SARS-COV-2 vaccine responses in renal patient populations

Background Dialysis patients and immunosuppressed renal patients are at increased risk of COVID-19 and were excluded from vaccine trials. We conducted a prospective multicentre study to assess SARS-CoV-2 vaccine antibody responses in dialysis patients and renal transplant recipients, and patients receiving immunosuppression for autoimmune disease. Methods Patients were recruited from three UK centres (ethics:20/EM/0180) and compared to healthy controls (ethics:17/EE/0025). SARS-CoV-2 IgG antibodies to spike protein were measured using a multiplex Luminex assay, after first and second doses of Pfizer BioNTech BNT162b2(Pfizer) or Oxford-AstraZeneca ChAdOx1nCoV-19(AZ) vaccine. Results Six hundred ninety-two patients were included (260 dialysis, 209 transplant, 223 autoimmune disease (prior rituximab 128(57%)) and 144 healthy controls. 299(43%) patients received Pfizer vaccine and 379(55%) received AZ. Following two vaccine doses, positive responses occurred in 96% dialysis, 52% transplant, 70% autoimmune patients and 100% of healthy controls. In dialysis patients, higher antibody responses were observed with the Pfizer vaccination. Predictors of poor antibody response were triple immunosuppression (adjusted odds ratio [aOR]0.016;95%CI0.002–0.13;p < 0.001) and mycophenolate mofetil (MMF) (aOR0.2;95%CI 0.1–0.42;p < 0.001) in transplant patients; rituximab within 12 months in autoimmune patients (aOR0.29;95%CI 0.008–0.096;p < 0.001) and patients receiving immunosuppression with eGFR 15-29 ml/min (aOR0.031;95%CI 0.11–0.84;p = 0.021). Lower antibody responses were associated with a higher chance of a breakthrough infection. Conclusions Amongst dialysis, kidney transplant and autoimmune populations SARS-CoV-2 vaccine antibody responses are reduced compared to healthy controls. A reduced response to vaccination was associated with rituximab, MMF, triple immunosuppression CKD stage 4. Vaccine responses increased after the second dose, suggesting low-responder groups should be prioritised for repeated vaccination. Greater antibody responses were observed with the mRNA Pfizer vaccine compared to adenovirus AZ vaccine in dialysis patients suggesting that Pfizer SARS-CoV-2 vaccine should be the preferred vaccine choice in this sub-group.


