Longitudinal nucleocapsid antibody testing reveals undocumented SARS-CoV-2 infections in patients with lung cancer

During the COVID-19 pandemic, monitoring hospitalizations and reported infections had a public health focus. However, the impact of “silent spread” by asymptomatic or unaware carriers and subsequent SARS-CoV-2 infection on vulnerable populations such as patients with lung cancer (LC) remains a major concern. Patients with LC are vulnerable to SARS-CoV-2 infection due to the lungs as shared target organ of disease as well as other factors that are often present such as advanced age, associated comorbidities, limited pulmonary reserves, and potentially immunocompromising therapies. In the early pandemic, prior to vaccination availability, LC patients had a high risk of severe outcomes from SARS-CoV-2 infection, with initial COVID-19 case fatality rates exceeding 30%.¹ Frequent exposure to healthcare facilities likely further intensified risk of infection. However, these alarming rates were based on hospitalization counts, likely overlooking the asymptomatic or mildly symptomatic infections.

The extent of SARS-CoV-2 spread in this population during later pandemic phases remains unclear. Monitoring SARS-CoV-2 antibodies against the viral nucleocapsid as a specific indicator of recent natural infection not attributed to vaccination offers a more accurate representation of the incidence of infection at the population level.

To determine the true burden of infection, we prospectively studied the prevalence and kinetics of anti-nucleocapsid antibodies (anti-N Ab) in blood collected every 3 months from a cohort of 175 LC patients in New York City from January 2021 to January 2023. The median age was 70 years with 57% female and 55% with stage IV cancer. Table S1A shows additional clinical characteristics. Overall, 35% (62/175 patients) had documented SARS-CoV-2 infections confirmed by standard testing (PCR or antigen test): 53 patients had a single infection and 9 patients had two infections (Figure S1B). In the cohort, 96% (168/175 patients) were fully vaccinated, receiving at least the primary vaccination series: two mRNA vaccine doses (BNT162b2 or mRNA-1273) or one adenoviral vector dose (Ad26.COV2-S). The majority had at least one additional booster shot. Among the vaccinated group, 30% (50/168 patients) had clinically documented breakthroughs infections after vaccination (Table S1B).

After analyzing the anti-N serology, 61% (107/175 patients) tested positive in at least one of their plasma samples over the study period (Figure S1A). The overall anti-N Ab positivity rate peaked in the latter half of 2021 and throughout 2022, aligning with the Delta (B.1.617.2) and Omicron (B.1.1.529) waves (Figures S1B and S1C).² Among the positive anti-N serology patients, only 46% (49/107 patients) had clinically documented infections and the remaining 54% (58/107 patients) had no documented SARS-CoV-2 infection (Figure S1E). Thus, anti-N serology revealed an additional 33% (58/175 patients) with evidence of SARS-CoV-2 infection but no clinical record. When including both documented infections confirmed by standard testing and positive anti-N serology data, the total estimated SARS-CoV-2 infection rate reached 68% (120/175 patients) by January 2023 (Figure S1D).

To understand the long-term humoral response to SARS-CoV-2, we evaluated the anti-N and anti-S antibody kinetics over a 365-day period following the first documented infection in our cohort. The estimated anti-N Ab half-life was 126 days (95% CI: 107–164) compared to 262 days (95% CI: 147-NA [not available]) for the anti-Spike (anti-S Ab) (Figures S2A and S2B). These findings guided us in selecting a 200-day window to assess the correlation between clinical documented infections and anti-N Ab detection. The assay showed a sensitivity of 90%, detecting positive anti-N Ab in 44 out of the 49 documented infections (Figure S2C). Furthermore, we investigated whether post-infection reads with low anti-N Ab also had compromised anti-S Ab, but this correlation was not observed (Figure S2D).

