Article Text
PDF
Abstract
Introduction To assess the incidence and risk factors for breakthrough COVID-19 infection in a vaccinated cohort of patients with autoimmune rheumatic diseases (AIRDs) and determine whether antibodies to receptor binding domain of spike protein (anti-RBD) serve as a reliable predictor of susceptibility to such infections.
Methods Patients with AIRDs who had completed two doses of SARS-CoV2 vaccines were included and anti-RBD antibodies were determined 4–6 weeks post the second vaccine dose and stratified into good responders (GR) (>212 IU), inadequate responders (IR) (0.8–212 IU) and non-responders (NR) (<0.8 IU). Patients who had completed a minimum of 8 weeks interval after the second dose of vaccine were followed up every 2 months to identify breakthrough infections. All sero converted patients who had contact with COVID-19 were also analysed for neutralising antibodies.
Results We studied 630 patients of AIRDs (mean age 55.2 (±11.6) years, male to female ratio of 1:5.2). The majority of patients had received AZD1222 (495, 78.6%) while the remaining received the BBV152 vaccine. The mean antibody titre was 854.1 (±951.9), and 380 (60.3%) were GR, 143 (22.7%) IR and 107 (16.9%) NR.
Breakthrough infections occurred in 47 patients (7.4%) at a mean follow-up of 147.3 (±53.7) days and were proportionately highest in the NR group (19; 17.75%), followed by the IR group (13; 9.09%) and least in the GR group (15; 3.95%). On log-rank analysis, antibody response (p<0.00001), vaccine(p=0.003) and mycophenolate mofetil (p=0.007) were significant predictors of breakthrough infections. On multivariate Cox regression, only NR were significantly associated with breakthrough infections (HR: 3.6, 95% CI 1.58 to 8.0, p=0.002). In sero converted patients with contact with COVID-19, neutralisation levels were different between those who developed and did not develop an infection.
Conclusion Breakthrough infections occurred in 7.4% of patients and were associated with seronegativity following vaccination. This provides a basis for exploring postvaccination antibody titres as a biomarker in patients with AIRD.
- Covid-19
- vaccination
- arthritis
- antirheumatic agents
Data availability statement
Data are available on reasonable request. Data will be available from the corresponding author on reasonable request.
This article is made freely available for personal use in accordance with BMJ’s website terms and conditions for the duration of the covid-19 pandemic or until otherwise determined by BMJ. You may use, download and print the article for any lawful, non-commercial purpose (including text and data mining) provided that all copyright notices and trade marks are retained.
https://bmj.com/coronavirus/usageStatistics from Altmetric.com
Key messages
What is already known about this subject?
A delayed and suppressed immune response to SARS- Cov2 vaccines has been observed in patients with autoimmune rheumatic diseases (AIRD).
What does this study add?
Breakthrough infections were observed in 7.4% of our cohort of 630 patients with AIRD.
Absent antibody response to receptor binding domain of spike protein (non-responders) at 1 month postvaccination was a significant predictor of breakthrough infections (HR: 3.6, 95% CI 1.6 to 8.0)).
How might this impact on clinical practice or future developments?
Postvaccination antibody response can be used as a biomarker for successful vaccination response in patients with AIRD.
This can help prioritise a subgroup of patients with AIRDs for booster doses of the vaccine.
Introduction
Patients with autoimmune rheumatic diseases (AIRDs) form a high priority group for vaccination against SARS-CoV-2.1 Although patients with AIRDs have been excluded from vaccine trials, there is ample proof that the vaccines are safe and efficacious in this group of patients.2 A delayed and suppressed response to SARS-CoV-2 vaccines, predominantly mRNA vaccines, have been demonstrated in patients with AIRD, especially for those on methotrexate, mycophenolate mofetil (MMF), rituximab (RTX), abatacept and glucocorticoids.3–7
Despite evidence of reduced immunogenicity of vaccines in AIRDs, initial data on breakthrough COVID-19 has been reassuring. Data from two large European registries showed a breakthrough rate of less than 1% from all patients with AIRD and a single centre American study found this to be 4.7%. However, breakthrough infection was associated with a much higher risk of death (8%–13%) and post-COVID-19 sequelae (8%).8 9 Thus, it is of utmost importance to identify and mitigate risk factors for breakthrough infections. Those at high risk may require booster vaccinations for better protection against COVID-19.