Introduction
The renal patient population is diverse, including individuals requiring renal replacement therapy in the form of dialysis, and those receiving immunosuppression as a result of renal transplantation or for autoimmune Open Access *Correspondence: diseases, such as vasculitis, systemic lupus erythematosus (SLE) or primary glomerulonephritis (GN). SARS-CoV-2 infection is associated with high morbidity and mortality in these complex patients [1]. The vulnerability of these individuals led to public health advice for them to adopt extended protective self-isolation measures early in the pandemic, and they were prioritised for vaccination against SARS-CoV-2. In the UK, most are being offered third doses of SARS CoV-2 vaccines in the autumn of 2021.
Vaccine trials have been successful in the general population and have shown that some vaccines are up to 95% effective at preventing symptomatic infection [2,3]. However, patients with end stage kidney disease and those receiving immunosuppressive medications were largely excluded from the initial SARS-CoV-2 vaccine trials. Historically, vaccination against other infections results in suboptimal responses in these patients [4][5][6]. Vaccines against SARS-CoV-2 approved for use in the UK include the mRNA vaccines Pfizer BioNTech BNT162b2 (Pfizer) and Moderna mRNA-1273,; and the Adenovirus vector vaccines Oxford-AstraZeneca ChAdOx1 nCoV-19 (AZ) and Janssen Ad26.COV2-S.
Several groups have now reported on the immunogenicity of SARS-CoV-2 vaccines in immunocompromised patients. The serologic response to mRNA vaccines in patients receiving dialysis has been encouraging. One study, which included 1256 dialysis patients, reported seroconversion rates of 85-95% after the second dose; comparable to healthy controls [7]. However, antibody titres are lower when compared with healthy controls despite high seroconversion rates [8]. Most studies have assessed the effectiveness of mRNA vaccines and few report the neutralising capacity of antibodies, but a UK multi-centre cohort study of 178 haemodialysis patients revealed that infection-naïve patients had a markedly reduced neutralising antibody response to the AZ vaccine compared to the Pfizer vaccine, although the difference was reduced in individuals that experienced an infection prior to or vaccination [9]. Risk factors for reduced immunogenic response in dialysis patients included older age, co-morbidities and dialysis vintage.
In contrast, antibody production after mRNA vaccination is poor in individuals with a renal transplant with less than 60% generating protective level antibodies after the second dose [9,10]. Comparison of the Pfizer and AZ vaccines in kidney transplant patients reported enhanced humoral responses with the Pfizer vaccine, however T-cell responses to both vaccines were attenuated when compared to healthy controls [11]. A reduced antibody response was associated with older age, more recent transplantation and anti-metabolite drugs, such as mycophenolate mofetil (MMF).
Less data is available on patients with autoimmune disease, and disease heterogeneity makes interpretation more challenging. The OCTAVE trial is assessing SARS-CoV-2 vaccine responses (to both Pfizer and AZ vaccines) in patients with chronic immune-mediated diseases on immunosuppression including individuals with inflammatory arthritis, ANCA-associated vasculitis, inflammatory bowel disease, hepatic disease and malignancy [12]. Preliminary data indicates that although 89% of patients seroconverted 4 weeks after the second vaccine dose, 40% generated lower levels of antibodies compared to healthy subjects. 72% of patients with ANCA-associated vasculitis (N = 30), who had all received rituximab, did not generate detectable antibodies, although T cell responses in all sub-groups were similar to those in healthy individuals. In addition to the poor antibody response following rituximab (particularly if rituximab has been administered within 6 months prior to the vaccine), glucocorticoids and methotrexate have also been associated with reduced antibody levels [13][14][15][16].
We conducted a prospective cohort study to assess differences and predictors of SARS-CoV-2 antibody responses following vaccination in three immunocompromised renal populations; dialysis, renal transplant and patients receiving immunosuppression for multisystem autoimmune disease. The use of the same assay across groups allowed direct comparison across renal patient populations as well as with healthy controls. Additionally, the majority of patients recruited to this study received the AZ vaccine, therefore addressing the relative paucity of data on responses to adenovirus vector vaccines in the existing literature. Our cohort includes over 100 individuals who have received the B-cell depletion therapy with rituximab, a population of particular interest.

Methods
This prospective observational cohort study included patients recruited from the Departments of Nephrology, at Cambridge University Hospitals NHS Foundation Trust, East and North Hertfordshire NHS Trust and University Hospitals of Leicester NHS Trust in the UK (ethics reference: 20/EM/0180). Healthy controls were recruited at Cambridge University Hospitals NHS Foundation Trust (ethics reference: 17/EE/0025). Patients receiving dialysis or immunosuppression following either renal transplantation or for autoimmune renal disease were eligible. Patients receiving IVIg or plasma exchange were excluded to avoid potential confounders of vaccine response. Blood sampling was performed approximately three monthly for up to 18 months. Flexibility in the timing of blood samples was permitted to align sampling with 4-6 weeks post COVID vaccine administration during the study, as well as (wherever possible) aligning with routine clinical tests and access to blood testing facilities. Clinical data were collected from electronic medical records and patient interviews and included baseline demographics, changes to immunosuppressive medication over time and data on episodes of SARS-CoV-2 infection. SARS-CoV-2 IgG antibodies to spike protein (wild type Wuhan variant) were measured using a UKAS accredited multiplex Luminex assay following the first and second doses of either Pfizer BioNTech BNT162b2 (Pfizer) or Oxford-AstraZeneca ChAdOx1 nCoV-19 (AZ) vaccine. Spike antibody MFI titres greater than 1896 were considered positive, with a sensitivity of 92% and specificity of 99% as determined by ROC analysis, nucleocapsid antibody titres of > 6104 were considered positive for previous natural infection [17,18]. Statistical analyses were performed using Stata version 14. Between group differences in proportional response to vaccine were compared using chi-squared test. Median spike antibody titres were compared using Wilcoxon rank-sum test. Multivariate logistic regression modelling was used to evaluate the odds of vaccine response.

