Discussion

In this single-center, US study of humoral response to 2 doses of SARS-CoV-2 mRNA vaccine BNT162b2, we observed a high seroconversion rate of 96% using the manufacturer recommended cutoff of anti-S >50 AU/mL in a cohort of predominantly Black dialysis patients. This finding is similar to that reported in studies in patients receiving maintenance dialysis from Europe and Israel. ^{Reference Paal, Arend and Lau10,Reference Grupper, Sharon and Finn21–Reference Broseta, Rodríguez-Espinosa and Rodríguez23} As seropositivity does not necessarily translate into protective immunity, we assessed humoral response using an anti-S level of 840 AU/mL as a cutoff for “sufficient response” because this value correlates with a neutralizing titer of 1:250, which corresponds to a 95% probability of viral neutralization and is likely predictive of protective immunity against COVID-19. ^{Reference Van Praet, Reynders and De Bacquer6,Reference Ebinger, Fert-Bober and Printsev19} We found that 20% of our study cohort fell below this threshold, suggesting that 2 doses of the vaccine may be ineffective in 1 of 5 patients in this dialysis population. Similar response rates were shown by 2 other US dialysis studies that utilized comparable neutralizing titers as the cutoff for protective immunity. ^{Reference Anand, Montez-Rath and Han7,Reference Mulhern, Fadia and Patel24}

Patients with prevaccination COVID-19 had heightened antibody response compared to those without prior COVID-19. This finding is consistent with published findings in both the general and dialysis population and supports the presence of immune memory and so called “hybrid immunity” to SARS-CoV-2. ^{Reference Ebinger, Fert-Bober and Printsev19,Reference Chan, Fuca, Zeldis, Campbell and Shaikh25} We also found a more vigorous antibody response in patients whose COVID-19 occurred >6 months prior to vaccination compared to those whose infection occurred <6 months before vaccination. This interesting phenomenon has been examined by Gaebler et al ^{Reference Gaebler, Wang and Lorenzi26} who have shown that memory B cells continue to evolve and display clonal turnover 6 months after COVID-19, producing antibodies with somatic hypermutation and more potent neutralizing activity. ^{Reference Gaebler, Wang and Lorenzi26,Reference Wang and Muecksch27} These B-cell clones expand markedly after vaccination, leading to a robust serological response. This observation may have implications in vaccine booster strategy in patients on dialysis.

In our cohort, 8 patients had a breakthrough COVID-19 infection 25–248 days after receiving the second vaccine dose (Table 2). The first 2 patients had their blood drawn for antibody titers 11–17 days after their COVID-19 diagnosis so the higher titers could be partly due to response to the infection itself. The third patient with breakthrough infection ∼4 months after the second vaccine dose did not attain sufficient antibody response, so infection was not surprising. For the rest, COVID-19 occurred >5 months after the second vaccine dose, implying that protective humoral immunity may have waned over time. This confirms a previous study in a maintenance HD population, which showed rapid antibody loss by 6 months after SARS-CoV-2 vaccination. ^{Reference Davidovic, Schimpf and Abbassi-Nik28} In a recent study examining breakthrough infection in dialysis patients after SARS-CoV-2 vaccination, antibody titers decreased significantly 5–6 months after vaccination and breakthrough infection correlated with preinfection circulating antibody level, ^{Reference Anand, Montez-Rath and Han29} emphasizing the need for a vaccine booster around 6 months after the second vaccine dose in these patients. Of our 8 patients with breakthrough infection, 2 developed severe disease and the rest only had mild symptoms, suggesting that a sufficient antibody response after 2 doses of SARS-CoV-2 vaccine may be effective in preventing severe illness at 6–8 months.

In our cohort, factors associated with insufficient antibody response by univariate analyses included HIV infection with CD4 count < 200 cells/mcL, active malignancy, prior nonresponse to HBV vaccination, and use of high-dose erythropoiesis-stimulating agent (ESA). Multivariate analyses identified ESRD comorbidity index ≥5 and absence of prior HBV vaccination response as predictors of insufficient vaccine response. These findings suggest that sicker and more immunocompromised patients are less likely to respond to the vaccine. Prior nonresponse to HBV vaccination is a marker of immune dysfunction in ESRD patients. ESRD is an immunocompromising condition with impairment in both innate and adaptive immunity which can lead to attenuated response to all vaccines. ^{Reference Kato, Chmielewski and Honda30–Reference Crespo, Collado and Mir32} High-dose ESA or ESA resistance is commonly associated with elevated inflammatory markers and has been negatively associated with response to HBV vaccination. ^{Reference Afsar33} However, the mechanisms by which ESA resistance leads to reduced vaccine responsiveness are yet to be elucidated. Also, patients with hypertension requiring blood pressure medications were more likely to develop a sufficient antibody response than those without hypertension. This has not been observed by others. In fact, hypertension has been shown to be a predictor of low vaccine response in nondialysis populations. ^{Reference Watanabe, Balena and Tuccinardi34,Reference Ruether, Schaub and Duengelhoef35} The likely explanation for our finding is that the patients with hypertension in our study population could represent a healthier and less immunocompromised subgroup, and this is supported by our observation that patients with hypertension were less likely to be HBV vaccine nonresponders and had a shorter dialysis vintage.

Unlike other studies, ^{Reference Van Praet, Reynders and De Bacquer6,Reference Anand, Montez-Rath and Han7,Reference Danthu and Hantz9,Reference Yanay, Freiman and Ma11,Reference Bertrand, Hamzaoui and Lemée36,Reference Longlune, Nogier and Miedougé37} older age, immunosuppressive therapy, longer dialysis vintage, and lower serum albumin were not independent predictors of humoral response in our cohort. One study showed that non-White race and Hispanic ethnicity were associated with lower response to mRNA vaccine. ^{Reference Anand, Montez-Rath and Han7} In our study, race was not associated with vaccine response.

Our study had several limitations. It was a single-center study with a small sample size and without a control group. Several studies have shown a blunted antibody response in patients on dialysis compared to nondialysis populations or healthy controls. ^{Reference Van Praet, Reynders and De Bacquer6,Reference Paal, Arend and Lau10,Reference Yanay, Freiman and Ma11,Reference Simon, Rubey and Treipl38} We did not measure prevaccination antibody levels; therefore, we could not rule out the possibility of asymptomatic COVID-19 cases and subsequent misclassification in our analyses. We did not assess cellular immune response or evaluate the durability of humoral immunity. As our vaccination program was implemented several months ago, humoral response to the current and emerging variants was not assessed. Also, we only examined response to BNT162b2 vaccine and whether response to mRNA-1273 would be significantly different in our cohort is unknown. Lastly, because our patients are predominantly Black, results from this study may not be generalizable to non-Black patients.

In conclusion, most dialysis patients in our cohort seroconverted after 2 doses of BNT162b2 mRNA vaccine. However, 1 of 5 patients remained at risk for COVID-19 due to the inability to achieve sufficient protective immunity. Although SARS-CoV-2 vaccination is strongly recommended for all patients on maintenance dialysis, the relatively high rate of low vaccination response shown in this study emphasizes the need for a more effective vaccine strategy in this vulnerable population, perhaps with modification of vaccine dose and booster frequency, and with mitigating strategies such as ring vaccination. Our results also highlight the need to identify the subgroup of dialysis patients at risk for inadequate vaccine response via serological testing and the importance of maintaining preventative measures such as social distancing and mask wearing until a protective immune response is attained.