Minnesota's broad COVID-19 testing under microscope

Minnesota lab directors are defending their COVID-19 diagnostic tests amid criticism that they are too broad and find insignificant viral material in nasal and saliva samples from some people who aren't infectious.

Critics have questioned how all positive results can be treated the same when some require more rounds, called cycles, of testing and the amplifying of millions more copies of DNA to find genetic proof of the COVID-19 virus.

A Harvard University epidemiologist recently suggested that positives produced with more than 30 cycles are unlikely to find infectious patients, but a leader of Minnesota's public health lab said there is no convincing proof and that her lab is confident in the federally required cycle threshold (CT) of 38 for its COVID-19 test.

"It's a qualitative test — it's positive or negative," said Sara Vetter, interim assistant division director of Minnesota's public health laboratory, which has processed 60,000 COVID-19 tests.

"A [higher or lower] cycle threshold can indicate how much virus is in your sample, but we really don't know enough about what that means for infectivity" in patients.

PCR testing was a breakthrough three decades ago for its ultrasensitive detection of viruses such as measles and HIV by isolating and then replicating genetic material from patient samples. Among other things, it replaced the slow and even hazardous process of diagnosing infections by growing viral cultures themselves — which can happen only in high-security labs when dealing with more dangerous viruses such as SARS-CoV-2.

High-cycle thresholds and the potential detection of low virus levels in specimens has long been understood in PCR testing. Each cycle doubles the amplification of genetic material and increases the chance of finding low virus levels or even fragments of dead virus.

But the PCR test for COVID-19 has taken on heightened scrutiny because the screening of thousands of COVID-19 cases has resulted in financially crippling restrictions on businesses and lives.

Gov. Tim Walz ordered a 51-day state shutdown this spring in response to rising cases and deaths, followed by social distancing restrictions for businesses and an indoor mask-wearing mandate. The school reopening strategy this fall guided districts on whether to pursue online or in-class learning based on county case levels.

The state reported nine COVID-19 deaths and 929 newly confirmed infections on Saturday, bringing the total to 1,906 deaths and 83,588 infections. Seven of the newly reported deaths occurred in long-term care facilities. The count is based on nearly 1.7 million PCR tests of more than 1.2 million people.

The response to the pandemic may have been excessive if it turns out that many of Minnesota's lab-confirmed cases were never threats to spread the infection, wrote Kevin Roche, a former health care industry executive and blogger. "People could reasonably disagree about trying to pick up every possible infectious case vs. being more cautious, given the consequences, for example in regard to closing a school."

Criticism grew after Harvard's Dr. Michael Mina told the New York Times last month about his concerns over test results with cycle levels of 30 or more. He argued for lower cycle thresholds but increased and more rapid testing, including of asymptomatic people who can spread the virus without knowing it.

A Canadian study this spring underscored his concerns, because researchers for the most part could not grow viral cultures from samples in COVID-19 patients whose positive PCR tests required more than 25 cycles or whose symptoms had occurred more than seven days prior to testing.

The takeaway shouldn't be to reduce PCR testing or cycle thresholds, though, because all positive cases inform health officials as they conduct contact tracing and try to contain an outbreak, said Dr. Jared Bullard, a lead author and assistant medical director of the Cadham Provincial Laboratory in Winnipeg that conducts COVID-19 testing.

Cycle thresholds vary and are set by manufacturers based on the validated limits by which their tests are accurate.

The state lab uses a CDC-derived test with a threshold of 38, while a test derived at Mayo Clinic has a threshold of 35. Thresholds vary from 40 to 45 for three commercial platforms used by HealthPartners' lab in Eden Prairie and by the private labs with which the health system has contracts.

A threshold also is just a maximum, Vetter said. Among 300 positive tests over the past three months, the state lab found that roughly 47% required fewer than 30 cycles to find one of two required viral targets in samples, while 21% needed 30 to 35 cycles, and 24% needed 35 to 38 cycles.

Switching to a lower threshold raises the risk of false negative results and missing cases, which could then send infected people back into their communities to spread the virus. Lab directors also cited several reasons that tests might require more cycles to produce positives even in highly infectious patients, including that doctors or nurses might not acquire high quality specimens with high viral loads for analysis.

Samples taken by three nurses from one patient would likely result in different cycle levels to produce positives. A high cycle result also could reveal a patient at the beginning of a COVID-19 case who will become more infectious.

Adopting new cycle thresholds is premature considering that nobody knows how much virus is needed to cause infection, said Michael Osterholm, director of the University of Minnesota's Center for Infectious Disease Research and Policy.

The U center is part of a global effort to agree on a minimal viral load for SARS-CoV-2, which wasn't discovered until late 2019.

"You still want to know about people who are in those higher PCR levels because they are infected," he added. "The question is, what do you do about them if you can find a substantial correlation between infectiousness and the CT level?"

UpToDate, a source of medical practice recommendations, advises doctors not to use cycle threshold data because there is no proof that it can improve their clinical management of patients.

Doctors don't receive cycle thresholds with positive test results but can call labs to discuss how the data could affect complex cases, said Dr. Glen Hansen, medical director for clinical microbiology and molecular diagnostics at Hennepin Healthcare in Minneapolis.

If a recovering patient is being denied access to an assisted-living facility based on a positive follow-up test, he said, it could help to show that the test was produced with a high cycle count.

Even if high cycle levels correlate to low infectiousness, they need to be considered in the context of patients' circumstances, said Dr. Bobbi Pritt, who leads Mayo's clinical microbiology division.

"Someone who has been sick for three days after going in to get tested, well, would you disregard a positive at that point?" she said. "Would you feel comfortable putting that patient in a room with an 85-year-old woman?"

Pritt said high-cycle positives are rarer in initial tests and more common in follow-up tests after prolonged illnesses — revealing evidence of persistent or old, dead virus. Some patients have recovered but still test positive as much as 120 days later.

Pritt led proficiency testing by the College of American Pathologists to ensure PCR tests in U.S. labs reached proper results for COVID-19. Sending equivalent specimens to labs, the college found that almost all reached the correct positive or negative findings — but that cycle levels varied even when labs used the same testing platforms.

Hansen said the existing level of PCR testing is guiding the state's pandemic response because sudden increases in cases or test positivity reveal local hot spots of virus activity.

The cycle thresholds of the test results don't matter as much for that purpose, he said. "PCR testing, particularly when its done by hospital labs, remains the best weapon we have in the tracking of COVID."

Jeremy Olson • 612-673-7744