Transmission from vaccinated individuals in a large SARS-CoV-2 Delta variant outbreak

Abstract

An outbreak of over 1,000 COVID-19 cases in Provincetown, Massachusetts (MA), in July 2021-the first large outbreak mostly in vaccinated individuals in the US-prompted a comprehensive public health response, motivating changes to national masking recommendations and raising questions about infection and transmission among vaccinated individuals. To address these questions, we combined viral genomic and epidemiological data from 467 individuals, including 40% of outbreak-associated cases. The Delta variant accounted for 99% of cases in this dataset; it was introduced from at least 40 sources, but 83% of cases derived from a single source, likely through transmission across multiple settings over a short time rather than a single event. Genomic and epidemiological data supported multiple transmissions of Delta from and between fully vaccinated individuals. However, despite its magnitude, the outbreak had limited onward impact in MA and the US overall, likely due to high vaccination rates and a robust public health response.

Keywords: Delta; SARS-CoV-2; contact tracing; genomics; transmission; vaccination.

Graphical abstract

Figure 1

Epidemiology of the COVID-19 Provincetown outbreak and overview of the dataset (A) SARS-CoV-2 genomes in this dataset by collection date. (B) Distribution of all cases in Barnstable County (gray bars) and estimate of reproductive number (R_t) (red line) over the course of the outbreak. (C) Flow diagram of sample sets presented here. Gray shading indicates sample sets reported and analyzed in this study. (D) Cycle threshold (Ct) value of the N1 gene for the 313 individuals known to be fully vaccinated by BNT162b2, mRNA-1273, or AD26.CoV2.S; includes linear regression with 95% confidence interval. All Ct values were collected using the same platform and assay. See also Figure S1.

Figure S1

Symptoms, Ct, age, and vaccination status, related to Figure 1 and STAR Methods (A) Ct values in outbreak-associated cases, including cases without genomes (465 individuals passing Ct thresholds). In individuals with multiple samples, the earliest collected sample was used. The presence or absence of symptoms was known for 263 individuals; of these, vaccination status was known for 251. Partially vaccinated individuals were excluded from the analysis at right. In each distribution, the mean is shown by a red line; the mean ± one standard deviation is shown by dashed red lines. All Ct values were collected using the same platform and assay. (B) Ct values by age and vaccination status, and age distributions by vaccination status and vaccine brand in 465 outbreak-associated cases passing Ct thresholds. In individuals with multiple samples, the earliest collected sample was used. Individuals with unknown vaccination status and partially vaccinated individuals are excluded from all six panels. Vaccination status was known for 355 individuals; of these, all had a known age and 290 were known to be fully vaccinated by one of BNT162b2, mRNA-1273, or Ad26.CoV2.S. In each distribution, the mean is shown by a red line; the mean ± one standard deviation is shown by dashed red lines. Scatterplot includes linear regression with 95% confidence interval.

Figure 2

Genomic epidemiology and limited onward impact of the Provincetown outbreak (A) Time tree of SARS-CoV-2 outbreak-associated genomes (blue dots) in the global context. Inset, frequencies of Pango lineages among outbreak-associated genomes. An interactive version of this tree is available at: https://auspice.broadinstitute.org/sars-cov-2/ma-delta/20211005/cluster-unique-usher. (B) Percentage of all Delta-lineage genomes from MA detected by baseline genomic surveillance with the mutational signature of the dominant outbreak cluster (red line). Percentages are calculated and shown per day based on sample collection date. Dashed line: cumulative new cases in MA over the same period (available at: https://www.mass.gov/info-details/covid-19-response-reporting). (C) Numbers of genomes (left) and percentage of Delta-lineage genomes (right) per state that are estimated to descend from the largest cluster in the Provincetown outbreak. Barplots show lower-bound (solid) and upper-bound (faded) estimates, calculated as described in STAR Methods. Note that the scale of the x axis in the plot on the right is in percentages, with a maximum of 1%. See also Figures S2 and S4; Table S1.

Figure S2

Genomic analysis of Delta lineages and mutations in Massachusetts, related to Figure 2 (A) The proportion by epidemiological week of Delta-lineage sequences among all publicly available baseline surveillance data from MA. Data shown are only those generated by the Clinical Research Sequencing Platform and Viral Genomics Group at the Broad Institute. Error bars denote 95% confidence intervals. (B) The frequency of the 50 most common consensus-level mutations among all outbreak-associated genomes (blue) compared with the proportion of Delta genomes in GISAID with the same mutation (gray). All AY.25 genomes had an amino acid change at position E239Q in ORF3a; however, although rare among publicly available Delta genomes, E239Q is shared across the AY.25 lineage and is of no known functional significance.

