Push toward pre-emptive kidney transplantation – for sure? Authors: Orsolya Cseprekal, Christian Jacquelinet, Ziad Massy Published: 2024 ABSTRACT ABSTRACTPre-emptive kidney transplantation (PKT) has long been considered the optimal treatment for patients with end-stage chronic kidney disease (CKD) seeking the most favourable long-term outcomes. However, the significant growth in transplant procedures over recent decades has led to a notable increase in wait-listed patients and a disproportionate demand for donor organs. This situation necessitates a re-evaluation of transplantation timing and the establishment of rational indications from both societal and clinical perspectives. An increasing number of retrospective analyses have challenged the universal benefit of PKT, suggesting that premature indications for living or deceased donor PKT may not always yield superior hard outcomes compared with non-PKT approaches. Conventional predictive models have shown limitations in accurately assessing risks for certain subpopulations, potentially leading to significant disparities among wait-listed patients. To address these challenges, we propose the following considerations. Prediction models should not only optimize the distribution of our limited donor resources, but should also illuminate foreseeable risks associated with a potentially ‘unsuccessful’ PKT. Therefore, this article seeks to underscore the necessity for further discourse on the smouldering concept of when and for whom living or deceased donor PKT should be considered. Is it universally beneficial, or should the clinical paradigm be re-evaluated? In the endeavour to attain superior post-PKT survival outcomes compared with non-PKT or conservative treatment, it seems critical to acknowledge that other treatments may provide more favourable results for certain individuals. This introduces the intricate task of effectively navigating the complexities associated with ‘too early’ or ‘unsuccessful’ PKT. INTRODUCTION Kidney transplantation has consistently been reported as the treatment of choice for end-stage chronic kidney disease (CKD) patients, providing enhanced survival rates, improved cardiovascular outcomes and better quality of life compared with dialysis [1, 2]. All-cause mortality decreases following transplantation but does not attain the survival rates of the general population [3, 4]. The substantial growth in kidney transplant procedures over the past 3 decades has led to gradual improvement of the surgical experiences, which along with the enhanced efficacy of immunosuppressive protocols have resulted in prolonged patient survival and better graft longevity [5–9]. Pre-emptive kidney transplantation (PKT) has been considered the foremost treatment choice, delivering optimal outcomes for individuals with end-stage CKD [10–12]. Nevertheless, PKT accounts for 13–19% of all kidney transplant procedures in Europe, with half of these being from living donors and the other half from deceased donors. Despite the efforts made by the transplant community, these numbers have been improving slowly for years [13, 14]. For now, the growing number of wait-listed patients face a disproportionate demand for donor organs. As a result of concerted efforts to promote living organ donation, the number of living donors has increased significantly in several countries. However, the number of deceased donors has not shown a comparable increase, particularly in the context of PKT. This disparity underscores the ongoing challenges in expanding the overall organ donor pool [15]. Despite progress in improving patient and graft survival, optimizing waiting times for transplantation remains a challenge [14, 16–19]. As transplant nephrologists, we advocate for ‘transplanting the patient first’ before initiating other renal replacement therapies (RRTs), given the negative impact of prolonged pretransplant dialysis exposure [20]. However, PKT decisions should not solely rely on estimated glomerular filtration rate (eGFR) thresholds but should also consider clinical and biochemical evidence, particularly uraemic symptoms, as recommended by the DESCARTES in 2016 [21–30]. Observational studies and medico-economic analyses suggest that PKT may not universally benefit all patients, with outcomes varying based on specific factors. Ongoing research is crucial to ensure optimal patient outcomes in this field [31–36]. It is apparent that living donor kidney transplantation is favoured over deceased donor kidney transplantation. This preference stems from factors such as the ability to schedule the transplantation at an optimal time, the availability of higher-quality kidneys, the flexibility offered by kidney exchange programs to enhance HLA and ABO matches and being the most favourable choice for PKT [37, 38]. Nonetheless, the fluctuating availability of living donors and the limited supply of deceased donors prompt us to contemplate the most efficient utilization of our resources to achieve optimal outcomes for our patients. This approach aims to steer clear of the worst-case scenario of PKT and adhere to the timeless