This study used data from the Scientific Registry of Transplant Recipients (SRTR). The SRTR data system includes data on all donor, wait-listed candidates, and transplant recipients in the US submitted by the members of the Organ Procurement and Transplantation Network (OPTN). The Health Resources and Services Administration (HRSA), U.S. Department of Health and Human Services provides oversight to the activities of the OPTN and SRTR contractors. The study was deemed exempt by the Mayo Clinic Institutional Review Board (INC8014532).

In the SRTR standard analysis file we identified all patients who received an SPK transplant between 1 January 2014, and 31 March 2024. During the study period, valganciclovir prophylaxis was routinely administered to SPK recipients for 3 months for R+ cases or up to 6 months for D+/R− cases. CMV-naïve recipients with a CMV-negative donor (D−/R−) did not receive valganciclovir prophylaxis [17, 23]. We excluded recipients on the basis of induction regimen (missing; mixed; or other than rabbit anti-thymocyte globulin, alemtuzumab, and interleukin-2 receptor agonist), maintenance regimen (missing, or other than tacrolimus and mycophenolate mofetil with or without corticosteroids), crossmatch data (missing, positive, or weakly positive), and CMV serostatus (missing). We categorized recipients into three risk groups on the basis of recipient and donor CMV serostatus: low-risk, D−/R−; intermediate-risk, R+; and high-risk, D+/R−, consistent with the risk stratification endorsed by major transplant and infectious diseases societies.

The primary outcomes of interest were recipient and overall allograft survival by donor-recipient CMV serostatus risk category. Death-censored allograft survival was also evaluated. Short-term outcomes included 1-year rates of hospitalization, rejection of kidney alone, rejection of pancreas alone, and rejection of either organ.

Continuous variables were summarized as means and SDs and compared by using analysis of variance or pooled t tests. Categorical variables were summarized as counts and percentages and compared with the χ² test.

Time-to-event data were summarized with Kaplan-Meier estimates of incidence through 7.5 years post transplant. The log-rank test was used to compare groups. Cox proportional hazards models (referred to as multivariable models) were used to evaluate the effect of CMV serostatus risk category on outcomes of interest, with adjustment for the following possible confounding variables: age, sex, ethnicity, diabetes type, preemptive transplant, dialysis duration, induction type, corticosteroid maintenance, HLA antigen mismatch, calculated panel reactive antibody, local vs. imported organs, pancreas donor risk index, transplant year, and donor-recipient Epstein-Barr virus status. The center was entered as a random effect in the multivariable models. Linearity in all tests was evaluated by using splines for continuous variables. We used Schoenfeld residuals plots to test the assumption of proportionality.

All analyses were performed with R version 4.2.2 (R Foundation for Statistical Computing). Statistical significance was defined as P < 0.05.

We identified 7,847 patients who underwent SPK transplant during the study period. The final analysis cohort consisted of 4,744 SPK recipients with complete data: 831 low-risk (D−/R−), 2,671 intermediate-risk (R+), and 1,242 high-risk (D+/R−) (Figure 1).

Table 1 details the demographic characteristics of recipients and donors by CMV serostatus. Recipients had a mean (SD) age of 42.3 (9.2) years and body mass index (calculated as weight in kilograms divided by height in meters squared) of 25.7 (4.2); these features were similarly distributed across risk groups. Overall, 61.6% of recipients were men, but the R+ group had a significantly lower share of men than the other groups (P < 0.001). Black and Hispanic recipients were significantly more represented in the R+ group (P < 0.001).

Recipients in the R+ group were significantly less likely than those in the D–/R–and D+/R–groups to receive a preemptive transplant (15.1% vs. 18.2% vs. 19.4%, respectively) (P = 0.002) and were on dialysis for significantly longer than recipients in the other groups (mean, 2.1 vs. 1.8 vs. 1.8 years; P < 0.001). The pancreas donor risk index and proportion of locally procured organs did not differ between groups. The proportion of organs procured after circulatory death was significantly higher in the D–/R–group than in the R+ and D+/R–groups (3.9% vs. 2.7% vs. 1.5%, respectively; P = 0.004). The groups also differed significantly in terms of diabetes type, preemptive transplant, calculated panel reactive antibody, induction type, Epstein-Barr virus status, and donor age and ethnicity (Table 1).

One-year outcomes post transplant are shown in Table 2. Kidney rejection rates and pancreas rejection rates did not differ significantly between risk groups, but the combined kidney or pancreas rejection rate did. The combined rejection rate was significantly higher in the D+/R− group than in the D–/R–and R+ groups (16.6% vs. 11.8% vs. 14.4%, respectively; P = 0.02). The D+/R− group was also hospitalized significantly more frequently than the D–/R–and R+ groups (67.2% vs. 60.0% vs. 63.0%, respectively; P = 0.005).

In the Kaplan-Meier analysis of recipient survival (Figure 2), the D+/R− group had the lowest overall survival (log-rank P = 0.03). The 7.5-year survival probabilities were 82.1%, 82.9%, and 76.4% in the D−/R−, R+, and D+/R–groups.

