Abstract
Retrospective serological analysis of discarded serum samples reveals that SARS-CoV-2 was introduced to New York City between late January and early February 2020. This led to a 4–5 week period of silent community transmission before the first official case was reported on March 1, 2020. By the time of the first official diagnosis, thousands of infections were already present, highlighting the critical need for early surveillance.
Introduction
In early 2020, New York City emerged as a global epicenter of the COVID-19 pandemic. The first confirmed case was reported on March 1, involving a traveler returning from Iran, followed by the first community-acquired case on March 3 in a resident of New Rochelle. However, the rapid escalation in hospitalizations and deaths that followed suggested that the virus had been circulating within the community much longer than official records indicated.
This study sought to fill the "knowledge gap" regarding the timing of the virus's entry into the city. Because early testing was strictly limited to symptomatic individuals with a history of travel to known hotspots, the "silent spread" among the general population went unrecorded. By using retrospective antibody testing (serology) on blood samples collected for unrelated clinical reasons, the researchers aimed to reconstruct the early timeline of the epidemic and determine when the virus actually established a foothold in the city.
Methods
The researchers developed and validated their own enzyme-linked immunosorbent assays (ELISAs) to detect IgG antibodies against two key components of the SARS-CoV-2 virus: the spike trimer (S) and the nucleocapsid protein (NP).
Assay Validation
To ensure the accuracy of the results, the assays were rigorously validated:
- Specificity: The tests were run against 106 "pre-pandemic" samples collected between 2015 and 2018. The S-IgG assay achieved a high specificity of 99%, ensuring that positive results were not due to cross-reactivity with other common coronaviruses.
- Sensitivity: The researchers tested 146 samples from PCR-confirmed COVID-19 patients to determine how well the test detected true positives over time. The S-IgG assay reached 93% sensitivity by day 22 post-symptom onset.
Sampling and Modeling
The study utilized "discarded" serum samples—residual blood left over from routine clinical testing—collected between February 24 and March 29, 2020. Two primary cohorts were analyzed:
- CUIMC-ED: 814 randomly selected samples from the Columbia University Irving Medical Center Emergency Department in Manhattan.
- CareMount Medical: 1,841 samples from a network of suburban clinics north of NYC (Westchester and surrounding counties).
To translate "antibodies detected" into an estimated "date of infection," the team employed epidemiological modeling. This model incorporated known viral parameters, including the incubation period, the time required for seroconversion (developing detectable antibodies), the serial interval (the time between successive infections in a chain), and the reproductive ratio ($R_0$).
Results
The analysis provided clear evidence that SARS-CoV-2 was circulating in New York City weeks before the first official case.
Early Seroprevalence and the "Silent Window"
In the very first week of sampling (February 24 – March 1), the NYC Emergency Department cohort already showed a 4.2% seroprevalence rate. Given that it typically takes 10–14 days for antibodies to become detectable, these individuals must have been infected in mid-February or earlier.
Mathematical modeling back-calculated the initial introduction of the virus to a window between January 23 and February 4, 2020, with a projected index case around January 28. This indicates a 4–5 week period of undetected community transmission.
Urban vs. Suburban Spread
The study highlighted a significant difference in transmission dynamics based on geography. Seroprevalence was notably higher in the dense urban environment of Manhattan (4.2% in the first week) compared to the suburban areas (1.4% to 1.9%). This disparity reflects the impact of population density and public transit reliance on the early, rapid spread of the virus.
The Tip of the Iceberg
By the time the first community-acquired case was officially diagnosed on March 3, the researchers estimate there were already approximately 3,888 infections within the CUIMC catchment area alone. Extrapolating this to the city at large suggests that tens of thousands of New Yorkers were already infected before the public was even aware that local spread was occurring. Over the following 18 weeks, the city reached a peak daily infection number of approximately 4,500 and a cumulative caseload of roughly 400,000.
Discussion and Limitations
The findings suggest that the initial public health focus on travelers allowed the virus to reach a critical mass undetected. Using discarded clinical samples for serological surveillance proved to be a powerful tool for retrospective analysis, though the study does have limitations.
The use of Emergency Department samples may introduce a selection bias, as this population may be sicker or have different exposure risks than the general public. Additionally, as the pandemic progressed and specialized "COVID clinics" were established, some positive cases might have been diverted away from the general pool sampled here, potentially leading to an underestimation of seroprevalence in later weeks. Finally, the geographic scope was limited to Manhattan and its northern suburbs, potentially missing different introduction timelines in other boroughs or near other major transit hubs like JFK Airport.
Conclusion
This study documents a critical window of "silent spread" that set the stage for New York City's massive initial outbreak. The results underscore the importance of early, widespread serological surveillance. Had such systems been in place in early 2020, they could have provided the "early warning" needed to implement social distancing and masking mandates sooner, potentially mitigating the devastating impact of the first wave.
Summary updated