principle of ‘primum non nocere’, or ‘do no harm’ [39–44]. PKT has the potential to alleviate the financial strain of dialysis costs [45], however, prioritizing deceased donor kidneys to those in the eligibility window rather than those who have just met transplantation criteria may shift the utilization of the deceased donor supply from optimal to exploitative. Differences in demographics and allocation policies rooted in varying health insurance systems lead to disparities among geographical regions [46–49]. Careful review of retrospective observational studies from diverse regions is essential, ensuring transparent methodologies to avoid overestimating the net advantages of living or deceased donor PKT compared with other transplant modalities, dialysis or conservative therapy. Healthcare professionals aim to address and analyse this issue systematically, developing refined practice patterns to achieve the ultimate goal of improving patient care and outcomes [50]. While post-transplant care has improved patient and graft survival rates, pretransplant predictive tools need careful reconsideration to optimize PKT availability in clinical practice. Determining ‘how’ and ‘when’ to estimate the risk of failure before deeming a patient transplantable remains crucial. Challenges in predicting unsuccessful transplants and optimizing scarce donor resource allocation continue to be critical areas requiring further elucidation [51]. In this narrative review and opinion paper, we examine the clinical indications and potential advantages of PKT over alternative therapies. We explore ambiguities, particularly regarding the timing, and emphasize the need for careful interpretation of retrospective studies on PKT. We also address societal, economic, clinical and decision-making challenges associated with PKT and highlight open research questions in this field. CLINICAL AND EXPLANATORY ASPECTS OF RECENT PRACTICE Challenging clinical aspects of pre-emptive transplantation It was about 14 years ago that the utility of PKT started to be negotiated and cultivated by clinicians for increasing the trend of wait-listing at a pre-emptive stage in order to transplant at a higher level of eGFR. However, no hard evidence proved that PKT had a net survival benefit of early versus late initiation of RRT. Even if eGFR itself appeared to not serve as an ultimate tool to indicate pre-emptive wait-listing or PKT, none of the studies provided reliable methods to predict the prognosis of PKT versus non-PKT [52]. During the last couple of years, more and more retrospective observational studies have prompted clinicians to consider the concept of ‘transplant first’ with more precautions. Table 1 summarizes studies published after 2010 that examine the potential advantages of PKT compared with other RRT strategies. While the majority of these studies provide supportive evidence for PKT, a few original papers and meta-analyses have yielded inconclusive results. These findings underscore the potential necessity for individualized assessment of certain patient populations prior to transplantation. This suggests that in specific cases, the implementation of a transplantation strategy may require careful reconsideration and a tailored approach. (The literature collection and searching strategy is available in the supplement material.) Table 1:Related evidence of PKT.EvidenceReferencesDesignSample sizeArms and donor type (N, L or D, comparator)Follow-upOGOMDCGLDWFGOther outcomeMain conclusionsIn favourMorales et al. (2015) [78]Single centre26DPKT versus non-PKT35.5 ± 20.1 monthsX–X–ARDCGL better in PKT. AR more frequent in PKTGoto et al. (2016) [79]Single centre786LPKT versus non-PKT61 months (35.3–94)––XXCVDPKT associated with lower risk of DWFG or DCGL. CVD associated with ABOi, duration of dialysis, DMJay et al. (2016) [80]UNOS47 108LPKT versus <1 year dialysis versus >1 year dialysis5 years–XX–Dialysis recipients of <1 year had similar OM to PKT but decreased DCGLPrezelin-Reydit et al. (2019) [54]Nationwide France22 345L + D, L/DPKT versus non-PKT4.7 yearsX–––PKT was associated with a decreased OG when compared with pretransplant dialysis, whatever the duration of dialysis, even in the first 6 months. The effect of PKT on OG was stronger in LMerzkani et al. (2022) [81]Multicentric5752L + DPKT + non-PKT3.5 years (2.0–6.4)––XXDWFG associated with age, dialysis, CKD, DM. DCGL associated with cPRA, CKD (GN, TIN)Rana Magar et al. (2022) [36]Meta-analysisa22 622L + D, L/DPKT versus non-PKTNAX–––ARPKT is superior versus non-PKTNaylor et al. (2022) [19]Nationwide Canada4461L + DPKT versus non-PKTNAXXXXHOSP-INF, CVDLonger time on dialysis was associated with an increased rate of several adverse post-transplant outcomesPrezelin-Reydit et al. (2022) [10]Nationwide Francea1911L + D, L/DPKT versus non-PKT7 yearsX–––PKT was associated with a 55% reduction of the OG at any time after KT independently of the duration of dialysisAzegami et al. (2023) [55]Meta-analysis719.5bL + D, L/DPKT versus non-PKTNA–X–XPKT reduces OM, DWFG, DCGL versus non-PKT regardless of L or DChallengeJung et al. (2010) [82]Single