COVID-19 mortality may be reduced among fully vaccinated solid organ transplant recipients

Background Solid organ transplant (SOT) recipients are at increased risk for morbidity and mortality from COVID-19 due to their immunosuppressed state and reduced immunogenicity from COVID-19 mRNA vaccines. This investigation examined the association between COVID-19 mRNA vaccination status and mortality among SOT recipients diagnosed with COVID-19. Methods & findings A retrospective, registry-based chart review was conducted investigating COVID-19 mortality among immunosuppressed solid organ transplant (SOT) recipients in a large metropolitan healthcare system in Houston, Texas, USA. Electronic health record data was collected from consecutive SOT recipients who received a diagnostic SARS-CoV-2 test between March 1, 2020, and October 1, 2021. The primary exposure was COVID-19 vaccination status at time of COVID-19 diagnosis. Patients were considered ‘fully vaccinated’ at fourteen days after completing their vaccine course. COVID-19 mortality within 60 days and intensive care unit admission within 30 days were primary and secondary endpoints, respectively. Among 646 SOT recipients who were diagnosed with COVID-19 at Houston Methodist Hospital between March 2020, and October 2021, 70 (10.8%) expired from COVID-19 within 60 days. Transplanted organs included 63 (9.8%) heart, 355 (55.0%) kidney, 108 (16.7%) liver, 70 (10.8%) lung, and 50 (7.7%) multi-organ. Increasing age was a risk factor for COVID-19 mortality, while vaccination within 180 days of COVID-19 diagnosis was protective in Cox proportional hazard models with hazard ratio 1.04 (95% CI: 1.01–1.06) and 0.31 (0.11–0.90), respectively). These findings were confirmed in the propensity score matched cohort between vaccinated and unvaccinated patients. Conclusions This investigation found COVID-19 mortality may be significantly reduced among immunosuppressed SOT recipients within 6 months following vaccination. These findings can inform vaccination policies targeting immunosuppressed populations worldwide.


Methods
The study population consisted of all consecutive SOT recipients diagnosed with COVID-19 between March 1, 2020, and October 1, 2021 at Houston Methodist, Houston Texas, USA. SARS-CoV-2 testing was performed as part of the routine screening on SOT recipients. Patients were followed up for outcomes up to December 1, 2021. The Houston Methodist J.C. Walter Jr. Transplant Center is located within the large urban, Texas Medical Center and maintains an active transplant and follow-up program including liver, heart, lung, pancreas, kidney, and multi-organ transplants. COVID-19 mRNA vaccines became available to highrisk populations in the study setting in December 2020, and vaccination programs utilizing the NT162b2 (Pfizer-BioNTech) or the mRNA-1273 (Moderna) vaccines for transplant patients began in January 2021. Demographic and clinical data were retrieved from the Houston Methodist COVID-19 Surveillance and Outcomes Registry (CURATOR), a COVID-19 specific electronic health records (EHR) data mining, surveillance and collection project [21]. Vaccination records and medication data were abstracted directly from the EHR. The Charlson Comorbidity Index was calculated from medical history components as a measure of overall comorbidity burden [22]. SOT recipients were included if they received a positive diagnostic result from a SARS-CoV-2 RNA polymerase chain reaction (PCR) assay or a viral antigen assay at Houston Methodist while undergoing immunosuppressive treatment.