A relevant issue for LC patients is the risk of SARS-CoV-2 breakthrough infection. Combining documented infection counts and the anti-N serology data, we estimated the cumulative SARS-CoV-2 infection incidence curve, finding an 18.3-month median time-to-infection from the date of full vaccination (95% CI: 16.5-NA) (Figure S2E). Cumulative infection rates at 6, 12, and 18 months after full vaccination were 3% (95% CI: 0%–5%), 18% (95% CI: 12%–25%), and 48% (95% CI: 37%–57%), respectively. Among the 168 fully vaccinated patients, 64% (108/168 patients) experienced breakthrough infections, with 87 patients having a single breakthrough infection, 20 patients with two breakthrough infections, and one patient with three (Figure S2F). In total, 130 breakthrough infections were counted during the follow-up period after full vaccination. This corresponds to a breakthrough infection incidence of 4.4 per 100 personmonths (95% CI, 3.9–5.0). Among the breakthrough infections, approximately 3% (4/130) were clinically considered severe, necessitating hospitalization and intensive treatments, and 38% (49/130) were clinically documented as non-severe, of which 47/49 were symptomatic and 2/49 asymptomatic. Thus, the majority of breakthrough infections, 59% (77/130), were unreported, detected solely by anti-N Ab positivity. These likely represent asymptomatic or mild infections.

Integrating clinical records and longitudinal anti-N serology revealed that by early 2023, around two-thirds of our cohort had evidence of SARS-CoV-2 infection. The addition of patients with a positive anti-N serology finding but no concurrent clinical record documenting infection through standard testing implies that many infections, likely mild or asymptomatic, went undocumented. Our overall infection rate aligns with recent reports on the general population data by the third quarter of 2022³ and with updates from the CDC (Centers for Disease Control and Prevention) regarding infectioninduced seroprevalence among individuals 65 and older in New York state.⁴

Despite high vaccination uptake, the cumulative incidence increased to almost 50% by 18 months after the primary doses. This period aligned with the emergence and wide spread of the Delta and especially the Omicron variants, likely facilitating immune-evasive breakthroughs. Supporting this, we observed anti-N Ab peaks coinciding with these variants’ waves, consistent with state-level studies reporting increased seroprevalence-to-case prevalence ratios during Omicron’s spread.⁵ Such findings suggest gaps in the rapid antigen test reporting, affecting accurate infections measurement.

The clinical implications of waning anti-N antibodies following natural infection remain unclear. Conclusive data on anti-N incidence and persistence after infection and re-infection in specific patient populations are needed. We determined a half-life of 126 days for anti-N Ab, and the detection assay showed 90% rate of sensitivity within 200 days after infection—results aligning with recent studies.^6,7 There is limited number of studies leveraging longitudinal anti-N tracking to estimate breakthrough infections overtime.⁸ A key aspect of our study was using longitudinal anti-N antibody measurements to comparatively assess increases indicating likely breakthroughs and reinfections after vaccination and documented infections. However, certain limitations may affect the result interpretation. The assay sensitivity and chosen cutoff, although guided by literature, could underestimate infections given the gradual antibody decline over time. Likewise, the assay may not have detected or distinguished reinfection cases in vaccinated patients whose anti-N levels fell below the assay’s arbitrary cutoff.⁹ Moreover, our single-center design and lack of probabilistic sampling may reduce generalizability. Despite these constraints, we estimated the number of breakthrough infections per patient, suggesting over half of the patients experienced at least one breakthrough and at least 12% had multiple breakthroughs. Overall, by incorporating anti-N serology, we estimated substantially more breakthrough infections than standard testing alone. However, further data on anti-N Ab kinetics after infections in specific patient populations are needed to validate the use of anti-N serology to detect breakthrough infections.

In contrast to initial reports of high risk among LC patients for severe SARS-CoV-2 infection outcomes,¹ we found fewer than 5% of breakthrough infections were severe. The widespread vaccine uptake likely contributed to this difference, reducing disease severity. Nonetheless, it is necessary to investigate the factors contributing to clinical severity versus asymptomatic cases despite infection.

Our results show viral immune evasion enabling breakthrough infections, highlighting the importance of effective monitoring mechanisms, such as anti-N serology, to track infection rates in specific vulnerable populations. It is crucial to recognize that SARS-CoV-2 may lead to significant long-term health implications¹⁰ and vulnerable groups likely still face increased severity risks. Future analyses should evaluate the protection, versus infection and illness, conferred by newer boosters in LC patients and other vulnerable populations. Likewise, the impact of symptomatic or asymptomatic SARS-CoV-2 infections on patients’ treatment efficacy, side effects, and outcomes needs to be evaluated.