centre476L + D, L/DPKT versus non-PKT5 years and 10 yearsX–––OG in PKT versus non-PKT is NS and does not associate with dialysis duration. Predictors are donor type, CKD, AR and recipient ageKessler et al. (2011) [83]Multicentric1585DPKT versus non-PKT5 yearsXX––Neither PKT nor time on dialysis was associated with OM or OGGrams et al. (2011) [52]Nationwide USA19 471L + DPKT by eGFR3 years–XX–No survival advantage to PKT at higher levels of eGFRLuo et al. (2012) [84]Single centre164DPKT versus non-PKT5 yearsXX––PKT was comparable to non-PKT in terms of OM and OG, however, PKT showed lower rate of AR and DGF and better QoLGram et al. (2013) [85]SRTR121 853L + DPKT versus <1 year dialysis versus >1 year dialysis15 years–XX–PKT survival is similar to <1-year dialysis groupSayin et al. (2013) [86]Single centre100L + DPKT versus non-PKT5 yearsXX––Non-PKT was non-inferior to PKT, but PKT accompanied with better comorbid conditionsNakamura et al. (2015) [87]Single centre96LPKT versus non-PKT with short-term dialysis versus non-PKT with long-term dialysis5 yearsXX––UTI, AR, CMV, HOSP-INFNo significant difference in OM and OG of PKT versus short-term dialysis group. UTI was more frequent in long-term dialysis group. AR was more frequent in PKTUnsal et al. (2015) [88]Single centre334L + DPKT versus non-PKT3 yearsXX––PKT does not provide better outcomes compared with non-PKTOkumi et al. (2017) [89]Single centre1060LPKT versus non-PKTPKT versus non-PKTc2 weeks––X–CVD, AR, CMV, PTDM, TUGraft failure and CVD was reduced in PKT in the entire cohort, but PKT was non-inferior to non-PKT after PSMFranco et al. (2019) [42]Single centre66L + DPKT versus non-PKTc55 monthsXX––Difference is NSFoucher et al. (2019) [31]DIVAT1138DPKT versus non-PKTc3.2 years (0–16.9)XX––Difference is NSKim et al. (2019) [32]Single centre1984LPKT versus non-PKTc68.9 ± 40.3 months–XX–OM and DCGL did not differ between PKT versus non-PKTc if dialysis duration <19 monthsFerreira et al. (2024) [90]Single centre589L + D groups by <12 months on dialysis versus >12 months on dialysis154 monthsXX––No association between dialysis time and outcomesKumar et al. (2022) [91]UNOS49 675L/D PKTEarly versus late PKTNA–X––Early PKT does not have benefit versus KT after <180 days of dialysisaPaediatric.bMedian of 719.5 (44–121 853).cPropensity score matched: /separately analysed; + jointly analysed.ABOi: ABO0 incompatible; AR: acute rejection; CI: confidence interval; CMV: cytomegalovirus; cPRA: calculated panel reactive antibody; CVD: cardiovascular disease; DCGL: death-censored graft loss; D: deceased donor; DGF: delayed graft function; DIVAT; Données Informatisées et VAlidées en Transplantation; DM: diabetes mellitus; DWFG: death with functioning graft; GN: glomerulonephritis; HOSP-INF: hospitalization due to infection; L: living donor; NS: not significant; OG: overall graft loss; OM: overall mortality; PTDM: post-transplant diabetes mellitus; TIN: tubulointerstitial nephritis; QoL: quality of life; SRTR: Scientific Registry of Transplant Recipients; TU: tumour; UTI: urinary tract infection. Related evidence of PKT. Paediatric. Median of 719.5 (44–121 853). Propensity score matched: /separately analysed; + jointly analysed. ABOi: ABO0 incompatible; AR: acute rejection; CI: confidence interval; CMV: cytomegalovirus; cPRA: calculated panel reactive antibody; CVD: cardiovascular disease; DCGL: death-censored graft loss; D: deceased donor; DGF: delayed graft function; DIVAT; Données Informatisées et VAlidées en Transplantation; DM: diabetes mellitus; DWFG: death with functioning graft; GN: glomerulonephritis; HOSP-INF: hospitalization due to infection; L: living donor; NS: not significant; OG: overall graft loss; OM: overall mortality; PTDM: post-transplant diabetes mellitus; TIN: tubulointerstitial nephritis; QoL: quality of life; SRTR: Scientific Registry of Transplant Recipients; TU: tumour; UTI: urinary tract infection. The matter of clinical decision-making A paradigm shift in demonstrating clinical decision-making in studies may help to distinguish different risk situations, predict individual risk and ultimately avoid early death or graft failure in pre-emptively wait-listed patients. Reliable risk prediction assets and comprehensive benchmarking tools for predicting the probability of death and graft loss estimating the potential success or failure of PKT versus remaining on regular CKD therapy or receiving other renal replacement modalities are currently unavailable, are needed and are yet to be explored. Fig. 1 provides the concept of the multifaceted issue of decision-making in PKT from a clinical and methodological point of view. Multifaceted aspects of clinical decision-making as of recent practice in pre-emptive transplantation. Societal perspectives Highlighting key examples demonstrates that transplanting patients pre-emptively, even either a living or deceased donor kidney, associates with better graft survival and all-cause mortality compared with non-PKT according to some results. Conversely, several recently published papers have not addressed the long-standing question regarding