which there is not symptoms that could clue the spreader in to the fact that he is a danger to others at the time of the spread.

Anyway, this would all be pedantic term hairsplitting, if it weren't for the fact that I'm aware of multiple studies that have come out showing that "true" asymptomatic spread is quite rare, which were then picked up on by various people and waved around as examples that the concern over asymptomatic spread is overblown.  Aside from the fact that those studies are not conclusive and have good counter-studies that should be considered, the main problem is that those studies are only considering a small portion of what most scientists are talking about when they talk about "asymptomatic spread".

The bottom line: be careful about those terms and make sure you understand how it is being used in the context of a specific study.  I'm going to continue using the term "asymptomatic spread" in the *broader* sense for the rest of this piece, and I'll make sure to say "true asymptomatic spread" if I mean spread caused only by people who never develop symptoms.

The main lines of thinking that show asymptomatic transmission is a major problem

As I see it, there are three main types of studies that show that asymptomatic transmission of Covid-19 is a major problem.

1. Molecular biology studies that show the typical progression of viral load in patients infected with Covid-19.  By first principles, the time periods in which a patient is shedding the most virus from their upper respiratory tract should be the times when that patient is most infectious.  These studies have shown that viral load spikes strongly right around the time that symptoms first arise, so that total virus shed is usually split something like 50/50 (I think 40/60 is closer to the consensus on this) between the periods before and after symptoms start.  The clear implication of these studies is that most people infected with Covid-19 will be quite infectious for the period of a few days before any symptoms arise *and* into the period when just the hints of symptoms are arising--both times when people are quite likely to still be going around doing their normal routine.
2. Large scale epidemiological studies tracing the course of the epidemic over large populations have been able to establish some characteristics of the disease that must be true, or else the disease could not be spreading in the way in which it is observed to do on the large scale.  Here epidemic timeline studies are very important because you can track the timeline of how many people are hospitalized and when (on average) their symptoms started.  Reasoning backward to the number of people who must be getting infected to produce these amounts of hospitalizations, it becomes impossible to explain the numbers seen without a large reserve of "silent cases" spreading the disease without becoming visible in the official numbers at all.  These studies are backed up by other large-scale serological surveys which have proven that substantial portions of the population are being infected by Covid-19 without ever having been aware of symptoms.
3. In certain very controlled locations where there have been outbreaks, it has been possible to do very thorough contact tracing and determine (with more or less certainty depending on the study) who caught the disease from whom and when.  There have been quite a few of these case studies  at this point, and they have shown convincingly that asymptomatic spread is real and a very significant portion of the total spread of Covid-19.

Some examples of each of these study types

Following are some examples of each of these three kinds of studies.  I want to emphasize again that this is not a complete list; it is meant to be a representative sample only.

Studies on viral load:

Temporal dynamics in viral shedding and transmissibility of COVID-19

This was a key early study on viral shedding with Covid-19.  The important data is conveyed in figure 2, which I am reproducing here:

This shows that viral load, as measured in 92 patients starting from the first day they showed symptoms, only decreased over time.  In other words, the latest possible date of peak viral load was the day of first onset of symptoms.  They didn't have measurements before symptom onset, obviously, because they weren't measuring viral load in these people before they were identified as sick.

The inference from this is that viral load peaked no later than onset of symptoms, and probably somewhere to the left of where the charts begin, a day or two before onset of symptoms.  Combining this viral load data with contact tracing in the same region these patients were from, they concluded that 37% to 48% of transmission for this outbreak occurred pre-symptomatically.

Side Note: The narrow spike

There is something very important to understand about these charts; namely, that the Y-axis of these graphs (labeled "Ct value") is of a metric that is an exponential measurement.  "Ct value" refers to the number of times the sample has to be amplified before the virus genetic material could be detected.  That's why the numbers go down as they go up the Y-axis of the graph: the fewer times you have to amplify the signal, the more signal there is to begin with.

Because each increase in an amplification cycle is a multiplication of the genetic signal, for every one unit you go *down* as you move *up* the graph, you are essentially raising the signal to a power one more time, not just increasing it by one (really, increasing the exponent of an e^x factor, to be most precise).  So this is another way of saying that these charts are logarithmic, and that means they visually under-represent how much more viral load there is at the left-hand side of the graphs compared to to the right.

