Chronic kidney disease (CKD) is a common multifactorial condition with a rising global prevalence. Current estimates suggest that about 850 million people worldwide live with kidney disease, underscoring its public health burden (1). In Poland, recent epidemiological data indicate that up to 4 million individuals have CKD, most of whom are undiagnosed, highlighting registry under-ascertainment and the need for system-level action (2, 3). The large population potentially requiring renal replacement therapy (RRT) makes it essential to optimize survival and quality of life. Kidney transplantation (KTx) remains the preferred treatment for end-stage renal disease based on its clinical benefits, economic considerations, and superior health-related quality of life compared with other RRT modalities. It is also the only RRT that restores endocrine kidney function (4, 5). Although short-term outcomes after transplantation have improved substantially, long-term graft survival, particularly death-censored survival, has modestly improved beyond the first decade. This has kept attention on ultra-long survivors and on the early, potentially modifiable determinants of late outcomes. Here, we examine pre- and peri-transplant factors associated with ≥25-year graft survival and consider how these signals could inform clinical quality improvement. Cohorts explicitly focusing on ultra-long survivors remain scarce; available ≥20–25-year reports are mostly single-center, use heterogeneous definitions, and are often underpowered for multivariable analyses (6, 7). Cold ischemia time (CIT) is one of the few clearly modifiable center-level determinants of transplant outcomes. Contemporary studies link a longer CIT to early complications such as delayed graft function (DGF) and, indirectly, to worse longer-term performance. In current practice, strategies such as hypothermic machine perfusion may mitigate risk when the CIT is prolonged (8). However, evidence linking pragmatic, center-level CIT targets to ≥25-year survival remains limited, which is a practical gap that our study addresses (6). Early immunologic injury is central: first-year biopsy-proven rejection (clinical or subclinical) is repeatedly associated with worse long-term trajectories, although granular, time-varying modeling of the rejection burden is rarely feasible in historical datasets. Recent multicenter work also indicates that the nature and timing of rejection episodes shape later risk (9). Cardiometabolic comorbidities are common after KTx. Hypertension affects approximately 70–80% of recipients at 1 year and is linked to cardiovascular disease, the dominant cause of late morbidity and mortality; however, optimal blood pressure targets and their translation into improved graft survival remain debated (10). Similarly, post-transplant diabetes mellitus (PTDM) is associated with worse patient and graft outcomes, supporting proactive screening and immunosuppression choices that consider glycemic risk (11). Simple early functional markers (e.g., 6–12-month serum creatinine) retain prognostic value and can serve as actionable waypoints; however, most studies focus on 3–5-year horizons rather than ≥25-year outcomes (12, 13). Accordingly, in a single-center cohort transplanted in 1980–1995, we assess whether early, potentially modifiable signals (CIT, first-year rejection burden, and early kidney function profiles) are enriched among recipients with ≥25-year graft survival and can be translated into auditable, center-level quality metrics (7).

The aim of this study was to delineate early and potentially modifiable factors associated with ultra-long kidney allograft survival (≥25 years) versus early graft failure (≤10 years) in a cohort transplanted in 1980–1995. We also sought to translate these associational signals—particularly CIT, first-year rejection burden, and early biochemical profiles—into explicit, auditable, hypothesis-generating targets for clinical practice.

This manuscript constitutes the third, analytical stage of a coherent research program conducted in a single center and within a single transplantation era (1980–1995). Stage I (14) was a descriptive study of a cohort of 45 recipients with well-functioning grafts ≥25 years post-transplant, without a comparator arm, focusing on associations between kidney function parameters and hemoglobin (HGB)/creatinine 10 years after KTx. Stage II (15) expanded this characterization to 59 ultra-long-term graft survivors (ULS) and reported detailed immunosuppression and long-term course without a control arm. Stage III does not repeat descriptive results; it introduces, for the first time, a designed comparison between the ULS cohort and a control arm with early graft failure (EGF), performs hypothesis-driven testing, and applies multivariable models—qualitatively distinguishing it from earlier publications and justifying a separate report.