the optimal timing of intervention, a topic that has been debated in the transplant community for ≈15 years [12, 53]. In a subsequent analysis, PKT versus non-PKT has been shown to be clearly associated with better graft and patient survival at any time after transplantation {n = 20 935; hazard ratio [HR] 0.55 [95% confidence interval (CI) 0.47–0.67], P < .0001}, although a subgroup of non-PKT patients registered on waiting lists even before starting dialysis benefited in graft survival [n = 4162; HR 0.59 (95% CI 0.45–0.78), P = .0002] but not in all-cause death with a functioning graft [HR 0.87 (95% CI 0.62–1.23), P = .44] [54]. The incidence of cardiovascular disease, biopsy-proven acute rejection, cytomegalovirus infection and urinary tract infection remain comparable between PKT and non-PKT patients. The certainty of evidence of the antecedent and the potential low-grade evidence of the latter secondary outcomes assume that there is supposed to be a subgroup of patients for whom PKT might bring disadvantages compared with remaining on regular therapy and getting a transplant at a later stage of CKD, or probably after the initiation of dialysis [55]. Recently some authors found that PKT decreases overall mortality [adjusted HR 0.78 (95% CI 0.66–0.92)] and death-censored graft survival [adjusted HR 0.81 (95% CI 0.67–0.98), P < .05] as compared with non-PKT patients, but others reported that pre-emptive living donor transplantation (PLKT) patients have a lower risk of death compared with non-PLKT patients, while pooled adjusted analysis did not show a significant difference between groups. It should be noted that in the same analysis, the risk of death of PKT with a deceased donor was not shown to be different from non-PKT [56]. Economic perspectives The expenses of dialysis are much higher than those related to PKT. The perspective of PLKT seems to be the best among all potential RRT modalities, both from a societal and economic point of view [57]. Deceased donor transplantation independent of the antecedent dialysis is the most economical way to gain quality-adjusted life years (QALY) compared with dialysis [58]. However, a Canadian study [59] recently reported that ‘although allocating deceased donor kidneys pre-emptively to patients seemed to be the best option from the patient perspective in terms of quality adjusted life years, allocating deceased donor kidneys to patients receiving dialysis was a better use of a scarce resource from a societal perspective.’ Authors forewarn to consider who and when to transplant pre-emptively, because the cost of transplanting pre-emptively wait-listed patients was found to be higher if the annual probability of a need for dialysis was <45%. Thus pre-emptively wait-listed patients with an eGFR of 15–20 ml/min/1.73 m2 were likely to gain fewer QALY at a greater cost. At the same time, deceased donor PKT resulted a net positive cost–benefit ratio if the patient spent at least 4 years on dialysis. Clearly seen was that the cost of PKT and the potential risk of harmful consequences of immunosuppression were not evident in patients with a higher eGFR at the time of transplantation compared those with 10 ml/min/1.73 m2. This study suggested the need of an individual risk assessment to define the safest and the most efficient transplantation window [59]. Challenging explanatory aspects of pre-emptive transplantation Who should be compared to whom? Our knowledge on the mortality risk of transplant patients compared with that of CKD patients on regular therapy at the same level of eGFR are incomplete. This knowledge gap was highlighted by Cheddani et al. [60], who reported that kidney transplant patients at the age of 60 years, at least 1 year after transplantation, had a 2.7-fold greater risk (95% CI 1.6–4.7; P = .0005) of mortality than propensity score–matched CKD counterparts with the same level of eGFR (≈41 ml/min/1.73 m2), independent of the fact that they had been wait-listed before or after the initiation of RRT. There was no difference in cardiovascular events (kidney transplant 29% versus CKD 22%), but the death rate due to infections was higher in the transplant group (19.4% versus 10%). This retrospective study demonstrated for the first time that the comparator at the same level of eGFR had better mortality indicators, whereas the article also highlighted the problem of whether eGFR alone is sufficient to indicate the need for RRT or additional clinical factors should be taken into consideration. This was the first study to posit the idea of comparing transplanted versus age, BMI, eGFR and baseline disease–based propensity score–matched CKD controls. In this study, the hypothesis could not have been examined without accepting the robust design and unaccustomed risk set and several type of biases (selection, immortal time and treatment bias), which hampers the generalizability of the results. Moreover, it seems that an intention-to-treat approach would be more appropriate to explore the real potential benefits of PKT at the first risk-evaluation phase in a transplant setting. Differences between living