As patients with AIRDs may respond inadequately to vaccines, it is important to identify a biomarker to assess the effectiveness of vaccination. This may allow prioritisation of patients for booster doses of the vaccine for those who have mounted an inadequate immune response. Estimation of antibodies to the receptor binding domain of the spike (anti-RBD) protein of SARS-CoV-2 is now widely available worldwide at a reasonable cost. Theoretically, it can be a robust biomarker to assess vaccination efficacy. Breakthrough infection in a cohort of healthcare workers was associated with levels of anti-RBD and neutralising antibodies.10 Again, not all laboratories have the capacity or the finances to measure neutralising antibodies. There is limited literature available on whether (total or IgG) serum anti-RBD could predict susceptibility to breakthrough infections. Current vaccination guidelines usually advise against estimating antibody levels post-vaccination.11 Also, there are data that even in the absence of such antibodies, cell-mediated immunity induced by vaccines might be protective against COVID-19.12
The two vaccines predominantly used in India are adenoviral vector-borne AZD1222 (ChAdOx1 nCoV-19, AstraZeneca COVID-19 vaccine “Covishield”) and the indigenous whole-virion β-propiolactone-inactivated BBV152 (Bharat Biotech COVID-19 Vaccine “"Covaxin”). Currently, India has crossed over one billion vaccination doses.13
We are prospectively following up a cohort of vaccinated AIRD patients. Their antibody titres against the SARS-CoV-2 Spike protein were measured at 4–6 weeks after the second dose of COVID-19 vaccination. We have previously shown that antibody titres have a good correlation with neutralisation assays in patients with AIRD.14 We have also explored the effects of past symptomatic COVID-19 on postvaccination humoral response.15 Our previous work has not included assays for cell-mediated immunity. Hence, one question remained whether postvaccination antibodies titres are predictive of susceptibility to breakthrough SARS-CoV-2 infections.
Thus, we prospectively followed up our cohort of 630 patients (for whom we had determined the antibody titres) to document postvaccination breakthrough infections. This survival analysis was to determine the strength of association between antibody titres and postvaccine breakthrough infections.
Methods
Objectives
Our objectives were to assess the incidence of breakthrough COVID-19 infection in a vaccinated cohort of patients with AIRDs and to study the relationship between anti-RBD antibody titres and serum viral neutralisation activity with the incidence of breakthrough infections.
Inclusion and exclusion criteria
Patients with AIRD who had completed both the doses of SARS-CoV2 vaccines were included from March 2021 onwards and followed up till October 2021 at the Centre for Arthritis and Rheumatism Excellence in Southern India. Patients with a prior diagnosis of COVID-19 infection were excluded to prevent confounding.
Clinical details
Demographic details, type of AIRD, immunosuppressive drugs, comorbidities, details of vaccination, were recorded. For RTX, exposure over the past 6 months was recorded as part of active treatment. The majority of our patients had received 500 mgof RTX for both induction and maintenance.16
Antibody assays
Serum samples for estimation of antibodies titres had been collected 4–6 weeks after the second dose of vaccine. IgG antibodies against the RBD of the spike protein were measured by ELISA using the Elecsys kit (Roche, Switzerland) as per the manufacturer’s instructions. Patients who had completed a minimum of 8 weeks interval after the second dose of vaccine were followed up every 2 monthly telephonically till the end of October 2021. For the analysis, patients were classified based on their anti-SARS CoV2-S antibody titres into good responders (GR) (>212 IU), inadequate responders (IR) (0.8–212 IU) and non-responders (NR) (<0.8 IU). This was based on our previous work where a receiver operator curve (ROC) had shown that antibody titres above 212 predicted more than 30% neutralisation by sera, with a sensitivity of 81.5% and a specificity of 83.6%.15
Assessment of breakthrough infections
Every patient was contacted telephonically at an interval of 2 months. Details of exposure to COVID-19 contact, testing for COVID-19, breakthrough infection and severity of the infections were recorded. A COVID-19 contact was defined as per WHO recommendations.17 The severity of infection was categorised as asymptomatic, mild, moderate and severe as per WHO criteria18
Neutralisation assay
Also, all sero converted patients who had known contact with a COVID-19 case had their sera analysed for neutralisation against the delta variant of SARS-CoV-2 virion particles using the SARS-Cov2 sVNT kit (GenScript, Piscataway, New Jersey, USA). This was to determine whether the presence of neutralising capability of the antibodies could give additional information regarding susceptibility to break-through infections. As seronegative individuals are unlikely to have neutralising antibodies, they were excluded from this subgroup analysis.
