By Grant Jacobs 21/02/2021

Sometimes it’s smaller, intensive studies that shed light on issues. Just reported results of daily sampling of COVID-19 patients indicate patients with the B.1.1.7 variant first observed in Kent, UK may have a longer infection compared to patients infected with non-B.1.1.7 variants. This is the variant seen in NZ’s most recent outbreak in Auckland.

If this observation bears out in further studies this may affect how we manage B.1.1.7 cases. Longer lag periods would signal more caution in identifying and shutting down outbreaks; longer infections (and likely a longer infectious period) suggest there may be a need to isolate these patients for longer.

It’s a call for more data to check this out. The results are from a small sample—just seven B.1.1.7 cases—but may explain this variant’s tendency to be hard to put down.

We know that B.1.1.7 has more transmission – this result might suggest a little of why.

Sampling the NBA

The sample group are althetes from the USA’s National Basketball Association, the NBA. It’s perhaps the premier sports competition of the USA along with baseball, featuring millionaire players including New Zealand’s Steve Adams, a quiet high-flyer in a family that includes Olympic shot-putter Valerie Adams.

These basketball players are tested daily for COVID-19. As a result there is detailed data over the period of time of their infections, something there isn’t for most patients.

Of 65 players infected with SARS-CoV-2, the virus that causes COVID-19, seven (7) had the B.1.1.7 variant. The tests are from nasal swabs, tested using PCR.

Longer infections

Here’s a graph summarising what the research team found:


Looking at the graph, you can see how the peak of the infection shifts to later, with a slower start and longer duration.

This data suggests the variant B.1.1.7 may cause longer infections compared to non-B.1.1.7 variants, and that the increased transmissibility seen in B.1.1.7 cases may be because of the longer infection.

One way to think of this is that earlier results from examining large populations might sense more transmission overall, but it may be a case of more days of transmission rather than more each day of transmission.

(I’ve given the actual numbers in the final section, for those that like to geek over numbers!)

Where from here?

More samples needed!

One problem with a sample that has so few people is that it’s possible you’ve gotten a few people who are unusual, and who do not represent what is typical. In this case it’s complicated by that these are a very specific group of people: highly trained professional athletes. They tall, very fit and young.

This needs a larger sample, and across a wider range of people. These athletes will mostly be in their 20s: a wider range of ages will be needed to be sampled, too. Height or size might matter, too: it can affect how the body copes with an illness.

Maybe management needs to reflect the variant a patient has

Epidemic management teams around the world will be interested in this data, as it may affect how we should best manage B.1.1.7 cases. We might anticipate a slightly longer ‘lag’ period when a patient is infectious but not symptomatic. (Symptoms aren’t shown in the data, but we’d expect these to have a similar trend to the viral load per day.) We might also expect longer infections, and a longer time the person can infect others (infectious period).

This would mean we’d want more diligence on B.1.1.7 outbreaks, as a longer ‘lag’ means we’d get a more delayed notice of an outbreak by a day or so compared to other variants.

It might also want caution about how long a person can transmit the disease. A 14-day quarantine just covers the mean acute infection, rather than the full range. That suggests longer isolation might be needed for patients infected with this variant.

If it turns out that management needs to reflect what variant a patient has, we’ll want one of the PCR tests designed to test for this and other variants. (While New Zealand tries to sequence the genome of every positive patient, and that will continue, it’s more work, and needs better samples than PCR.)

Geeking the numbers

More formally for those that like to geek out on numbers –

Comparing side-by-side, athletes with B.1.1.7 variants compared to those with non-B.1.1.7 (and 90% credible intervals[1]):

Mean duration of proliferation phase:

5.3 days (2.7 – 7.8) v. 2.0 days (0.7 – 3.3)

Mean duration of the clearance phase:

8.0 days (6.1 – 9.9) v. 6.2 days (5.1 – 7.1)

Mean overall duration of infection (proliferation plus clearance):

13.3 days (10.1 – 16.5) v. 8.2 days (6.5 – 9.7)

Peak viral concentration (estimated from PCR cycle threshold, Ct[2]):

19.0 Ct (15.8 – 22.0) v. 20.2 Ct (19.0 – 21.4)

i.e. 8.5 log10 RNA copies/ml (7.6 – 9.4) v. 8.2 log10 RNA copies/ml (7.8 – 8.5)

(Tip: To keep it simple, worry more about the range of along the horizontal axes, than the height of the graphs per se. These are density curves, with the same area under each graph! All the horizontal axes are per day, except the first, which is by the amount of RNA in the mixture.)

Other articles in Code for life

For a sample list of earlier writing see the lists at the end of my 1,000th article milestone piece.

A few to explore:


Journalists: a plea. I’m a freelance science writer (and scientist). If your newspaper or magazine is interested in this story, or related stories, I would much prefer you let your editors know I’m available for covering science rather than have others write copy based on my work.

1. I’ve dispensed with the formal notation for ranges, as it’s confusing for non-scientists.

2. Cycle thresholds are the number of PCR cycles it took for the amount of RNA in the reaction mixture to rise about the level (threshold) that is considered to indicate a ‘signal’. Lower Ct numbers indicate there was lots of virus genome in the sample (it was easier to multiply up to the threshold amount from the sample, indicating the sample had a lot of virus genome to start with).

References & Sources

A copy of the abstract is available online.

A short twitter thread on the research has been posted by Yonatan Grad. You’ll see he calls out for feedback, and data from other patients who have been sampled frequently.

The research team are from several institutions including Harvard University’s T. H. Chan School of Public Health, the Yale School of Public Health, IQVIA, and associated researchers.

Featured image

Graph of duration of infection (cropped from copy of figure Yonatan Grad posted to Twitter).