and deceased donor transplantation have been evidenced by several studies thus far. However, all of them used different comparators. The most comprehensive analysis, conducted by Magar et al. [49], reported that living donor PKT results in better overall mortality compared with living non-PKT; however, pooled adjustment did not show the same advantages. Concerning deceased donor transplantation, PKT does not demonstrate advantages over non-PKT. Thus it is notable that donor type (living or deceased donor) might influence the overall mortality, but the pre-emptive or non-pre-emptive manner of transplantation does not seem to be associated with overall mortality. This study focuses on a specific aspect of the multifaceted clinical decision-making process faced by clinicians. Do biomarkers exist to predict post-transplant outcome? Randomized controlled trials could not be implemented to show the net benefit of PKT versus other modalities because of the parachute effect [61]. Only retrospective data showing that no evidence are available regarding who to transplant pre-emptively at a certain level of eGFR. Although some evidence and clinical practice has taught us the unwritten rules of transplantation and immunosuppression, thus defining absolute and relative contraindications of transplantation, we still use eGFR as the objective indicator for placing patients on the waiting list or starting dialysis [62, 63]. Although PKT is generally recommended when the eGFR is <15 ml/min/1.73 m2, the optimal timing for PKT is still unclear [39]. Some studies have reported that Black patients are less likely or are likely to be wait-listed at a lower level of eGFR. To approach equity in PKT, race-free eGFR was introduced. Although this step ameliorates the imbalance, further efforts are needed to find criteria for wait-listing and transplanting pre-emptively aside from eGFR and thus achieve the desired balance [64]. When eGFR reaches a critical point, other uraemic factors and the progression of CKD necessitate RRT. eGFR trajectories could theoretically help our decision-making, however, it has not yet become a part of our everyday practice [65–67]. Nevertheless, eGFR alone right before transplantation could scarcely predict post-transplant survival. Rather, it may refer to the relative risk of mortality at the time of the clinical decision [68]. Clear evidence for indicating pre-emptive wait-listing or PKT is not yet available. Some retrospective data analysis emphasizes patient groups who may benefit from earlier or even later transplantation. Thus current clinical guidelines should be individualized when indicating a particular type of RRT for a patient with end-stage renal disease. Additionally, this practice should follow locally available, medico-economically reasonable approaches while taking into consideration the estimated risk of post-transplant patients’ mortality. Challenges of estimating the risk difference Several approaches have been implemented to address the lack of randomized clinical trials in solid organ transplantation, aiming to avoid the known and accepted biases of retrospective observational studies, which might hamper approaching the evidence that PKT ensures the best quality life and longevity for all patient populations compared with other replacement modalities. Whereas none of them provides a simple tool to overcome the pitfalls of false positive or false negative risk estimation of PKT patients, it remains questionable which clinical parameters are favourable for predicting the potential risk and when is the right time for risk estimation in patients not on dialysis awaiting transplantation. In kidney transplantation, the primary evidence relies on retrospective observations due to ethical and practical constraints on randomization. Causal inferences are drawn from clinical databases and registries, but the lack of unified data and limited pretransplant information hinders comprehensive risk assessment. A crucial challenge is developing tools to estimate post-transplant risk at treatment initiation or organ allocation. Table 2 illustrates the complexity of predicting outcomes at wait-listing, considering various clinical scenarios that could influence PKT success or failure. Contradictory scenarios exist where transplantation might increase mortality risk but reduce graft loss probability, requiring careful consideration. Given the disproportionately low PKT rate and given that the intention to prepare patients and donors for a living donor PKT might lead to patient selection bias in clinical studies, further investigation is necessary to address these issues and improve pretransplant risk assessment. Table 2:Risk classification from the clinical point of view. Difference in the risk of mortality after pre-emptive transplantation or other renal replacement strategies (increased risk of transplantation is highlighted in red, low risk of transplantation is highlighted in green, transitional colours refer to obscure clinical certainty).