April 2, 2020 - 2:57pm
Testing for COVID-19 is very much on people’s minds at the moment. I had an email from someone who works in a virology lab, who is doing a lot of testing for the disease at the moment, and wants to remain anonymous. I thought it was interesting; I don’t mean to suggest it’s the whole story, but I thought I’d pass it on.
Her lab does polymerase chain reaction (PCR) tests, the current standard. They work by finding a unique length of viral RNA, and multiplying it lots and lots. To extract the RNA, you need a specific chemical, called a reagent.
The pre-PCR extraction machine her lab uses to extract the RNA for the PCR process is made by the Swiss firm Roche. “It can test 96 samples in 90 minutes,” she says, “which is great for testing large numbers quickly.” Lots of other labs use it too.
But the reagents come in a specific kit that only Roche makes, and — unsurprisingly — Roche is unable to meet demand at the moment. We’ve heard that UK companies could make the required chemicals, but that’s not the whole story. “The chemical components of the reagent may well be very easy to make in the UK,” she says. “But the reagent has to be in a Roche kit for a Roche machine to take it.” It’s like saying you could use Hewlett Packard toner in a Canon printer — maybe you could, in theory (the inks might be the same), but the cartridge wouldn’t fit, so you can’t.
While her lab is waiting for more Roche supplies, they’re using other testing platforms, which can only do 24 or 48 samples at a time. “Some other labs are moving to slower platforms, like Qiagen,” she says, “which can do 96 samples [at a time], but take about 5 hours rather than 90 minutes.”
She speculates that, since Qiagen is a German company and German labs use a mix of Roche and Qiagen machines, that might explain why Germany is further ahead on testing.
I’m not suggesting this explains the whole problem, or that it means the government is blameless. If the UK virology industry is hugely over-reliant on Roche machines for testing, then I guess you could say that’s a foreseeable problem — but “the virology industry is small, and in normal non-pandemic times there is always a reason behind scientists picking the machines they pick, and doing things the way they do,” says my correspondent. This will surely be the biggest boom in virology-test demand in history by several orders of magnitude; I doubt any firm could keep up with it.
No doubt there are more things the government could be doing, and it’s important to keep pressing them for more openness. At the moment, though, I’m most interested in rolling out serology tests that could tell us who’s already had the disease. Then we can start answering the really tough questions.
Join the discussion
Join like minded readers that support our journalism by becoming a paid subscriber
To join the discussion in the comments, become a paid subscriber.
Join like minded readers that support our journalism, read unlimited articles and enjoy other subscriber-only benefits.
SubscribeThis sounds similiar to the problem Apollo 13 had with the square and round filters and the work around they had to come up with.
Your comparison to inkjet cartridges is an interesting one. The old cartridges can easily be refilled with new (non-genuine) ink, so I would have thought something similar could take place with refilling reagent into the Roche kit.
people have suggested this! I don’t know how possible it would be, but I can imagine (since false positives/false negatives could be so harmful) that people would be very wary of using them if they weren’t as reliable. But I have no idea if that’s the case
I ran clinical laboratories here in the U.S. for forty years, and participated in the advent of both automation of testing – including the Roche Cobas Bio – and PCR testing. I have personally performed proof-of-concept and routine testing of many clinical analyzers and many PCR reagent systems. The average layperson has no clue how complicated it is to design, manufacture, distribute and supply laboratory systems and reagents and to assure their reliability. Especially when it comes to very sophisticated tests like PCR. The inkjet cartridge analogy is germane, but understates the problem by orders of magnitude.
Clinical laboratory tests and equipment are not like soda-vending machines. They require trained, dedicated operators and a robust supply chain. Quality controls are crucial. Even then, there are limits to the testing, both in quantity and quality.
See the article in today’s Wall Street Journal about the (un)reliability of testing now being done worldwide. That will give you a glimpse into the reality. We are a long, long way from highly-sensitive, highly-specific viral testing of COVID-19 on anything like a high percentage of the population. And, from an epidemiological standpoint, it isn’t a panacea anyway.
Here’s an excellent, enlightening discussion of this topic:
https://www.hoover.org/rese…
Thank you for this article, Tom. It needs a wide audience but I doubt it will get one. Perhaps you could tweet the link to Peston?
I understand the reagents include enzymes, which are “grown”, for want of a better word, using bacteria or yeasts. You can’t hurry a yeast. I doubt the industrial chemicals industry would be much help with this.