Baseline recipient characteristics were broadly comparable between groups (Table 1). The main between-group contrast was a shorter cold ischemia time (CIT) in the ULS, with the total ischemia time showing a concordant pattern. Preoperative cytotoxic antibody levels were low and similar in both groups.

Donor and procedural characteristics were broadly similar between groups among cases with available documentation (Table 2).

Living-donor transplants were rare in both groups. Follow-up in ULS extended beyond 25 years by design, whereas EGF was defined as graft failure within 10 years. The distribution of recipient sex and primary kidney disease was broadly similar, with chronic glomerulonephritis predominating among cases with a documented etiology. Pre-transplant RRT patterns differed descriptively in the available records, but dialysis exposure (duration of hemodialysis and peritoneal dialysis) did not differ significantly between groups (Table 1). Peritoneal dialysis–related peritonitis occurred in both groups without a clear between-group signal. ABO compatibility was complete in both groups, except for one ULS pair with a missing donor blood type. Rh compatibility was documented in a similar proportion of pairs, and HLA matching categories were broadly comparable between groups (Table 3). Where documented, donor age and sex distributions were similar between groups, as were donor–recipient sex concordance and procedural descriptors such as graft laterality and implant site (Table 2). Early postoperative courses were broadly similar between groups, including rates of acute tubular necrosis and the need for post-transplant dialysis (Table 4). Within the first year, clinically recognized rejection was less frequent in ULS than in EGF, and the treatment indicators suggested a lower overall rejection burden in ULS (Table 5).

Longitudinal laboratory profiles favored ULS at later time points (Table 6). Although the creatinine level on postoperative day 1 was higher in the ULS group, serum creatinine levels at 6 months, 1 year, and 5 years were consistently lower in the ULS group, consistent with better, longer-term graft function. Hemoglobin was higher in the ULS group at 5 years, whereas leukocytes and platelets did not show meaningful between-group differences over time. Electrolytes were largely similar, with only small differences at isolated time points (Table 6).

At the last follow-up, the ULS recipients were older, whereas the EGF recipients experienced graft failure at a younger age. Chronic rejection was the leading cause of graft loss in EGF, with a smaller subset showing concurrent or recurrent glomerular disease, while acute rejection was the terminal event in a minority of cases. Other causes were infrequent.

The univariate analyses of the binary and nominal factors are summarized in Table 7. The denominators varied because of incomplete archival documentation. Two signals were notable: ULS more often lacked documented post-transplant carbohydrate metabolism disorders and more often had no documented/treated rejection within the first year, consistent with a lower early immunological and metabolic burden.

Consistent with Table 5, the ULS experienced fewer rejection-related events during the first post-transplant year. This was evident both for clinically recognized rejection episodes and for treatment-based indicators, supporting a lower first-year rejection burden in ULS (Table 5).

Consistent with the baseline contrast, CIT differed between groups in univariate analyses. In an exploratory univariable logistic regression with recognized rejection within 1 year as the outcome, a longer CIT was associated with higher odds of recognized rejection (Figure 1). The fitted curve is presented to visualize the observed association and should not be interpreted as a validated prediction model.

End-stage renal disease continues to increase, making the improvement of long-term outcomes after kidney transplantation a central clinical goal. Better late outcomes reduce return to dialysis, alleviate waitlist pressure, and improve quality of life while reducing costs (19). The distinct contribution of our study is its focus on ultra-long survivors (≥25 years) and on early, observable signals that can be operationalized for within-center audits and quality improvement, rather than reiterating broad, non-actionable associations.