=: clinical worthiness is questionable (unnecessary or ‘too early’ transplantation); +: clinically beneficial; −: clinically detrimental; S: clinically significant objective decision; NS: clinically non-significant, subjective decision; QoL: probable improvement in quality of life. Risk classification from the clinical point of view. Difference in the risk of mortality after pre-emptive transplantation or other renal replacement strategies (increased risk of transplantation is highlighted in red, low risk of transplantation is highlighted in green, transitional colours refer to obscure clinical certainty). =: clinical worthiness is questionable (unnecessary or ‘too early’ transplantation); +: clinically beneficial; −: clinically detrimental; S: clinically significant objective decision; NS: clinically non-significant, subjective decision; QoL: probable improvement in quality of life. Rationale of potential biases Fig. 2 depicts the risk of mortality from the diagnosis to CKD. At the time of transplantation, the relative risk of death sharply increases to near 4-fold higher than that of those on the wait-list. The risk stabilizes ≈3 months after transplantation and clearly improves only ≈1 year after transplantation [3]. The amelioration depends on patient-, donor- and transplantation-specific features over time. This time dependence of the treatment cannot be omitted from the risk assessment in retrospective observational studies. One method is choosing the moment of risk assessment, i.e. defining clinical decision-making points. Clinical decision-making occurs first upon wait-listing and second at the moment of transplantation. The time of the PKT can only be anticipated in living donor transplants; for deceased donor transplants, this is occasional. Estimation of relative risk of mortality during the (A) peri-transplant course and (B) potential biases. RR: relative risk; DG: diagnosis; WL: waiting list; KTX: kidney transplantation. In retrospective databases, earlier entries may result in information bias due to the alteration of general patient profiles and improvement of care over time, leading to several consequences that must be addressed. Most of the time, pretransplant risk estimation remains incomplete, as registries suffer from a lack of details regarding CKD or uraemic burden, resulting in lead-time bias. Thus, missing information on the CKD burden in PKT patients may lead to false positive advantages when compared with those on dialysis and non-PKT patients. This problem with the control group accompanies all retrospective observational studies, raising the question of whether a CKD cohort at the same level of eGFR would be preferable as a control group. In recent retrospective studies, PKT is usually the exposure versus other treatment modalities, therefore, beyond lead-time bias, it creates selection or treatment bias. Since the relative risk of mortality worsens from wait-listing until transplantation, this factor cannot be included in statistical analyses without further aggravating lead-time bias [69]. Another potential reason for obscure, off-target predictive models could be due to unmeasured confounders. Certain variables are limited to the study period in which they were available and only the multiple imputation method can be used to reduce missingness at random [70]. PKT itself is a dynamic treatment. Positioning and fitting the right comparator is not so evident when using retrospective datasets. The proper risk set (population, time zero, exposure and outcome) is of clear importance. The optimal time of transplantation can be determined only in case of living donor transplantation. Studies combining living and deceased transplantation do not take the time dependence of the exposure into consideration. Presumably the risk prediction could be started at wait-listing in these cases as a definitive point of clinical decision-making. Emerging clinical challenges and refining methods Fig. 3 illustrates two critical junctures for estimating the probability of death or graft loss: at the time of wait-listing and at the time of transplantation. Notably, even pre-emptively wait-listed candidates with potential living donors may receive offers of deceased donor organs, which could be accepted. This significant step is incorporated into allocation systems in several European countries as a measure to safeguard potential living donors [14, 71]. However, the decision to accept a deceased donor kidney for a high-risk patient rather than continuing organ allocation and proceeding with living donor transplantation in these clinical scenarios is likely to necessitate individualized decision-making processes. This approach requires careful consideration of multiple factors specific to each patient's circumstances. At the time of transplantation, there is still an option to choose among different renal replacement strategies. The decision depends on the risk of death and graft failure for the recipients, the donor characteristics and the potential risks of the transplantation itself (transplant-specific risk factors), which become known at the time of allocation. Since similar risk stratification has not been implemented in recent