In the Netherlands, confirmed infections are reported to the local Public Health Service (PHS), and source and contact tracing is performed with a telephone interview. Data are obtained on sociodemographics, date of symptom onset, symptoms, vaccination status (and, if applicable, vaccine type, number of doses, and dates of administration) and locations the index-patient visited during the incubation and contagious periods. Medical ethics clearance for this study was not required (Appendix 1).

We defined a case as illness in a person who visited the nightclub on June 26, tested positive for SARS-CoV-2 within 14 days, and whose status was reported to the PHS. Cases were identified passively: persons were included only if they indicated during their PHS interview that they had visited the nightclub on June 26. The nightclub has an estimated capacity of 150 persons and was reported to be at full capacity that evening with attendees dancing and singing to loud music. A total of 60 confirmed COVID-19 cases were linked to the nightclub, raising suspicion of a superspreading event. Onset of symptoms occurred during June 27–July 3. Most case-patients were not fully vaccinated (defined as 14 days after completion of the vaccination series [Appendix 1 Figures 4, 5]): 4 (7.4%) persons were fully vaccinated and 41 (76%) were unvaccinated (Table). Most cases were in young adults (mean age 21.1 years [SD 3.3 years]) and women (60%), and most persons reported COVID-19–associated symptoms (93%). In 61% of cases, no other potential source for transmission besides the nightclub event was indicated. Of the 60 confirmed cases, 33 persons lived in the Amsterdam region and 27 resided in other regions (Table).

Figure 1

Figure 1. Clustering of severe acute respiratory syndrome coronavirus 2 cases related to nightclub event, Amsterdam, the Netherlands, June 2021. Red circles indicate sequences linked to the nightclub. A) Maximum-likelihood tree of...

Samples from 23/60 cases were available for sequencing, of which 3 were not eligible because of high cycle threshold values (>32). For 19/20 samples, we successfully obtained full genome sequences; all belonged to PANGO-lineage B.1.617.2 (4), which was denoted as variant of concern Delta by the World Health Organization (5) (Appendix 1 Table). To provide for phylogenetic context, we included weekly surveillance samples from the Amsterdam region (n = 421) in the analyses, as well as all Delta variant sequences from the Netherlands available in the GISAID database (https://www.gisaid.org; n = 4,465) (Appendix 2 Table) on August 1, 2021. All nightclub-associated genomes showed characteristics of a superspreading event: a tight phylogenetic cluster closely related in time (June 27–July 3) (Appendix 1 Figure 4) and genomic diversity (Figure 1). The pairwise genetic distance between all sequences was <2 single-nucleotide polymorphisms (Appendix 1 Figure 6), comparable to previously observed superspreading events (6). In addition, all sequences formed a monophyletic cluster marked by a specific single-nucleotide polymorphism combination: a Delta variant with C4321T in the presence of (wild-type) 22792C. In our dataset, all viruses with this combination collected before July 1 were sampled from persons who were at the nightclub. This combination was not observed in our dataset or in any Netherlands Delta sequences (n = 4465) from the GISAID database before June 26 (Appendix 1 Figure 7). Furthermore, randomly collected surveillance samples in the region from the weeks preceding the nightclub event showed diverse viruses circulating in the Amsterdam region (Appendix 1 Figure 8), and samples collected from 2 other nightclubs on June 26 also showed different lineages (Appendix 1 Figure 9). This finding makes multiple introductions at the nightclub with a highly prevalent, highly similar variant unlikely. In all, these findings strongly suggest a single introduction of the C4321T + 22792C variant, which was amplified by superspreading at the nightclub.

Figure 2

Figure 2. Detected increase of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) sequences with signature first detected in nightclub samples, Amsterdam, the Netherlands. A) Absolute number of weekly cases in randomly selected...

Since the introduction of C4321T + 22792C, the variant has been increasingly detected in random genomic surveillance from the Amsterdam region: no surveillance samples were detected in week 26 compared with 33% of samples in week 28 (Figure 2). This lineage was introduced the weekend nightclubs were opened and has clearly propagated in the community, where subsequent transmission of the lineage occurred.

This study illustrates the amplification of a specific linage in a largely unvaccinated group under circumstances such as those observed in a nightclub where social distancing measures and facemask requirements were lifted, despite a testing-for-access policy. In addition, our results highlight the consequence of superspreading events on subsequent transmission dynamics of SARS-CoV-2 in the community. Investigating an outbreak on June 26, 2021, the first date that social distancing measures were lifted under testing-for-access conditions, enabled us to isolate a single SARS-CoV-2 transmission event.

The role of superspreading in SARS-CoV-2 transmission has been highlighted previously (6,7), also in the context of nightclubs (8,9). Considering the potential of SARS-CoV-2 to be transmitted through aerosols (10,11), nightclubs can be a high-risk setting because of poor ventilation and sustained overcrowding. Our findings suggest that the rapid surge in cases in July 2021 was at least partially driven by superspreading events such as the event we describe.

In particular, testing-for-access, as it was put in place in the weeks following June 26, provided opportunity for infectious persons to slip through. Access was provided immediately after a single Johnson & Johnson/Janssen vaccination (https://www.janssen.com) (too soon), a negative antigen test result was valid for 40 hours (too long), and checking of QR codes was reported to be inconsistent at some venues (12,13).

This study used data collected for nonresearch purposes during scaled-down source and contact tracing and has limitations. First, cases were passively included, which could underestimate the true extent of the outbreak, because asymptomatic cases or cases tested only by self-administered antigen tests might have been missed. This factor could also explain the high percentage of symptomatic cases (14). Nevertheless, we believe this factor did not result in a biased selection of cases. Second, we conducted genomic analysis for only 1/121 detected hospitality sector–related clusters, limiting generalizability of our findings.

In conclusion, testing-for-access did not prevent superspreading at this event, indicating the need for caution when easing social distancing measures in night life, even under more optimal testing-for-access conditions. This finding is particularly relevant in a population where vaccination coverage is low or when new variants circulate that are associated with lower vaccine effectiveness.