**4. Discussion**

During the study period, 103 solid-organ-transplant patients were diagnosed with SARS-CoV-2 at our institution, with a 9.7% SARS-CoV-2-specific mortality rate within three months of diagnosis. This finding was similar to our early institutional experience and to reports by others during the first year of the pandemic, before vaccinations had become broadly available [7–9,11,17,18]. In our cohort, most of the infected patients were kidney-transplant recipients, which aligned with the higher prevalence of this transplant subgroup. Similar to our preliminary reports [8,18], the kidney-transplant-recipient SARS-CoV-2 mortality rate was 11.8% vs. 4.3% among the liver-transplant recipients, with a calculated relative risk of 2.7 (95% CI 0.36–20.3). There were no reported deaths among the 4 combined liver–kidney-transplant recipients who had tested positive for SARS-CoV-2.

Hypertension and diabetes were present in 12.7% and 15.9% of the deaths. Despite early reports by others, these comorbidities were not associated with increased mortality in our cohort. SARS-CoV-2 mortality increased with advancing age, a finding described in general population outcomes [34].

Mirroring the practice of decreasing or discontinuing MMF in the presence of a viral infection, such as Cytomegalovirus, the antimetabolite dose was decreased or held for two weeks from the time of SARS-CoV-2 diagnosis. Our study failed to demonstrate any significant MMF effect on SRS-CoV-2-related mortality. However, more than 60% of patients who were taking MMF at the time of diagnosis had this medication held or decreased.

In our patient cohort, Tacrolimus demonstrated a protective effect (HR = 6.1, *p* = 0.022), an observation already reported by others [35]. A meta-analysis of 11 cohort studies investigating the impact of immunosuppression on SARS-CoV-2 suggested that Tacrolimus usage did not impact mortality or SARS-CoV-2 infection severity [36]. In our cohort, only eight (7.8%) patients had been on Cyclosporine at the time of the SARS-CoV-2 infection, three of whom died. In the univariate Cox regression, Cyclosporine was associated with a 6.1 (95% CI 1.6–24) death risk, contrary to a favorable 0.23 (95% CI 0.064–0.81) when using Tacrolimus as a Calcineurin inhibitor. These findings do not necessarily imply causation and should therefore be interpreted with caution; the findings may be attributed to the small patient sample and/or lack of control of confounding variables, including, but not limited to, the underlying indication for the switch to Cyclosporine from Tacrolimus, which has been the standard of care in our institution.

Two out of five (60%) patients who had been on Belatacept at the time of SARS-CoV-2 diagnosis eventually succumbed to the disease (HR = 6.1, *p* = 0.022). The literature is largely limited to case studies on the impact of Belatacept on SARS-CoV-2 outcomes. As a T-cell co-stimulation inhibitor, Belatacept is theorized as a potential mitigator of the cytokine storm caused by SARS-CoV-2 infection; however, it has also been shown to potentially increase the risk of severe opportunistic infections [37,38]. Similar to Cyclosporine, it remains unclear if this apparent positive correlation of Belatacept with severe SARS-CoV-2 infection reflects causation; since Belatacept is a choice often reserved for patients intolerant to CNIs and/or with a significant cardiovascular burden or recent cardiac events, there might be confounders that have not been identified in this small population sample, such as the underlying indication of the patient being switched to Belatacept. Like in the case of Cyclosporine, it may be the underlying comorbidities that led to the immunosuppressionregimen switch rather than the immunosuppression choice *per se*, as the factors impacting the disease outcome.

As scientific evidence evolved along with the pandemic progression, treatment for SARS-CoV-2 for both inpatients and outpatients at this institution changed over the course of this study, in alignment with the federal guidelines and transplant organizations' recommendations. Monoclonal-antibody therapy was recommended for SARS-CoV-2-positive transplant recipients managed in the outpatient setting and became available near the end of the study period in December 2020. A total of 21 (20.38%) patients in this study received monoclonal-antibody therapy. Remdesivir and convalescent plasma were also used for inpatients meeting certain criteria. A total of nine (8.74%) patients received Remdesivir, and five (4.85%) received convalescent plasma. While the impact of these treatments was not analyzed as part of this portion of the study, it is reasonable to consider that their use may have mitigated the mortality in these patients, particularly towards the latter stages of the cohort, when antibody treatments became standardized and broadly available, particularly for the higher-risk subgroups.

These data were collected over a period when SARS-CoV-2 vaccination was not widely available, therefore providing an opportunity to assess the viral infection fatality in our immunosuppressed population prior to the broad implementation of SARS-CoV-2 vaccines.

Age cohorts stood out as remarkable predictors of outcome and provided for a more robust analysis. No patients died in the 20–51-year age group and patients in the >72-year group had the least survival probability (75%, *p* < 0.001). Other studies have found age to be one of the most important factors in predicting SARS-CoV-2 mortality. This study, combined with data from existing works, is perhaps suggestive of the need to provide more robust, earlier intervention in the older transplant population [39–42]. Novel treatments such as MABs, antiviral agents, and most importantly preventative measures, could prove particularly life-saving in this older group of unvaccinated SARS-CoV-2 positive transplant recipients.

Statistical limitations existed in this study due to the small sample size. This led to severe model instability when a multivariate Cox regression analysis was attempted, as well as some instability of the univariate regression model. Model instability is particularly prevalent in fields with zero covariates in the fatality group. An extension of this study is currently ongoing to capture a larger study population in the attempt to build a stable model for analysis.
