Another Harvard study, and COVID-19 immunity

It is also an MIT and MG, etc, study (https://www.medrxiv.org/content/10.1101/2020.08.23.20178236v1.full.pdf), but still.  It has a day and night difference from the other one mentioned on my June 10, 2020 blog.  It examines how COVID-19 outbreak occurred in MA.  Media like the headline that a Biogen meeting in late Feb contributed to 20K cases in MA, although the authors didn’t exactly say that in the manuscript (https://www.bostonglobe.com/2020/08/25/business/biogen-conference-likely-led-20000-covid-19-cases-boston-area-researchers-say/).  But they said a lot more else.

The manuscript is one of the few that looked at how COVID-19 started in various states by sequencing viral genomes, but it is the most complete story.  MA has world-class universities and hospitals well suited for the task.  So do CA and NY, but CA suffers from lack of access to patient info (doi: 10.1038/d41586-020-02478-z), while NY had so many initial cases that sampling was good enough to give an overall picture but not enough to pinpoint transmission chains (DOI: 10.1126/science.abc1917, 7/18/2020 blog).  Since MA cases were lower than NY or NYC, the authors were able to identify ~ 80 introductions to MA and compare them to the subsequent ~ 700 cases.  As another I-told-you-so, the MA results reaffirmed the assertions from my blogs (e.g., 7/18/2020) one more time: no evidence Chinese travelers causing lasting transmissions in the West, and that chances and superspreading play a major role in COVID-19.  

Regarding the first point, the authors wrote: “Most introductions of SARS-CoV-2 into MA occurred early in the pandemic, in March and early April, primarily from elsewhere in North America and from Europe”, and “Based on tMRCA estimates for major Boston-area clades, we do not find evidence of undetected, or “cryptic,” transmission before mid-February”.  These results agree with those in NY (DOI: 10.1126/science.abc1917, 7/18/2020 blog).  The earliest three MA cases were one from Wuhan on Jan 29, 2020, and two from Europe, but none had any decedents, indicating “quarantine and contact tracing efforts appear to have prevented spread from the first known introductions into MA”.  In other words, again, an infection doesn’t always spread, if people do the work.  

Similar studies and consistent results from WA, NY, MA and other places are actually truly amazing.  While data early on, even in Feb and Mar, already showed that Chinese nationals/tourists contributed little to sustained transmission in the West, one could still reasonably assume that they played a small but not insignificant role.  Yet all genome data, the gold standard, have found zero so far.  Not in NY.  Not in MA.  The Feb WA outbreak was not from the first WA case in Jan, either.  While one can always argue that some early imported cases were missed and might have infected others, (how do you prove a negative?) they clearly had not led to any lasting transmission.  So from a practical standpoint, they were as good as none.  

The Biogen international meeting in Boston on Feb 26-27 lead to > 90 known infections with a viral strain first identified in France.  The authors believe that the strain first appeared in Europe in mid Feb and got amplified in the US due to the Biogen meeting.  A large percentage of the MA genomes in the manuscripts evolved from this Biogen strain, hence the eye-catching 20K estimate.  The authors further proposed that it was “exported from Boston to several US states, including Virginia, North Carolina, and Texas, and to other countries, including Australia, Sweden, and Slovakia”.  Thus, an infection can stop, but can also superspread.  COVID-19 is super-dependent on superspreading. 

 

DNA doesn’t lie.  Ample data from genome analyses show that Chinese tourists didn’t seed any long-term infections in the West, where such studies have been performed.  Whenever and wherever early Chinese importations were identified, infections were all contained.  There wasn’t any evidence or suggestion of local transmission beginning until at least mid-late Feb in the US, while the Chinese ban was effective at least 2 weeks earlier.  This timeline already rules out the Chinese connection.  DNA sequencing further kills it.

Democrats are airing ads against Trump.  One shows on Feb 10 Trump saying that the virus will disappear, and China is working hard against it.  The irony is that Trump was not very wrong: the virus did mostly go away in China soon after, only not so in the US.  China was able to stem the tide domestically by Feb 20, so COVID-19 was not destined to be the destructive force it has become.  But perhaps due to bad luck, a few random superspreading events later, it got a hold in some places, then a pandemic.  Yet even with superspreading, it is not inevitable: see South Korea.   

