Reply: Chronic kidney disease after liver transplantation: Accurate measurement, early acute kidney injury, and renal-protective strategies

Abstract

In this article, we comment on the article by Sessa et al published in the recent issue of the World Journal of Transplantation, who revisit the three-decade single centre cohort of Muñoz-Serrano et al and conclude that chronic kidney disease (CKD) after liver transplantation (LT) is a modifiable determinant of long-term survival rather than an unavoidable consequence of calcineurin inhibitor exposure. We share their central position and welcome the weight they give to perioperative acute kidney injury (AKI), to calcineurin inhibitor minimisation through basiliximab induction and mycophenolate, and to the disciplined control of hypertension and metabolic disease. Our intention is to carry the argument one step further, because three problems lie beneath the literature they assemble and constrain what any inventory of risk factors can deliver. The first is metrological. Almost every reported association, including the signal attached to female sex, and the reported incidence itself, which varies according to the definition of CKD applied, the time point of ascertainment and whether creatinine-based or measured filtration is used, derive from creatinine-based estimates of glomerular filtration, a measure the same authors concede performs poorly in patients who are sarcopenic, oedematous and metabolically deranged. The second concerns timing, since the most powerful and most modifiable determinant of chronic injury is the passage from early AKI to fixed nephropathy, a transition that draws surveillance away from isolated tacrolimus trough concentrations, which did not predict one year disease in the index cohort, towards intrapatient variability and time within the therapeutic range. The third is therapeutic, because the prevailing synthesis omits the sodium glucose cotransporter 2 inhibitors, agents with established renal benefit in non-transplant CKD populations that deserve formal evaluation after LT. Measured accurately, protected early and evaluated pharmacologically, the kidney of the liver transplant recipient is more defensible than current practice assumes.

Key Words: Chronic kidney disease; Liver transplantation; Calcineurin inhibitors; Cystatin C; Acute kidney injury; Sodium glucose cotransporter 2 inhibitors; Tacrolimus

Core Tip: Sessa et al recast chronic kidney disease after liver transplantation as a preventable threat to survival. We press their case further. The creatinine-based estimates that underpin every reported risk factor, the female sex signal included, should give way to cystatin C, measured glomerular filtration and the albuminuria axis. Surveillance should follow intrapatient tacrolimus variability rather than isolated troughs that fail to predict chronic disease. And nephroprotective pharmacotherapy, sodium glucose cotransporter 2 inhibitors foremost among them, deserves formal trial in this population rather than continued neglect.

TO THE EDITOR

Few complications shape the long-term fortunes of a liver transplant recipient as quietly and as decisively as the slow attrition of renal function. In the article to which we respond, Sessa et al[1] published in the recent issue of the World Journal of Transplantation, survey the contemporary evidence and conclude that chronic kidney disease (CKD) after liver transplantation (LT) is neither uncommon nor inevitable, but a modifiable hazard that the transplant community already possesses the means to soften. Their commentary takes as its anchor the retrospective analysis of Muñoz-Serrano et al[2] who followed 594 adults transplanted across more than three decades and found that close to one half had developed CKD by the first anniversary of their operation, with older age, female sex, preexisting renal dysfunction and cyclosporine standing out as independent correlates. Around this scaffold the authors reconstruct the familiar architecture of post-transplant nephrology, a substantial early incidence that approaches thirty percent at one year and rises beyond forty percent by the fifth, a clear preponderance of stage G3 disease, and a prognostic penalty so steep that the presence of CKD at one year has been linked to a more than fourfold increase in the hazard of death[3,4]. These estimates should be read with care, because the reported incidence of CKD after LT, and the apparent divergence between the near one half observed in the index cohort and the lower figures cited elsewhere, vary according to the definition of CKD that is applied, the time point at which it is ascertained, and whether kidney function is expressed as a creatinine-based estimate or as a measured glomerular filtration rate.

We find little to dispute in this account and a good deal to admire in its clarity. The reframing of renal decline as a preventable rather than an inescapable cost of immunosuppression is both correct and overdue, and the authors are right to locate much of the opportunity in the perioperative period and in the design of the maintenance regimen. Yet a synthesis is only as secure as the measurements on which it rests, the chronology it assumes and the therapies it is willing to consider. It is along these three axes that we wish to extend the conversation. We argue first that the discipline must repair the instrument before it trusts the reading, second that it should treat the acute to chronic transition as the organising principle of prevention rather than as one risk factor among many, and third that it should widen its therapeutic imagination beyond the adjustment of immunosuppressive drugs.

The instrument before the reading

There is an unresolved tension at the heart of the evidence Sessa et al[1] review, and it is worth naming plainly. In their discussion of monitoring they acknowledge, correctly, that serum creatinine is an insensitive marker in sarcopenic and metabolically disturbed patients, and that cystatin C based estimation offers greater accuracy in precisely these contexts[5-7]. Yet every quantitative claim that precedes that acknowledgement, the incidence figures, the odds ratios, the very definition of who has CKD, is built upon creatinine based estimating equations applied to a population in which creatinine is least trustworthy. The decompensated cirrhotic patient is frequently sarcopenic, often oedematous, and generates creatinine at a reduced rate, so a creatinine-based estimate of filtration tends to flatter the kidney before transplantation and to behave erratically afterwards as muscle mass and fluid status shift. Studies that have placed these equations against measured filtration in solid organ recipients have shown considerable and directionally inconsistent error[5], and work using measured glomerular filtration has demonstrated that the relationship between renal function and mortality after LT is both real and underestimated when creatinine alone is used as the lens[6].