Statistical analysis
Data are expressed as mean and SD or median and IQR based on the Shapiro-Wilk test for normality. Baseline characteristics were compared across the three groups (GR, IR and NR). A p<0.05 was deemed as statistically significant, all reported values were two sided.
For the survival analysis, ‘Survival’ and ‘Survminer’ R packages were used for the survival analyses. Kaplan-Meir (KM) survival curves were used to illustrate proportions of survival among the three groups (GR, IR and NR). Univariate analysis for age, sex, diagnosis, the vaccine used and various immunosuppressant drugs against break-through infections were analysed with the log-rank test. Drugs that were used in less than 20 individuals were not analysed. All parameters that had p<0.10 in the univariate analysis were included in the multivariate analysis using Cox regression. The censuring events for both the cox models and the KM models were only breakthrough infection.
Results
Baseline characteristics
We studied 630 patients of AIRDs with an average age of 55.2 (±11.6) years and a male to female ratio of 1:5.2. Table 1 contains details of the cohort including the background rheumatic disease, vaccine received, comorbidities and immunosuppressants used. The majority of patients had received AZD1222 (495, 78.6%) while the remaining received the BBV152 vaccine. Around a quarter (179, 28.4%) had at least one other comorbidity beyond AIRD. Methotrexate and hydroxychloroquine were the most commonly used disease modifying anti-rheumatic drugs (DMARDs). Of 360 patients who were on methotrexate, 21 patients withheld it 1–2 weeks postvaccination whereas the others continued it.
Baseline characteristics, N=630
Response to the SARS-CoV-2 vaccine
The mean antibody titres were 854.1 (±951.9)with the majority being GR (380, 60.3%). Of the remaining, 143 (22.7%) were classified as IR and 107 (16.9%) as NR.
Predictors of antibody response
On univariate analysis, the type of vaccine received was a significant determinant of response. 70% of patients developed a GR to AZD1222 whereas 56% of patients were NRs among recipients of BBV152 vaccine (p=0.001; Fisher Exact test). Based on the subtype of AIRD, the majority of patients with RA were GR (261, 62.9%) as opposed to patients with SSc (8, 44.1%) in whom the majority were NR. Systemic lupus erythematosus (SLE), vasculitis, other CTDs did not show a significant difference in the response rates.
Among the drugs, only MMF was significantly different between the three groups (table 2). Analysing antibody levels using a generalised linear model with age, disease, gender, presence of comorbidities, the vaccine used and drug usage as predictors, only the vaccine used (AZD1222 vs BVV152) and the use of methotrexate were significantly associated with lower antibody titres (online supplemental table 1).
Supplemental material
Comparison of characteristics among good, inadequate and non-responders
Breakthrough infections
At a mean follow-up of 147 (±53.7) days, breakthrough infections had occurred in 47 patients (7.4%) of which 4 (8.5%) were asymptomatic, 37 (78.7%) had mild, 4 (7.4%) moderate and 2 (3.7%) severe disease. Breakthrough infections were highest in the NR group (19/107, 17.75%), followed by the IR group (13/143; 9.09%) and least in the GR group (15/380; 3.95%).
An additional 22 patients had a positive COVID-19 contact but tested negative on RT-PCR for SARS-CoV-2.
Predictors of breakthrough infections
As mentioned above, the proportion of breakthrough infections was highest in the NR group and lowest in the GR group (p=0.01; analysis of variance). The KM curve illustrating the probabilities of survival from breakthrough infection in the three groups is provided in figure 1 with overall survival of 96% for GR, 91% for IR and 82% for NR. The overlap between the 95% CIs (shaded colours) demonstrated how similar rates are between the three groups.
Kaplan-Meier survival curves for survival from breakthrough infections in three types of vaccine responders (non-responder: antibody titres <0.8 IU/mL; inadequate responder: titres 0.8–212 IU/mL; good-responders: titres >212 IU/mL).