A key constraint is the limited effective sample for multivariable adjustment owing to archival missingness. Therefore, we treat the adjusted estimates as supportive and emphasize consistency across complementary analyses (group contrasts, regression, and time-to-event approaches). Accordingly, the findings are presented as hypothesis-generating and suitable for a within-center audit, pending external validation. Although one-year outcomes after KTx have improved substantially, longer-term gains have been less consistent across studies, and late graft loss remains multifactorial (6, 19–23). Some cohorts report continued improvement in 10-year graft survival despite an increasingly complex recipient profile, whereas others emphasize that progress beyond the first decade has been modest and uneven (19–23). This tension in the literature supports a pragmatic focus on early, observable signals that may be amenable to center-level optimization. Contemporary ultra-long survivor reports (>20 years) further suggest that excellent late function clusters with specific donor and immunologic characteristics, yet detailed evidence beyond 20 years remains limited (7). Against this background, our dataset isolates a small set of actionable levers observable within the first post-transplant year that are enriched among ultra-long survivors.

Across analyses in our cohort, fewer first-year rejection events and a lower rejection-treatment burden were consistently associated with the ultra-long survivor phenotype. This aligns with prior work linking acute rejection—particularly later or more severe episodes—to worse long-term trajectories, with risks shaped by timing, persistence, and histologic phenotype (19, 24–26). Importantly, our rejection metrics reflect recognized and treated episodes captured in historical records rather than protocol-biopsy-based subclinical injury, which likely attenuates the measured associations. Overall, the direction of the effect is consistent with the concept that minimizing early immunologic injury is a plausible pathway toward better long-term graft durability. Cardiometabolic comorbidities (hypertension and PTDM) were highly prevalent and documentation-sensitive in this archival dataset and are therefore discussed primarily as long-term care targets and context variables, rather than as independent determinants of ultra-long graft survival based on our multivariable evidence. A minority of reports have suggested paradoxical associations in which early treated rejection correlates with better long-term outcomes, potentially reflecting treatment intensification and residual confounding rather than the protective effect of rejection itself (27). In our data, lower first-year rejection burden consistently characterized ULS; therefore, we interpret rejection primarily as an avoidable early injury signal and propose that simple tracking of 0/1/≥2 antirejection treatment cycles could serve as a feasible unit-level audit metric rather than a prescriptive clinical recommendation (9). Recent data from the latest era similarly link longer CIT, DGF, and immunologic contexts to early rejection risk and inferior early function, reinforcing the rationale for a first-year rejection-free pathway for prevention and rapid treatment.

Histocompatibility and immunosuppression are well-established determinants of graft outcomes. In our cohort, HLA matching is presented descriptively, but stable multivariable testing was not feasible because key covariates were incompletely captured, and complete-case restriction reduced the effective sample. Maintenance immunosuppression regimens varied from 1980 to 1995 and were not uniformly documented. Therefore, we avoid underpowered regimens or HLA comparisons and instead emphasize early, auditable signals (CIT, first-year rejection burden, and early graft function profiles) that warrant validation in contemporary multicenter datasets. In our study, the pre-transplant RRT modality and duration of HD/PD did not differ meaningfully between the groups in the available records, and any descriptive contrasts should not be over-interpreted. Prior literature on PD versus HD before KTx remains mixed; some studies report better patient survival among those treated with PD, while others find no consistent association with long-term graft outcomes (5, 28, 29). Given the descriptive nature of our data and limited statistical signals, we treat the RRT modality primarily as a context for education and logistics, rather than as a causal lever for ultra-long survival.

In this study, the key factor for long-term KTx outcomes was the shorter CIT. The literature includes many studies describing the relationship between CIT and transplanted kidney function. CIT is a modifiable pre-transplantation risk factor that affects graft survival. Debout et al. (30), in their research, demonstrated that every hour of CIT can impact both graft and patient survival after KTx. Hariharan et al. (19), in their work, state that reducing the severity of perfusion injury with colder ischemia time may contribute to improving long-term kidney survival outcomes in the case of deceased donors. Similarly, Rego et al. (16), in their conclusions, note that in many scientific studies, one of the main factors influencing long-term survival of recipients is a low CIT. The results of Dziewanowski et al. (31), who analyzed the outcomes of Polish patients over a long-term period of more than 15 years post-transplantation and a short-term period of less than 6 years, also indicate that CIT is a strong independent factor inversely correlated with long-term graft survival. Newer paired-kidney analyses confirm that even relatively small increments in CIT increase the risk of DGF and suggest that machine perfusion may mitigate CIT-related injury—evidence that supports our proposed CIT time-budget as a pragmatic, center-level quality target (8).

