Disclaimer: Early release articles are not considered as final versions. Any changes will be reflected in the online version in the month the article is officially released.
Author affiliations: University of Texas at Austin Department of Integrative Biology, Austin, Texas, USA (K.E. Johnson, S. Woody, S.J. Fox, L. Ancel Meyers); Santa Fe Institute, Santa Fe, New Mexico, USA (M. Lachmann, L. Ancel Meyers); University of Texas at Austin Office of the Vice President for Research, Austin (J. Klima); University of Texas at Austin Department of Population Health, Austin (T.S. Hines)
Mutations in the B.1.1.7 spike protein result in a failure to detect the spike gene probe in standard SARS-CoV-2 quantitative reverse transcription PCR (qRT-PCR). In estimating the prevalence of B.1.1.7 from local quantitative PCR data, we initially assumed US estimates for the proportion of spike gene target failures (SGTF) attributable to B.1.1.7 (4) and, in our retrospective analysis, update that proportion on the basis of local sequencing data. We used a Bayesian model to estimate the local growth rate of B.1.1.7 among all SARS-CoV-2 infections and applied a compartmental susceptible-exposed-infected-recovered model of SARS-CoV-2 transmission to project the effect of B.1.1.7 on future COVID-19 prevalence.
We previously estimated that the relative frequency of B.1.1.7 among positive SARS-CoV-2 samples was growing logistically at a daily rate of 0.077 (95% CI 0.017–0.140), corresponding to an early doubling time of 9.0 days (95% CI 5.0–41.0 days) (K.E. Johnson et al., unpub. data, https://doi.org/10.1101/2021.03.05.21252541). At the time, we projected that B.1.1.7 would comprise most cases at UT by March 5 (95% predictive interval [PI] February 20–March 28) (Figure, panel A).
Subsequent estimates of B.1.1.7 prevalence based on quantitative PCR data from February 20 and April 9 fell within 95% PIs of the early projections (Figure, panel A) but suggested a lower daily growth rate of 0.037% (95% CI 0.026%–0.048%) and a corresponding doubling time of 18.7 days (95% CI 14.3–26.7 days). As of April 9, we estimated that B.1.1.7 comprised 61.2% (95% CI 48.5%–72.6%) of SARS-CoV-2 infections, consistent with our initial projections that B.1.1.7 would become the dominant variant by March 28 (95% CI March 20–pril 10) and that B.1.1.7 is 24% (95% CI 17%–32%) more transmissible than the wild-type virus.
Based on those local estimates, scenario-based projections suggested that B.1.1.7 might cause 6.2% (95% PI 3.7%–8.4%) more cumulative infections during April 9–May 23, 2021, than if it were not more transmissible than the wild-type virus (Figure, panel B). When we assume a higher published estimate for the relative transmissibility of B.1.1.7 of 59% (95% CI 56%–63%) (2), we projected that B.1.1.7 would increase overall incidence by 14.3% (95% CI 10.8%–18.0%) during this period (Figure, panel B; Appendix Figure 5). We provide projections as total infections, rather than hospitalizations or deaths, because the primary concerns of the university at the time of this analysis were anticipating increased demand for isolation facilities, testing, and contact tracing. In either scenario, if behavior remains constant for the remainder of the semester, then we would not expect B.1.1.7 to drive a major surge in infections in the university community during this period (Figure, panel B). The relatively small effect derives from 2 factors that constrained future growth of B.1.1.7. We estimated that, by April 9, 47% (95% CI 39%–57%) of the student community was immunized by prior infection (either virus type providing complete immunity) and that B.1.1.7 already comprised most (61.2%) new cases. This result hinges on the assumption that previous infection from either virus type confers immunity to both types and therefore would not apply to any type able to evade vaccine- or infection-acquired immunity. Our projections, which do not consider future behavioral change or reflect the full range of uncertainty, were not intended as forecasts but rather as plausible guideposts to help the university anticipate the severity of B.1.1.7.
UT surveillance testing indicates that B.1.1.7 rapidly became the dominant variant during the spring 2021 semester. Our methodology enabled rapid detection of B.1.1.7 emergence from widely available quantitative PCR data when sequence confirmation was not available or delayed, while quantifying uncertainty in the variant growth rate and fraction of SGTF samples that were positive for B.1.1.7. During January 16–March 5, UT confirmed 22 of 23 sequenced SGTF SARS-CoV-2 specimens as the B.1.1.7 variant, corroborating our reliance on SGTF data (Appendix).
Our findings reinforce the urgent need for expanded molecular surveillance capacity. In the absence of widespread and rapid sequencing efforts, quantitative PCR data from large-scale testing efforts have provided sentinel warning of B.1.1.7 emergence in cities throughout the United States.
Dr. Johnson is a postdoctoral researcher in the Department of Integrative Biology at the University of Texas at Austin. She uses mathematical models to aid in decision-making for control of infectious diseases. Dr. Woody is a postdoctoral researcher in the Department of Integrative Biology at the University of Texas at Austin. He develops mathematical and statistical models to elucidate transmission dynamics, surveillance, and control of infectious diseases.
The conclusions, findings, and opinions expressed by authors contributing to this journal do not necessarily reflect the official position of the U.S. Department of Health and Human Services, the Public Health Service, the Centers for Disease Control and Prevention, or the authors’ affiliated institutions. Use of trade names is for identification only and does not imply endorsement by any of the groups named above.