Online supplemental figure 1 is the KM survival curve for breakthrough infection in patients who had been administered the two different vaccines (94% for AZD1222 and 87% for BBV152). Table 3 summarises the results of univariate analysis (log-rank test) of different variables versus breakthrough infections. Univariate analysis is not reported for tacrolimus and tumour necrosis factor inhibitors since there were less than 10 patients on these drugs.
Univariate analysis showing the association of the mentioned variable with postvaccination breakthrough infections
Antibody response, vaccine, gender, MMF, RTX and steroid use had an association of p<0.1 with breakthrough infection, and these were modelled in the Cox proportionate hazards regression. Figure 2 shows the HRs of which only NR was significantly associated with breakthrough infections. ROC analysis did not reveal any cut-off antibody titre that could predict breakthrough infections since the area under curve of the model was less than 0.5. This is likely because the ROC does not incorporate the time element (which determines exposure to the virus) as in survival analysis.
HRs from Cox regression modelling for survival from breakthrough infections including types of vaccine responders (non-responder: antibody titres <0.8 IU/mL; inadequate responder: titres 0.8–212 IU/mL; good-responders: titres >212 IU/mL), gender (1=male), rituximab (RTX=1 implies exposure within last 6 months), mycophenolate mofetil (MMF=1 implies exposure within the last 3 months of vaccination) and steroid use (1=any steroid use within 3 months of vaccination). GR, good responders; IR, inadequate responders; NR, non-responders.
Neutralisation assays
Patients with a history of exposure with detectable antibodies in the sera (GR or IR) underwent estimation of virion particle neutralisation by their sera. The average neutralisation percentage by sera was significantly higher (p<0.01) for those who did not develop infection (42.9, 95% CI 16.8 to 59.6) compared with those who developed the infection (14.8, 95% CI –12.6 to 39.5) (figure 3). A minimum of 30% neutralisation by sera was achieved by 7 of 28 (25%) of those infected (despite having positive antibody titres) and 13 of 20 (65%) among those exposed but not infected had (p<0.01, OR 2.1 (95% CI 1.2 to 4.2)). The median anti-RBD antibodies were numerically higher in patients who remained negative (1091±947 IU) versus those who tested positive (654±864 IU) but the difference was not significant (p=0.08).
(A) Antibody titres in COVID-19 exposed who had breakthrough infections versus those who did not have. (B) Neutralisation assays in COVID-19 exposed who had breakthrough infections versus those who did not have.
Discussion
The cohort of 630 vaccinated patients had been divided into GR (60.3%), IR (22%) and NR (16.9%) based on their anti-RBD antibody titre 4–6 weeks postvaccination. Breakthrough infections occurred in 7.4% of patients and were associated with non-response to vaccination. This provides evidence for using post-vaccination antibody titres as a biomarker to assess successful vaccinationin patients with AIRD.
Breakthrough infections were higher (7.4%) in our study as Kerala (a state in Southern India) was facing the second wave due to delta variant during this period when the study was conducted.13 There are two studies, one each from Europe and the United States, which reported lower breakthrough infections (<1%, 4.7%, respectively) in patients with AIRD postvaccination.8 9 They had a similar patient profile with the most common disease subtype being RA. However, there are differences: first, most of these patients had received mRNA vaccines whereas our patients had been vaccinated with AZD1222 and BBV152 as per national guidelines and vaccine availability. Second, the majority of our patients were on oral conventional DMARDs as compared with the other two cohorts. Also, many other factors would be different between the three continents such as infection rates, population density, usage of public transportation, local travel restrictions and behavioural patterns. These factors can explain the difference in breakthrough infection rates. The severity of infection in our cohort was similar to those in the other cohort with over 90% being symptomatic.
Anti RBD antibody response to vaccines was similar to other cohorts with 16.9% being seronegative in ours as compared with 14% from Israel (post-BNT162b2 vaccine).19 In contrast, two German cohorts reported lower rates of sero negativity (6% and 0%) post-BNT162b2 vaccine3 6 however their sample sizes were smaller. One of the determinants of antibody response was the type of vaccine used, with poorer responses in those who received the inactivated BBV152 vaccine consistent with results previously reported in healthy controls as well as AIRD patients.14 20 21 The number of our patients on RTX might be proportionately smaller and hence it might not reach a statistical significance due to a type 2 error. As we follow a low dose RTX protocol at our centre, most of our patients had received a lower dose of RTX which also could have contributed to a better humoral response to the vaccine. Furthermore, we have observed that patients with detectable B cells mount a good humoral response to the vaccine despite having received RTX.16 22 However, in our cohort, the proportion of patients on MTX was higher and we interestingly found MTX to be associated with lower antibody levels even on generalised linear modelling. Other cohorts have also reported similar responses with MTX users.3 7 Most of our patients had not withheld MTX before vaccination as there were no national or local guidelines recommending this at that time.
