More contagious variant is spreading fast in US, even as overall cases level off

The New York Times
Washington, DC, United States of AmericaWritten By: Lauren Leatherby and Scott Reinhard © 2021 The New York Times CompanyUpdated: Mar 11, 2021, 02:42 PM IST


Story highlights

There is not yet enough genomic sequencing — the process required to screen positive coronavirus samples for variants — to be certain of how widely that variant, known as B.1.1.7, is spreading. But data suggests its share of total cases is growing at a trajectory similar to that seen in countries where it has fueled surges.

As U.S. coronavirus cases remain at a low not seen since October, a more contagious variant first reported in the United Kingdom has likely grown to account for more than 20% of new U.S. cases as of this week, according to an analysis of data from Helix, a lab testing company.

There is not yet enough genomic sequencing — the process required to screen positive coronavirus samples for variants — to be certain of how widely that variant, known as B.1.1.7, is spreading. But data suggests its share of total cases is growing at a trajectory similar to that seen in countries where it has fueled surges.

Still, experts note that low total case counts in states with a high share of B.1.1.7 are an encouraging sign. It remains to be seen, they say, whether the variant will cause a significant resurgence here or whether widespread vaccinations and virus-control measures can keep case counts at bay.

The B.1.1.7 variant is just one variant of concern found in the United States, in addition to a variant first identified in Brazil and another discovered in South Africa. But those variants make up a tiny fraction of total cases compared with B.1.1.7, which experts say is likely to become dominant across the country this month. The variant is doubling as a share of all new U.S. cases approximately every 10 days, continuing a pattern first identified by researchers in early February.

Screening for variants requires sending a positive sample to a lab to be sequenced, which takes time and resources. The Centers for Disease Control and Prevention says it often takes about 10 days to sequence a sample and analyze it for reporting. Estimates for the cost of sequencing a sample range from $25 to $400, according to a report from the Johns Hopkins Center for Health Security. And the large amount of data each sequence generates requires computational and data science expertise that many health departments struggle to afford.

Only 0.5% of cumulative cases in the country have been sequenced since the beginning of the pandemic, but, with new efforts, about 1% of new cases in February and about 3% in the past two weeks were sequenced, according to CDC data and Jasmine Reed, a spokeswoman for the agency.

Experts’ recommendations on how many cases should be sequenced vary, but many say sequencing around 5% of cases would begin to adequately show the spread of variants circulating at low levels.

The United Kingdom, by contrast, has sequenced about 8% of its total cases and is currently doing so for about a third of its cases, according to its health secretary. It has about as many new cases per day as Florida. Denmark, another world leader in coronavirus genomic surveillance, had sequenced about 12% of its total cases as of January. It now sequences almost all new positive cases, which are now at daily levels comparable to Mississippi or Missouri.

The share of cumulative cases sequenced varies widely by state, from as high as 3.5% in Hawaii to 0.05% in Tennessee. Some states have had an advantage: It’s easier to sequence a higher share of cases if the total number of cases is lower. But those vast differences in genomic surveillance mean it is especially difficult to estimate how widely a given variant is spreading in states where there are not many cases being monitored.

According to estimates based on the Helix data, there were most likely tens of thousands of cases attributed to B.1.1.7 in the United States in the past week alone. But because sequencing is so low, only 2,672 cases have been officially recorded.

Helix has analyzed nearly half a million positive samples for an anomaly indicative of the B.1.1.7 variant. Helix then sent samples to Illumina, a gene-sequencing company, to have their genomes sequenced to confirm the variant. Based on results from those samples and the share of cases with the anomaly, Helix is able to make inferences about how widely B.1.1.7 might be spreading.

The relative share of the variant grew exponentially over the past several weeks in almost every state where Helix had enough data for a trend to be evident. In Florida, the share of cases caused by the B.1.1.7 variant was highest: well over an estimated 30% of cases. Nationwide, the figure is probably more than 20%, athough Helix’s testing is not representative of population distribution.

But there could be some good news: Even Florida, where the variant is spreading most widely as a share of total cases, has not yet seen a resurgence in total new coronavirus cases.

“I am encouraged by the declining case counts in the most heavily affected states,” said Caitlin Rivers, an public health researcher at Johns Hopkins University. “I’ve been watching Florida closely, which has the highest share of B.1.1.7. Case counts have plateaued there in recent days but are not resurging. The longer we can hold the line, the more time we have to roll out vaccines, which will protect individuals, particularly those at highest risk of severe illness, and slow transmission overall.”

B.1.1.7 is not the only variant of concern in the United States, but it has by far the biggest foothold. According to the CDC, which is tracking confirmed cases of the variants, there are more than 30 times as many cases of B.1.1.7 as of the other two variants of concern first reported in South Africa (called B.1.351) and Brazil (called P.1). So far, those other two have been found in a minority of states, but given the low rate of sequencing in many states, they are most likely spreading undetected in others.

“The other variants are circulating at very low levels now compared to B.1.1.7, but over the long run it will be important to watch those because they have the potential to make our vaccines less effective,” Rivers said.

“Although total cases have been going down in recent weeks, B.1.1.7 cases have not, which means that it may not be surprising to see cases start rising again in regions that have high levels of B.1.1.7 transmission,” said William Lee, vice president of science at Helix.

But even if another resurgence in cases begins, Lee says, it is reasonable to hope the situation would be different from when the United Kingdom faced a sharp spike around the holidays. “I think even if cases start going up again, the impact on hospitalizations and mortality may still be mitigated by vaccinations and higher levels of natural immunity than we had in the past.”


Helix identifies and reports a testing anomaly indicative of the B.1.1.7 variant called S gene target failure (SGTF). Estimates of cases attributable to B.1.1.7 were calculated following the methodology in a preprint using Helix’s data. The share of cases with SGTF was multiplied by the share of observed B.1.1.7 sequences in the SGTF samples Helix sent to be sequenced. For some later dates, if the share of observed B.1.1.7 was not available, the share of cases with SGTF was multiplied by the most recent moving average for which data was available.