The overall kidney survival probability (Figure 3) was significantly lower in the D+/R− group than in other groups (log-rank P = 0.01). The 7.5-year overall kidney survival probabilities were 77.1%, 73.8%, and 68.1% in the D−/R−, R+, and D+/R–groups. Figure 4 shows the death-censored survival of kidney allografts. The D−/R− group had a significantly higher probability of graft survival than the other groups (log-rank P = 0.001). The 7.5-year death-censored kidney graft survival probabilities were 90.8%, 84.6%, and 81.5% in the D−/R−, R+, and D+/R–groups.

The overall pancreas allograft survival probability (Figure 5) was significantly lower in the D+/R− group (log-rank P < 0.001). The 7.5-year overall pancreas survival probabilities were 73.8%, 71.0%, and 62.8% in the D−/R−, R+, and D+/R–groups. Figure 6 shows the death-censored pancreas graft survival probabilities. The D−/R− group had a significantly higher probability of graft survival than the other groups (log-rank P = 0.002). The 7.5-year death-censored pancreas allograft survival probabilities were 87.3%, 82.5%, and 78.4% in the D−/R−, R+, and D+/R–groups.

In the multivariable model for recipient death and graft loss, compared with R+ serostatus, D+/R− serostatus was significantly associated with a 28% higher risk of death (P = 0.045) and a 25% higher risk of pancreas graft loss (P = 0.009) (Table 3). It was also associated with a 20% higher risk of kidney graft loss, but the effect was not significant (P = 0.06). CMV D−/R− serostatus was not associated with altered risks of death or pancreas or kidney graft loss. However, in the death-censored graft loss model, D−/R− serostatus was significantly associated with a 25% lower risk of pancreas graft loss (P = 0.04), and a 34% lower risk of kidney graft loss (P = 0.03), compared with R+ serostatus. In the death-censored model, D+/R–serostatus was significantly associated with a 31% higher risk of kidney graft loss (P = 0.04) compared with R+ serostatus.

The causes of death and kidney and pancreas allograft failure are detailed in Supplementary Tables S1–S3. Notably, the specific cause of death was not reported for most recipients. Cancer as a cause of death was reported more frequently in the D+/R–group than in the other groups (Supplementary Table S1).

The causes of kidney and pancreas graft failure were not reported for approximately half of the recipients. Kidney graft rejection was the leading documented cause of kidney graft loss in the D+/R− group (Supplementary Table S2). Primary nonfunction was the most common documented cause of pancreas graft loss (Supplementary Table S3). Pancreas rejection was the cause of graft loss more frequently in the D+/R− group than in the other groups.

To our knowledge, this study represents the largest and most comprehensive to date in documenting the long-term outcomes of SPK transplant recipients on the basis of donor-recipient CMV risk profiles. It was designed to isolate the effects of donor-recipient CMV risk profiles by including only conventional-risk recipients with crossmatch-negative transplants and by using a standardized maintenance regimen. Primary outcomes were based on well-documented metrics from the SRTR.

However, the study has limitations. First, the retrospective design prevents full adjustment for unmeasured confounders. Second, the SRTR standard analysis file has substantial variability in center reporting practices. For example, although the cohort was limited to recipients discharged on a standard maintenance regimen of tacrolimus and mycophenolate mofetil, the standard analysis file lacks consistent data on postdischarge changes to immunosuppressive regimens. This restricts the ability of researchers to analyze variances in immunosuppression exposure, or to assess tolerability of SPK transplant recipients to immunosuppression. Third, the SRTR does not capture granular longitudinal data on the duration of CMV prophylaxis, the magnitude of CMV infection (e.g., viral load) or disease, antiviral drug resistance or management, donor-specific antibody formation, or late rejection episodes. Therefore, it is difficult to determine whether the outcomes observed in this study were primarily influenced by active CMV infection, given that delayed-onset CMV infection remains a common phenomenon in SPK transplant [19, 22]. Finally, the lack of access to a biorepository or T-cell profiling limits the ability of researchers to examine the relationship between primary CMV infection and the immune system in SPK transplant recipients.

In this large cohort of SPK transplant recipients, having a high-risk donor-recipient CMV serostatus discordance (D+/R−) was associated with a significantly higher risk of death and overall graft loss. In contrast, concordant-negative CMV serostatus (D−/R−) was associated with significantly higher death-censored survival of both kidney and pancreas grafts. The high-risk group also had the highest rates of complications, including rejection and hospitalization, whereas the low-risk group had significantly lower rates of these outcomes.

These findings underscore the potential benefit of matching CMV-seronegative transplant recipients with organs from CMV-seronegative donors. Implementing such matching strategies could improve overall survival rates for recipients and allografts and help with CMV prevention. However, further research is needed to evaluate the potential effects of extending wait times for seronegative organs and to explore the feasibility of revising the allocation policies.