Figure S3

Within-host variation in outbreak-associated samples, related to STAR Methods (A) Total number of iSNVs per individual grouped by vaccination status. (B) The number of observations of each iSNV across all samples. iSNVs are labeled by their gene and amino acid change (if nonsynonymous) or nucleotide position (if synonymous). Bars are colored by the vaccination status of each individual in which a mutation was observed.

Figure S4

Outbreak introductions and onward spread, related to Figure 2 (A) Left: time tree of outbreak-associated SARS-CoV-2 genomes in a global context colored by association with the Provincetown outbreak (as in Figure 2A). Right: each introduction into Provincetown as inferred from the phylogenetic tree based on a change in ancestral inference of a branch to “outbreak-associated.” Gray dots represent the most recent common ancestor of the clade that was inferred to be from outside of Provincetown. All outbreak-associated samples downstream of each node are shown in the full phylogenetic tree. (B) The percentage of all Delta-lineage baseline surveillance genomes from MA with the mutational signature of the five Delta-lineage clusters among outbreak-associated cases. Three mutations (C8752T, C20451T, and A26759G) are shared by the majority of all outbreak-associated genomes, referred to as the dominant cluster. The remaining four clusters were defined by their characteristic mutations, with cluster 2 defined by G4124A and A5608G; cluster 3 by T7858C and A29257C; cluster 4 by T23131C; and cluster 5 by A26759G, C20451T, C7600T, and T27940C. Percentages per day based on sample collection date and new cumulative cases in MA were plotted over time from July 3rd to August 31st, 2021.

Figure 3

Identification of putative transmission events, including from and between vaccinated individuals (A) High-confidence transmission links (among the 467 cases with genomes) from contact tracing investigations alone, prior to the incorporation of genomic data. (B) Predicted transmission links based on genome sequence, intrahost variants, and symptom onset date. Three transmission links identified by contact tracing but not strongly supported by genomic predictions are shown as dots without connecting lines (see Figure S5A for an example). (C) Genomics-predicted transmission links (from B) further corroborated by additional epidemiological follow-up. Transmission links are indicated as confirmed known links (black lines) or possible links (gray lines) depending on whether contact was confirmed (e.g., members of the same household) or likely (e.g., lived in the same building but not the same home). A confirmed transmission pair that is part of a larger cluster of putative links, described in the text, is marked with an asterisk. In all panels, circles, representing unique individuals, are colored by the vaccination status of the individual at the date of sample collection. See also Figure S5.

Figure S5

Genomic support for transmission links and transmission predictions by vaccination status and symptoms, related to Figure 3 (A) Maximum-likelihood phylogenetic tree of the only high-confidence transmission pair from contact tracing without strong statistical support in outbreaker2 transmission reconstruction. This pair was in a cluster of six identical consensus genomes with very similar collection dates. No symptom onset date was known for either of the individuals in the pair. Even when incorporating contact tracing information into the model, another sample was predicted as almost equally likely to have been the ancestor of this case. (B) Maximum-likelihood phylogeny of the large cluster of cases associated with a single index case in a close-contact setting. Cases from this epidemiological cluster are colored in dark red. (A) and (B) are part of a larger phylogenetic tree available at https://auspice.broadinstitute.org/sars-cov-2/ma-delta/20211005/cluster-unique-usher. (C) Gray bars, fractions, and 90% confidence intervals indicate the proportion of individuals that were the origin of at least one transmission event predicted by outbreaker2 with a probability of >70%. Individuals are separated by vaccination status (left), days from final vaccination date to collection date (middle), or presence or absence of symptoms in vaccinated individuals (right). 90% binomial confidence intervals were calculated using the exact method through the binom package in R. Using simulations incorporating outbreaker2’s confidence in putative transmission links, we calculate that an infected unvaccinated individual was 0.18–2.11 times as likely to transmit as an infected vaccinated individual. Among fully vaccinated individuals, an infected asymptomatic individual was 0–0.99 times as likely to transmit as an infected symptomatic individual. Our estimates of relative risk are predicated on outbreaker2 correctly estimating the probability that it has chosen the correct index case of each putative transmission.