Patients were considered 'fully vaccinated' in this study if they had received the second dose of either the NT162b2 or the mRNA-1273 vaccines at least 14 days prior to their first positive SARS-CoV-2 diagnostic test, according to CDC definitions [23]. Patients who tested positive for COVID-19 before the completion of the vaccination course were considered 'unvaccinated'. Demographic and clinical data were reported as frequencies and proportions for categorical variables and as median and interquartile range (IQR) for continuous variables. Differences in covariates across vaccination status groups were evaluated using the chi-square or Fisher's exact tests for categorical variables and Kruskal Wallis test for continuous variables, as appropriate. Timepoints for primary and secondary outcome analyses were determined from the distribution of time-to-event data. Cox proportional hazards modeling was performed to determine the characteristics associated with COVID-19 mortality within 60 days. Multivariable logistic regression modeling was utilized to determine the characteristics associated with an ICU admission within 30 days. Variables for the multivariable models were selected on the basis of potential clinical relevance and using the Bayesian Information Criterion. Only baseline characteristics were assessed in regression analyses, as eligibility for specific COVID-19 treatments, including remdesivir and monoclonal antibodies depended on additional factors, such as disease severity upon presentation and time from symptom onset [24,25]. Given the vaccine effectiveness drop after 6 months [26,27], we classified the vaccination status as follow: vaccination status was presented as 'vaccinated <180 days before COVID-19 diagnosis', 'vaccinated 180+ days before COVID-19 diagnosis', 'unvaccinated, diagnosed with COVID-19 in 2020', or 'unvaccinated, diagnosed with COVID-19 in 2021.' Unvaccinated patients were categorized by whether they were diagnosed with COVID-19 in 2020 or 2021 as a proxy for advances in COVID-19 treatments which could have influenced survivorship. We conducted a non-replace propensity score (PS) matching (ratio 1:1, caliper 1) between patients with and without vaccination. The matching criteria included age, race, Charlson Comorbidity Index (CCI), year of transplantation, COVID-19 treatments administered at diagnostic encounter (remdesivir, monoclonal antibodies, azithromycin, methylprednisolone, ribavirin, tocilizumab, dexamethasone), immunosuppressant treatments at COVID-19 diagnosis (antithymocyte globulin, tacrolimus, cyclosporine, mycophenolate, azathioprine, sirolimus, everolimus, belatacept, prednisone). Balance of the matching criteria between groups was evaluated using standardized bias percent. Multivariable analysis was also conducted on the PS cohort for mortality with 60 days and ICU admission within 30 days. All analyses were performed on Stata MP version 17.0 (StataCorp LLC, College Station, TX, USA). This retrospective registry-based study was approved by the Houston Methodist institutional review board (PRO00025320) and granted a waiver of informed consent.
COVID-19 incidence among SOT recipients followed local epidemiologic trends, displaying four distinct peaks in April 2020, June 2020, January 2021, and August 2021 [28] (Fig 1). Following the initial peak, during which testing and treatment policies varied widely [21,29], 30 day mortality and 30 day ICU admission rates remained relatively stable, even as the proportion of patients diagnosed with COVID-19 after completing a vaccination course increased.  (Fig 1). The primary outcome of COVID-19 mortality within 60 days of diagnosis was used in survival analyses. Of the 66 patients who died from COVID-19 within 60 days of diagnosis, 57 (86%) were unvaccinated, 4 (6%) had been vaccinated less than 180 days before being diagnosed, and 5 (8%) had been vaccinated 180 days or more before being diagnosed. In multivariable Cox proportional hazard models, increasing age was a risk factor for 60 day COVID-19 mortality with hazard ratio (HR): 1.04 (5% CI: 1.01-1.16); p-value: 0.001, while vaccination within 180 days of diagnosis was protective, compared to unvaccinated patients diagnosed in 2021, HR: 0.31 (95% CI 0.11-0.90); p-value: 0.03. Notably, vaccination at 180 days or longer was not protective against COVID-19 mortality, HR 0.70 (95% CI 0.26-1.85); p-value: 0.47; furthermore, mortality risk did not vary significantly between SOT recipients diagnosed in 2020 and unvaccinated SOT recipients diagnosed in 2021 (Table 2).
Patient characteristics in the PS matched cohort was present in Table 4, which showed all evaluated criteria were balanced between patients with and without vaccination. The Cox regression model run on the PS matched cohort confirmed the association between the vaccination and lower mortality within 60 days with a multivariable HR of 0.28 (95% CI 0.08, 0.94), p = 0.04 ( Table 5). The multivariable logistic regression model run on the PS matched cohort also found patients who received vaccination with 180 days from diagnosis had lower odds of ICU admission, OR 0.21 (95% CI 0.08, 0.51), p = 0.001 (Table 6).