analyses, the comparison of high-risk and low-risk patients in the same cohort could presumably result in contradictions regarding the utility and potential benefits of PKT versus non-PKT. Probability of death and graft loss in different clinical scenarios. Abbreviations: WL, waitinglist; KT, Kidney transplantation; D, deceased; L, living; PKT, preemptive kidney transplantation; PLKT, preemptove living kidney transplantation; PDKT, preemptive deceased kidney transplantation; deWL, delisting; nonWL, not waitlisted. There are several statistical methods for eliminating the biases listed in Fig. 2, providing useful approaches to avoid statistical distortions that could lead to false clinical inferences in the risk estimation of PKT. One potential statistical method suggested for evading the previously mentioned problems is the target trial emulation framework [72, 73]. Cloning, censoring and inverse probability weighting of treatment can appropriately reduce some of the biases. Propensity score matching could be built into the framework [74]. Although the approach almost evades the problems of the above-mentioned biases, time dependence of the treatment remains to be resolved to some extent. For instance, inverse probability weighting Kaplan–Meier curves of the treatment and control groups (pre-emptive transplant versus dialyzed and transplanted patients) may overestimate the survival of PKT patients due to information bias. Some groups have suggested applying a semi-parametrically efficient estimator of the time-varying effects for survival data with time-dependent exposure. Moreover, varying coefficients of Cox models has been suggested based on the global partial likelihood approach, remedying the statistical issues of the time dependence of the exposure (namely, that the time of transplantation is not predictable except for living donor transplantation) [75–77]. It is presumable that common analytical standpoints should be accepted and followed to conduct meta-analyses on the potential indicators associated with the outcome in PKT over other renal replacement modalities or over patients on regular CKD treatment with the same level of eGFR. Prediction modelling for PKT could have both medical and economic impacts. Identifying subpopulations likely to benefit more from early transplantation could guide PKT decisions, while estimating the potential risks of unnecessary transplantation at the time of wait-listing. The most important key clinical questions to resolve remain: How should patients be selected for PKT? Does PKT benefit everyone compared with later-stage transplantation? How should patients be selected for PKT? Does PKT benefit everyone compared with later-stage transplantation? How can complex prediction tools be developed by using markers beyond eGFR for PKT patient selection? How can complex prediction tools be developed by using markers beyond eGFR for PKT patient selection? How do we decide between conservative therapy and transplantation at specific eGFR levels? How do we decide between conservative therapy and transplantation at specific eGFR levels? How can concerns be addressed while evaluating results of retrospective observational cohort studies in PKT? How can concerns be addressed while evaluating results of retrospective observational cohort studies in PKT? CONCLUSION This article examines a small but significant aspect of PKT practice that could potentially lead to substantial improvements in outcomes: the importance of making informed decisions both at the time of wait-listing and at the time of kidney transplantation. While all forms of transplantation provide better outcomes than dialysis, our practical goal remains to minimize the necessity for dialysis treatment and prioritize kidney transplantation. However, this review underscores the critical need to achieve the optimal distribution of our limited donor resource while equally avoiding unsuccessful pre-emptive kidney transplantation. Based on unintentionally blurred reports of retrospective observational studies, PKT provides the best outcome, but there is a call for a more refined, clinically relevant, medico-economically balanced evaluation method to guide the utilization of scarce deceased donor resources and to appropriately identify the right timing for PKT in eligible patients. Routinely pursuing accurate survival prediction models for PKT versus non-PKT does not dismiss the potential advantages of prolonged conservative treatment before transplantation. The complex issue and potential pitfalls of ‘too early PKT’ is acknowledged as a significant clinical challenge and needs to be scrutinized more comprehensively in the future to move towards a practice of ‘PKT in due course’. Further studies and comparative risk estimations are needed to determine whether all individuals should be pushed toward PKT or if this inclination is just a speculative belief. More robust hazard prediction approaches focused particularly on the indications of deceased donor PKT are essential, with special consideration given to the enigmatic timing of the procedure. Supplementary Material