Incidentally, this new Time article (https://www.yahoo.com/news/exclusive-chinese-scientist-sequenced-first-020759886.html), while still clinging to certain Western misinformation, like which doctors were “detained” and which “Chinese counterparts said” what, dispelled the major theme of the AP piece that China hid the viral sequence (https://www.yahoo.com/news/china-delayed-releasing-coronavirus-frustrating-040707689.html), while confirming my June 2, 2020 blog arguments.  Frankly, such nonsense is no longer worth refuting, except for recording history.

  

As daily confirmed cases in the US and Brazil have dropped, India has seen a surge for weeks.  But the surge is in large part due to increased testing, including Ab results, so it is often comparing oranges and apples.  India for sure will surpass Brazil in total case numbers.  In fact, India likely also leads the US in cases already, if there is a magical way to count everybody ever infected.  But the deaths in India are so low that either they are vastly under counted, or for various reasons COVID-19 just doesn’t affect Indians much.

Another major scientific news is that HK reported the first strong evidence that one can get COVID-19 twice.  The person in question recovered from COVID-19 in March/April.  He then went to Spain and UK in Aug.  When he returned and were tested at the airport, he was found to carry a viral strain different from the one he had before, very convincing data that he was re-infected.  It should be noted that he was/is asymptomatic.

There are a few things to take from this piece of news, most duly explained in concurrent media analyses.  The most urgent one is that what does it mean to the prevailing, hopeful immunity to COVID-19?  The short answer is not zero, but not much either.  Biology allows exceptions, due to stochasticity and population heterogeneity.  Even the best vaccine is not 100%.  We can all agree that re-infection is rare because only after 20 million infections worldwide, the first clear example of reinfection emerges.  And in animal studies with mice and monkeys the first infection always prevented a subsequent infection or reduced its severity.  On the other hand, all we know thus far doesn’t tell us how long the immunity lasts (https://www.the-scientist.com/news-opinion/cold-causing-coronaviruses-dont-seem-to-confer-lasting-immunity-67832?utm_campaign=TS_DAILY%20NEWSLETTER_2020&utm_medium=email&_hsmi=93878755&_hsenc=p2ANqtz-8I5-EIUaMRx_tv0ZfzjrFi80SGtq-UtCG29o0kM32GJ_ExqAsFygmzCGm2WcHZlWIANVvV4QpzT_IOk6uZOP5WXP7eKQ&utm_content=93878755&utm_source=hs_email).  This goes to the heart of vaccine usefulness, and only time will tell.  

Another thing is that the HK person was/is asymptomatic, so at least in this group of one, the first infection might have helped prevent/reduce pathogenesis from the second.  Perhaps the most important takeaway is that immunity works.  Or, what matters is not whether reinfection occurs, but whether a person can get sick/infectious twice or better than the first time.  As long as there is some protection from the first infection, it should be viewed as good.  

A side issue concerns asymptomatic, which has been explained extensively.  Most countries follow the WHO guideline which counts the asymptomatic in total cases.  China has the asymptomatic in a separate category: if he/she develops symptoms later, he/she will be counted as confirmed, otherwise the case is asymptomatic but not included in the total, confirmed cases.  Since all are medically monitored, there is no leakage, but a scientific case can be made whether an asymptomatic should be considered a patient or not.  There are reasons to include asymptomatic: their symptoms might have been too transient or mild to notice (not really asymptomatic), or they might still be infectious (evidence is, however, that the true asymptomatic don’t transmit much).   There are also precedents not to include: HPV, HBV, HCV, etc, have asymptomatic infections, who don’t need doctors or carry a tag.  

The recent episode in Xinjiang, China tells an interesting story, or experimentation in the sense that it was not the purpose but nevertheless offers a new perspective.  First discussed in my Aug 6, 2020 blog, Urumqi, the capital of Xinjiang, had an outbreak from mid July to mid Aug.  It semi-officially reportedly originated from a wedding.  The response in Urumqi is similar to that in Dalian, China, whose smaller outbreak happened around the same time, but differed from the previous ones in China in that Urumqi/Dalian tested the “focused” group repeatedly, defined as those in close contact with the infected, or simply living in the same communities.  In Beijing in June, for example, those people were tested only once, or twice.  But in July and Aug, many more people were tested 3X, 4X, or even more.  According to news reports, most initial Urumqi infections were young people, who were pre-symptomatic or asymptomatic.  They went home and infected family members later.  So only by saturated testing these young people, their families, and others several times, could Urumqi get a picture as complete as possible of extent of the outbreak.  At the end there were ~ 800 confirmed and ~ 200 asymptomatic cases.  No other cities in the world, with the exception of Dalian, can claim such thoroughness. 