Discussion
Our findings add to the growing body of literature demonstrating that, in spite of chronic immunosuppression and ostensibly a poor humoral response to the mRNA vaccines, mortality may be reduced among fully vaccinated solid organ transplant recipients who are exposed to COVID-19, compared to partially vaccinated or unvaccinated patients [30][31][32]. The presence of neutralizing antibodies could plausibly contribute to the observed reduction in COVID-19 case fatality among fully vaccinated SOT patients. However, new research indicates the production of neutralizing antibodies may also be hindered by the immunosuppressive drugs used in SOT recipients [33,34]. Therefore, future research is needed to identify additional mechanisms, such as a potentially protective role of cellular (T-cell) immunity, to explain the protective effect of vaccination without accompanying immunogenicity [35]. Additionally, further investigation should determine the effect of novel SARS-CoV-2 variants on disease dynamics within immunosuppressed populations. We also demonstrated that vaccination may offer waning protection from poor COVID-19 outcomes over time in an immunosuppressed population; SOT recipients in our cohort diagnosed with COVID-19 six months or more after being vaccinated were at similar risk for both COVID-19 mortality and ICU admission as patients who were never vaccinated. This study was restricted to patients without a known history of previous COVID-19 infection, and further research is needed to determine how vaccine immunogenicity may vary among immunosuppressed survivors of COVID-19.
Our study has several limitations. First, given this is a retrospective study, some clinical parameters were not collected in our dataset. For example, time from the first symptoms to diagnosis of COVID-19 was not available which may affect the outcome as we could not determine if the vaccinated patients were more likely to seek medical care earlier than unvaccinated patients. However, the analysis on the PS matched cohort confirmed that vaccination within 180 days of diagnosis was associated with a lower mortality within 60 days of diagnosis or ICU admission, independently with the treatment of remdesivir or monoclonal antibodies. Second, our data were obtained from one hospital system, which might not be generalized to other populations. Third, data of patients who had at least three doses of vaccines were not available in our patients for the time being. Of note, while patients in our study were considered as "fully vaccinated" with 2 vaccine doses as per the CDC definition, SOT recipients ages 5 years and older are recommended by the American Society of Transplantation to receive an initial three-dose series of mRNA vaccine followed by one or two booster doses if ages 12 or older [36]." Fourth, although we tried our best to rule out patients with a known history of previous COVID-19 infection, we may have missed some previous infections given lack of the confirmation of anti-N SARS-CoV-2 antibodies. Our analysis may also underestimate the patient outcome as we could not rule out completely the possibility that the patient sought further care at a different institution after their initial encounter at the Houston Methodist Hospital System. Finally, our findings may need to be considered with a gap between the data collected during the pandemic period at a single center and the current COVID-19 situation, especially with the fast-pace of evolving new SARS-CoV-2 variants and the change in the vaccine recommendations for SOT recipients from 2 doses to a primary 3-dose course followed by boosters.
Despite the limitations, our study has notable strengths. Our study is strengthened by a robust sample size of SOT recipients with COVID-19, a diverse, heterogeneous population, precise EHR-derived clinical information, and significant longitudinal follow-up of COVID-19 patients. While overall COVID-19 case fatality rates have decreased over time following the introduction of more effective treatment strategies [10], among our population of immunosuppressed SOT recipients, COVID-19 mortality and ICU admission rates did not vary significantly after the initial peak in April 2020. This investigation was a single-center, registrybased chart-review, so while patients could have experienced un-observed or un-documented outcomes at other institutions, all SOT recipients routinely received SARS-CoV-2 testing for surveillance purposes during the study period, regardless of symptoms. The observed difference between the fully vaccinated and unvaccinated patients can inform clinical practice and warrants additional studies as large-scale vaccination efforts continue. Moreover, the observed loss of protection within months of vaccination lends crucial evidence to discussions of booster recommendations in immunologically at-risk populations. Future investigations may prioritize longitudinal follow-up of vaccinated patients to determine incidence of additional markers of morbidity and poor health outcomes over time. This study demonstrated an association between mRNA vaccination and reduced mortality in solid organ transplant recipients; these results can contribute to the development of comprehensive vaccination programs targeting high-risk, immunosuppressed populations.