Antibody response by MMF and prednisolone status
Within the transplant cohort, median s-antibody titres were lower in those taking MMF (1145 [IQR

Discussion
Nephrologists are involved in the care of a heterogeneous group of complex patients. These individuals often have comorbidities other than their kidney diagnosis, and are vulnerable to severe Covid-19 infection. This multi-centre study provides real world comparable data on vaccine efficacy in three immunocompromised renal populations compared to healthy controls. More than 50% of patients in this cohort had received the AZ SARS-CoV-2 vaccine, enabling comparison between mRNA and adenoviral vaccines to be made. In all patient groups, antibody titre and the proportion of patients with positive antibody responses incremented following a second dose of vaccine. Although positive antibody responses were observed in 97% of dialysis patients, only 52% of renal transplant recipients and 70% of autoimmune patients had positive titres after the second vaccine dose. Median titres in transplant recipients were lower than both autoimmune and dialysis patients, and median titres in autoimmune patients lower than dialysis patients. In dialysis patients, antibody responses were higher with the mRNA Pfizer vaccine compared to AZ adenovirus vaccine and lower overall compared to healthy controls. A lower spike antibody titre following 2 doses of vaccine was associated with a higher likelihood of breakthrough infection.
Key predictors of poor vaccine response were receiving triple immunosuppression or concurrent MMF in transplant recipients, immunosuppression use in patients with CKD stage 4/5 and B cell depletion (prior rituximab therapy) in autoimmune patients. There was a clear positive correlation between time since last rituximab and antibody titre with 94% of patients achieving a positive antibody response when a second vaccine dose was administered more than 12 months after last rituximab dose. These data suggest that repeating vaccination more than 12 months after last Rituximab dose may have a beneficial effect on antibody response to vaccination. For stable patients receiving ongoing repeat rituximab dosing, where clinically appropriate, postponing routine rituximab dosing to optimise vaccine timing and response may be considered, provide close monitoring of patients is in place to avoid relapse [19]. The majority of patients in our autoimmune disease population received rituximab for ANCA-associated vasculitis (AAV), where for remission maintenance fixed interval 6 monthly rituximab dosing is frequently used [20]. Alternative dosing strategies with longer dosing intervals and dosing according to peripheral blood CD19 B cell return and PR3/MPO-ANCA rise may be reasonable, temporarily, in stable vasculitis patients with low relapse risk and experienced clinician oversight, to help optimise vaccine responses [19].
Strengths of this study are the inclusion of three renal patient populations and healthy controls, with evaluation of antibody response on the same diagnostic platform enabling results to be directly compared. In addition, the large number of rituximab treated patients is of particular interest and has enabled the time dependent effect of rituximab to be demonstrated on vaccine response. The frequency of prior COVID 19 infection was low in our population which allowed assessment of vaccine response without the confounding effect of prior infection in the majority. Limitations of the data presented include lack of data on T cell response and neutralising capacity of the antibodies. This is particularly important for those individuals with sub-optimal antibody titres, in whom a weak positive result may offer a false reassurance/overestimate of protection and a borderline negative antibody result may falsely imply no protection at all. Work is ongoing to assess the durability of response and the impact of third dose vaccinations in these immunocompromised renal patient populations.

Conclusion
These data provide insight into patient subgroups at greatest risk for suboptimal SARS-CoV-2 vaccine response. Although, it is not possible to make individual patient recommendations based on this data, patients on triple immunosuppression or MMF following renal transplantation, autoimmune patients receiving rituximab, and patients with CKD stage 4/5 receiving immunosuppression are at greatest risk for poor vaccine response. These groups should be prioritised for repeated vaccination. At the time of vaccination, temporary reduction of immunosuppression could be considered, but a careful balance between transplant rejection risk/autoimmune disease flare and achieving adequate vaccine response is required, and decisions should be made on an individual basis only by an experienced nephrologist. Reducing prednisolone dose alone, where possible, may have a beneficial effect on response to vaccination. For autoimmune patients, the positive correlation between vaccine response and time since rituximab suggests that delaying routine remission maintenance dosing could be considered; provided relapse risk is deemed to be low and adequate relapse monitoring is in place. Although third and fourth SARS-CoV-2 vaccine doses may allow a further increment in the proportion of patients with detectable antibodies, it remains vital that antibody durability and function is assessed in immunosuppressed patients to identify the most vulnerable sub-populations in whom additional COVID-19 prevention strategies will be of greatest benefit.