There are multiple reasons to test to saturation.  Transmission takes time.  The infected may not have enough virus to be tested positive at first (https://www.yahoo.com/huffpost/coronavirus-testing-symptoms-doctor-130000666.html).   So to identify every infection, this is the only way.  China may want to do it also simply because it can.  The flipped side is it uses up a lot of materials and work.  But it constitutes a unique opportunity to study COVID-19 infections and an outbreak.  

Tellingly, when Urumqi started testing in mid July, the daily tallies included dozens of confirmed and dozens of asymptomatic.  On the next days, many of the confirmed cases were conversion of the asymptomatic from the previous day(s), indicating those “asymptomatic” were actually presymptomatic.  As days went by, testing still found asymptomatic, but the conversion ratios decreased.  By early Aug fewer confirmed and asymptomatic were found every day, and even lower conversions, and soon, only asymptomatic were reported.  By mid Aug, no more positive tests.  

These evolving numbers make perfect sense.  When you act fast enough, you will catch people who are infected but yet to exhibit symptoms, hence the higher conversion ratios early on.  As time goes by, and because you don’t test everybody every day, you will find confirmed cases more directly, but for the mild or asymptomatic cases, they can recover quickly or remain asymptomatic.  Since you quarantine, transmission stops, then after 3-4 weeks, no more confirmed cases, and only asymptomatic, who might have been mild but recovered cases, or asymptomatic throughout.   Due to saturated testing, one can calculate the asymptomatic ratio in a large population (million) as 200/(200+800)=20%.  Because some of the asymptomatic might have been self-reported or recovered, the true asymptomatic ratio of COVID-19 is likely less than 20%.  Again, no other estimates of asymptomatic in the world come close to the power of this Urumqi dataset, simply because of the latter’s sheer size.  Note that this 20% number is not affected by “confirmed” vs “actual” COVID-19 cases.  For example, a country confirms 300K cases, but the actual infections may be 2 million.  The 20% applies to both scenarios, to the first approximation.  

Of course, all these were from the news.  Only the doctors have the data regarding those positive individuals, their relationships, and their locations.  Mining such data, truly the most detailed and comprehensive so far, will tell us more about how COVID-19 transmits, when one develops symptoms after infection, how long a person remains infectious, and what counts as the asymptomatic.  

On the other hand, there remains a vexing question about how deadly COVID-19 is.  Since April mainland China has reported 0 death from COVID-19, after ~ 5000 cases.  That is, no person has died, or has a positive RNA test at death.  Maybe more young adults were infected, we know more about COVID-19, patients received care early, and the medical system is not overwhelmed.  A reduced death rate is also observed in the US in the summer.  In India, if we assume its infections lead the world, then unless its deaths are vastly undercounted (not much evidence to the extent), COVID-19 fatality rate is much lower than 1%.  If so, one should rethink what societal measures to tackle COVID-19.  

Back to the HK case.  Media often explain it from the angle of declining Ab levels in recovered COVID-19 patients, as if it indicates loss of immunity.  Such media thinking follows the results of studies in China, Spain, UK, and other countries since April.  While the implications are not necessarily wrong, they are not necessarily right, either, and they become more and more annoying over time.  Those studies reveal Ab levels do decrease, but they never say no more Ab, no more immunity, because humans have cell immunity too.  But the most important point missed by most reports is that declining specific Ab and immune cell levels are normal biological processes, which says nothing about immunity memory against a future infection.  Because producing/maintaining Ab and cell populations spends energy, if you no longer need them (you recover, and you can’t foretell when you will be re-infected), of course you will make less, so over time, Ab and cells will decline.  This pretty much happens to every disease and every vaccine.  What is important is whether once you encounter COVID-19 again, you will be able to produce the Ab and cells anew quick enough.  No studies have shown humans can or cannot.  Animals studies say they can, but for how long, nobody knows.  We just have to wait.  At the meantime, no more misleading.