Our most important finding is the relationship between antibody levels and breakthrough infections. The KM curve (figure 1) demonstrates the rising probability of breakthrough infections with lower antibody levels. Though not significant in multivariate analysis, even those who had antibody titres less than the cut-off of 212 IU had numerically more infection. This implies that the antibody titre may predict risks for further infections.
Telephonic follow-up of patients at regular intervals and a state government policy of testing all the primary contacts even if they are not symptomatic enabled us to have a cohort who did not develop disease despite having a high-risk contact. Neutralisation assays were done in the subset of patients with positive antibodies who were exposed to COVID-19 as per WHO definition. The data clearly showed that COVID-19 contacts who did not develop disease had a higher proportion of neutralisation of virion particles by their sera. They also had higher titres of total antibodies though not significantly different. This might be a type II error due to limited numbers (48) in this subgroup analysis. A good correlation between neutralising antibodies and anti-RBD has been reported previously.14 15 Though neutralisation assays may provide some additional data about susceptibility, it remains to be seen what the practical benefit of such data is.
One caveat of this study is that we have not assessed the role of T cell immunity in the protection against COVID-19. This study was not aimed at looking at T cell factors. However, antibody titres being sufficient to predict breakthrough infections do not exclude the role of T cells. T cell responses may go hand in hand with humoral responses.12 What is more relevant in the clinical context is that the antibodies alone are sufficient as a biomarker for the risk of future infections.
Our study has some limitations. We have assessed antibody levels at only a single time point (4–6 weeks after the second dose). Thus, we are unable to comment on how the rate of change in antibody titres may influence SARS-CoV-2 infection risks. Although waning of immunity may be a risk factor for breakthrough infection after 6 months of the second dose, it is unlikely to have contributed to the breakthrough infections in our cohort as mean follow-up was less than 6 months. Second, we have not analysed the effects of disease activity scores on breakthrough infections due to heterogeneity between the different diseases and drugs used. Underlying immune deregulation and immunosuppressant use can undermine the vaccine efficacy leading to higher breakthrough infections compared with healthy controls.23 And lastly, our cohort is different from other previously described cohorts in the DMARD usage. The majority of the patients were on MTX or HCQ while those on tofacitinib or biologicals were less than 10%. Due to lower numbers in these sub-groups, they might not have reached statistical significance when their association with vaccination response was compared.3 5 24 25
In conclusion, antibody titres appear to predict susceptibility to breakthrough infections, and the absence of an antibody response is the strongest predictor of breakthrough infection on multivariate analysis. This should propel us to explore the routine use of post-vaccination antibody titres as a biomarker. Patients with inadequate humoral response to two doses of vaccines can be prioritised for a booster dose.
Supplemental material
Data availability statement
Data are available on reasonable request. Data will be available from the corresponding author on reasonable request.
Ethics statements
Patient consent for publication
Ethics approval
The study was approved by the Ethics Committee of Sree Sudheendra Medical mission (IEC/2021/35) and written informed consent of all the participants was taken.
References
Supplementary materials
Supplementary Data
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Footnotes
Handling editor Josef S Smolen
Twitter @sakir_rheum, @PanktiMehta24, @drdpshenoy
Contributors PS, VS and SAh designed the study . Patient enrolment and data collection was done by PS,AP, SC, SAn, SJ, VS and KKN. Manuscript was draft by SAn and PM. PS accepts full responsibility for the work and/or the conduct of the study, had access to the data and controlled the decision to publish.
Competing interests SA has received honorarium as speaker from Pfizer, DrReddy’s, Cipla and Novartis (unrelated to the current work). The other authors declare no conflicts of interest.
Provenance and peer review Not commissioned; externally peer reviewed.
Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.