Cardiovascular disease is the dominant cause of late morbidity and mortality after KTx, and hypertension is highly prevalent in this population (17, 32). In our archival dataset, hypertension was common in both groups, and any between-group contrast was borderline and sensitive to documentation completeness; therefore, we treat it as descriptive rather than as evidence of an independent determinant of ultra-long graft survival. Nevertheless, prior studies suggest that higher blood pressure (BP) after transplantation is associated with inferior graft outcomes, and BP control remains a central long-term management target (33–37). Importantly, among ultra-long survivors, patient outcomes may become increasingly driven by cardiovascular comorbidity even when graft survival is favorable, reinforcing the need to integrate BP and cardiometabolic stewardship into long-term pathways (7). We also note the behavioral dimension of long-term graft success. Depressive and anxiety symptoms are common in kidney replacement therapy and can undermine adherence to immunosuppression and follow-up, with downstream consequences for graft and patient outcomes (38–41).

Post-transplant diabetes mellitus (PTDM) and related carbohydrate metabolism disorders are common after KTx and are associated with cardiovascular complications, graft loss, and mortality (11, 42). In our study, carbohydrate metabolism disorders were more frequent among early failures but did not consistently remain significant after parsimonious adjustment; accordingly, we treat this signal as hypothesis-generating and documentation-sensitive. Regardless, proactive screening, PTDM-conscious immunosuppression, and metabolic management remain reasonable vigilance strategies. These conclusions are consistent with recent pooled evidence indicating higher all-cause and cardiovascular mortality and increased graft failure risk in PTDM (11). Early graft function markers have long been linked to downstream outcomes, and several studies report that lower creatinine in the first year after KTx characterizes longer graft survival (21, 27, 43). In our cohort, the 6-month creatinine level showed the most useful within-sample prognostic separation, whereas the 1-month creatinine level and other routine laboratory measures did not meaningfully discriminate between outcomes. This supports treating 6- and 12-month creatinine levels as feasible auditable key performance indicators (KPIs) for within-center monitoring and hypothesis-generating risk stratification (12). More broadly, landmark or dynamic prognostic frameworks can help operationalize repeatedly measured markers in future prospective datasets (13). Beyond baseline determinants, 6-month creatinine functions as a practical landmark marker that can support within-center audits and hypothesis-generating follow-up prioritization. Recipients with higher 6-month creatinine levels had markedly greater odds of falling into the EGF ≤ 10 group, whereas 1-month creatinine levels and other routine laboratory measures did not significantly discriminate outcomes. This supports embedding 6- and 12-month creatinine levels as auditable quality indicators alongside CIT reduction and first-year rejection prevention.

Cold ischemia time emerged as the most consistent and actionable early factor associated with ultra-long graft survival, supporting its prioritization as a modifiable, center-level quality target. Six-month serum creatinine levels may serve as a pragmatic landmark indicator for within-center audits; however, they should not be interpreted as a validated decision threshold. Accordingly, we propose an auditable framework centered on minimizing the CIT, preventing first-year immunologic injury, and routinely monitoring early kidney function markers (e.g., 6–12-month creatinine) to support long-term horizon care planning. Because ultra-long survivors consistently showed a lower first-year rejection burden, centers should emphasize prevention, timely diagnosis, and simple audit-friendly tracking of rejection events and treatment intensity.

Metabolic risk warrants proactive management through early screening and PTDM-conscious immunosuppression, whereas early biochemical markers (particularly creatinine in the first year) can help guide follow-up intensity within audit programs. Other baseline procedural descriptors showed no meaningful signal in this dataset and should not be prioritized over the core levers: minimizing CIT, preventing first-year rejection, and optimizing cardiometabolic risk.