To illustrate this difference, I got hold of the data table for this paper, reproduced one of these graphs as a chart in excel, and then corrected the Y-axis values to indicate their true logarithmic nature.  This is what the viral load signal looks like if you make that correction:

From here you can see that almost all of the viral load is concentrated in the 3-4 days after the onset of symptoms, plus presumably a similarly narrow spike to the left of the chart before symptoms begin.  I believe that the profile of Covid-19's viral shedding explains a few key things that were previously puzzling to me before I saw this:

1. Covid-19 transmission has a propensity to happen in "super-spreader" events.  It's not spread evenly by everyone who gets it.  Rather, it appears that people who get it and are then unfortunate enough to be in a relatively crowded place during this period in which they are spreading the virus much more than otherwise are the primary spreaders of the disease.
2. Transmission among family members at home has been high, as one would expect, but not as high as I personally was expecting (I don't currently have the numbers at hand, but I was shocked at how low they were compared to my expectations).  Actually, what studies they have done in the West suggest that restaurants are a larger factor in disease spread than individual-to-individual spread in the family.  I think the "narrow spike" explains this.  When a person sheds virus liberally while they are sick for a week at at time, everyone in the home is practically guaranteed to get it.  However, if the sick person in the household sheds most of the virus in just a few days time, and they are out and about for half of that time period, then I can see that there is a chance that all of the other family members will luck out and miss catching the disease during its most contagious period.  This seems to be what is happening fairly often.
3. Some random sample population surveys have seemed to indicate that true asymptomatic people are extremely unlikely to spread Covid.  This was on the basis that when random people are sampled, and some who have had no symptoms are nevertheless found to test positive for the virus, the virus sampled from them was detected in a fragmented or low level state at which it was deemed very unlikely to be infectious.  But if there is only a short time span when you get Covid in which you are very infectious, and then a longer time period in which you are much less infectious but still clearing the virus from the system, this would make sense.  You would expect that most truly asymptomatic individuals you catch on the street would be currently non-infectious--but that wouldn't therefore allow you to judge that they had never been infectious.

Distribution of Transmission Potential During Nonsevere COVID-19 Illness

This was a study in Cleveland showing essentially the same distributions of viral load over time, though focused on nonsevere Covid.  From their conclusions: 

"This study shows that in otherwise healthy subjects with nonsevere illness from COVID-19, the SARS-CoV-2 viral load is very high within 2–3 days of onset of symptoms and falls rapidly by orders of magnitude within a few days . . . The extremely high viral loads within 2 or 3 days since onset of symptoms suggest that viral loads may be almost as high in the immediate presymptomatic period, suggesting substantial transmission potential in the presymptomatic period. "

Epidemiological studies:

Serial interval of novel coronavirus (COVID-19) infections

This was an important early study based on the Wuhan infection timeline.  You can look at a timeline of an epidemic and plot how fast new infections are occurring compared to the number of existing infections, and from that deduce (with good statistics) how long it takes (in the average case) between a person getting infected and that person passing along the infection to the next person.  This is called the "serial interval" of a disease.  In the case of Covid, when they did this they found a serial interval of about 4 days.  The incubation period (the time after someone is infected but before they start showing symptoms) is a full 3-7 days.  This implies that a lot of spread must be happening before symptoms show up.

Quote from the paper:

"The serial interval of COVID-19 is close to or shorter than its median incubation period. This suggests that a substantial proportion of secondary transmission may occur prior to illness onset."

Asymptomatic Cases, the Hidden Challenge in Predicting COVID-19 Caseload Increases

This was a large random-sample study in Ontario showed that large amounts of asymptomatic infections (16-30% of all infected people, depending on age category).  Further, 47% of people who tested positive on a PCR at what was considered infectious levels were asymptomatic at the time.

Prevalence of SARS-CoV-2 in Spain (ENE-COVID): a nationwide, population-based seroepidemiological study

This was a very large study in Spain that just verifies how there is a high rate of asymptomatic infections overall: 22-35% of all infections, in this case.

Substantial undocumented infection facilitates the rapid dissemination of novel coronavirus (SARS-CoV-2)

Very thorough statistical analysis revealed that undocumented infections had to account for the bulk (~80%) of the spread in China.  Now, "undocumented" isn't the same as "asymptomatic", but at the least this proves that the greatest part of the spread of the disease happens from people who at least don't feel sick enough to come forward for medical care.

Contact tracing studies:

Now we can look at a series of case studies in which authorities had high control over patient movement and could decisively track actual point-to-point infections.  This has been the case in contained environments (nursing homes, prisons, etc.) and also in some nations with a much less exalted view of personal privacy rights regarding medical data than we have!

Contact Tracing Assessment of COVID-19 Transmission Dynamics in Taiwan and Risk at Different Exposure Periods Before and After Symptom Onset

Taiwan locked things down very rapidly at the start of the pandemic, and used aggressive contact tracing to identify infections early and get people into isolation.  The conclusion from this study was that to get things under control, you need to assume that spread starts 4 days before symptom onset:

  "The findings of this study suggest that most transmission of COVID-19 occurred at the very early stage of the disease or even before the onset of symptoms"  

In addition, they mention in the "Limitations" section of the study that they might have actually "underestimated the importance of early transmission" in the study for various reasons.

Mass SARS-CoV-2 Testing in a Dormitory-Style Correctional Facility in Arkansas

Study in Arkansas correctional facility which reached the conclusion: 

"If testing remained limited to symptomatic individuals, fewer cases would have been detected or detection would have been delayed, allowing transmission to continue."

Presymptomatic SARS-CoV-2 Infections and Transmission in a Skilled Nursing Facility

Study on an outbreak in a nursing home.  Conclusion: 

"More than half of residents with positive test results were asymptomatic at the time of testing and most likely contributed to transmission. Infection-control strategies focused solely on symptomatic residents were not sufficient to prevent transmission after SARS-CoV-2 introduction into this facility."

Rapid asymptomatic transmission of COVID-19 during the incubation period demonstrating strong infectivity in a cluster of youngsters aged 16-23 years outside Wuhan and characteristics of young patients with COVID-19: A prospective contact-tracing study

Case study of transmission within a cluster of 8 young people (16-23 years old).  Serial interval was just 1(!) day.  All ended up with just mild symptoms.  Demonstrates rapid transmission before any symptoms arise, and from people with only mild Covid (frequently unreported or not identified as Covid).

Presumed Asymptomatic Carrier Transmission of COVID-19

Case study of "true" asymptomatic spread in China.  One completely asymptomatic woman infected 5 family members.  So yes, this does happen as well.

The high prevalence of asymptomatic SARS-CoV-2 infection reveals the silent spread of COVID-19

Case study in Bahrain.  320 infectors and 1289 people infected.  In this study, they found that the asymptomatic infectors spread the disease at just as high a rate as the symptomatic infectors. 

"In conclusion, these data show that the high asymptomatic incidence of SARS-CoV-2 infection in Bahrain and subsequent positive contacts from an index case are more likely to be asymptomatic, showing the high “silent” risk of transmission and the need for comprehensive screening for each positive infection to help halt the ongoing pandemic."

Personal Anecdotes

Here, I want to add a final category, which is not a study but which is more easily accessible to most people.  At this point, there have been enough people who have actually had Covid that it is not difficult to do a little bit of "gum-shoe" detective work on your own.  Just ask the people who have had it whether they were aware of being around anyone sick before catching it, or if they are pretty sure they passed it along to someone else despite not showing any symptoms at the time of their known contact.

I have two siblings who have had it now, one of whom was quite careful of whom she was around before she became infected.  She's also pretty sure she infected someone else, despite not having any symptoms at the time they were in contact.

Likewise, my wife has two siblings who have been infected now, both of whom were being *extremely* cautious to avoid associating with anyone exhibiting any symptoms.  They say that anecdotes don't add up to data, but as I said, these anecdotes are multiplied millions-fold at this point and are fairly easy to access.

Conclusion

Again I stress, this is only a lazily compiled sample of the studies out there that all point in the same direction.  I spent very little time looking for studies beyond the ones with which I was already familiar, and I'm positive I could have easily doubled this list without effort.  The conclusion is clear: large amounts of asymptomatic spread of Covid-19 is now a firmly established fact and should be treated as such.

A secondary conclusion would be, there are better places to get information about Covid-19 than the blogs, social media, and news sites some people are relying upon.  It is a mystery to me why some of these outlets persist in fighting this specific point, given how much clear evidence is in its favor.

In the case of this article, (and previous posts), I have found that searching this online database of publications is very helpful:

Global literature on coronavirus disease

I would recommend spending some time searching through and reading relevant studies published in this database if you care to be informed about the best state of knowledge on Covid-19 currently.