This is not a pedantic quibble, because the measurement problem propagates directly into the risk factors the field treats as settled. Consider the association with female sex, which recurs across cohorts and which Sessa et al[1] attempt to explain through the differential accumulation of metabolic comorbidity with age[8,9]. A more parsimonious and more testable explanation begins with physiology. Women carry less muscle and generate less creatinine than men of comparable size, and although estimating equations include a sex term, that correction is calibrated in stable ambulatory populations and is unlikely to hold in the catabolic, fluid shifted state that follows a major hepatic operation. An apparent excess of CKD in women defined by a creatinine-based threshold may therefore reflect, in part, systematic misclassification rather than a distinct biological susceptibility. The claim is falsifiable, which is its virtue. Re estimation with cystatin C, or measurement of filtration directly, would either dissolve the signal or confirm it, and either result would be informative.

The remedy is available and is already endorsed in general nephrology practice. Cystatin C based estimation, measured filtration for pivotal decisions such as candidacy and the timing of combined transplantation, and the routine incorporation of albuminuria as a second and independent axis of risk, are all recommended by current guidance for the evaluation of CKD[8]. Cystatin C is not itself a perfect instrument, and it should not be adopted uncritically. Its concentration is influenced by systemic inflammation, corticosteroid exposure, thyroid dysfunction, adiposity and malignancy, each of which is common after LT, and these non-glomerular determinants can bias estimation in either direction. The reasonable position is therefore not that cystatin C should displace creatinine, but that the two should be used together, ideally through a combined creatinine and cystatin C equation, with measured filtration reserved for decisions of consequence and for the endpoints of trials[5-8]. It is striking that albuminuria, which carries prognostic weight at least equal to filtration in the wider CKD population, is almost entirely absent from the post-transplant cohorts on which these editorials draw, including the index study, which classifies disease by filtration alone[2]. Table 1 sets out how the principal one-year associations reported in that cohort might read once a measurement aware lens is applied, and the exercise suggests that the field has been confident about numbers it should hold more lightly.

Table 1 Reinterpreting the principal one-year renal associations of the index cohort under a measurement aware lens.

Reported association	Conventional reading	Measurement caveat or alternative reading	Practical implication
Female sex (OR 1.88)	A distinct biological susceptibility	Lower muscle mass and reduced creatinine generation depress serum creatinine, and sex coefficients calibrated in stable populations may fail in a catabolic postoperative state, inflating apparent filtration loss	Re estimate with cystatin C or measured filtration before accepting sex as a stratifying target
Older age (OR 1.03 per year)	Nephron senescence	Largely valid, although age also tracks sarcopenia, which biases creatinine	Retain as a stratifier and confirm with cystatin C in the frail
Pre transplant renal dysfunction (OR 2.69)	True baseline vulnerability	Real, but baseline creatinine overestimates filtration in decompensated cirrhosis, so the true burden is under recognised	Use measured or cystatin C based filtration at listing
Cyclosporine vs tacrolimus (OR 3.77)	Greater intrinsic nephrotoxicity of cyclosporine	Confounded by era and partly mediated through blood pressure rather than direct toxicity	Prefer tacrolimus while treating blood pressure as a modifiable mediator
One year incidence near 49%	A fixed disease burden	Inflated and imprecise when anchored to a single creatinine based threshold without albuminuria	Define chronic disease by the full guideline axis, including albuminuria

OR: Odds ratio.

From acute injury to chronic disease, and what the trough cannot tell us

If measurement is the first correction, timing is the second. The most consistent and most actionable observation across this literature is that CKD after LT is, to a substantial degree, the sequela of an acute event. A high serum creatinine before transplantation is among the strongest predictors of subsequent chronic disease, and acute kidney injury (AKI) in the first postoperative days, which afflicts more than half of recipients, independently forecasts later chronic injury[10]. In the index cohort the relationship was unmistakable, since renal dysfunction at one month was the dominant antecedent of CKD at one year[2]. The biology is coherent. An ischaemic and haemodynamically turbulent perioperative course initiates tubular injury, and where repair is incomplete it proceeds through maladaptive fibrosis to fixed nephron loss, a trajectory that the immediate introduction of full dose calcineurin inhibitor can only aggravate, given the structural and vascular toxicity these agents exert on an already vulnerable kidney[11].

Reframing chronic disease as the downstream expression of an early insult relocates the point of greatest leverage to the operating theatre and the first postoperative weeks. Meticulous attention to perioperative haemodynamics, the avoidance of avoidable nephrotoxins, and the strategic delay of calcineurin inhibitor exposure in those entering transplantation with compromised kidneys are interventions of immediate plausibility. The evidence for induction with basiliximab, which permits that delay, is encouraging. Delayed introduction of calcineurin inhibitor under basiliximab cover has been associated with preserved renal function and no penalty in rejection[12], and single centre experience has shown a fall in medium term renal dysfunction from roughly one fifth to one fourteenth of recipients when induction is used to shield the kidney through its most fragile interval[13].

The same reframing exposes the limits of how the field monitors calcineurin inhibitor exposure. It is a quiet but consequential finding of the index cohort that tacrolimus trough concentrations did not distinguish those who developed CKD at one year from those who did not, even though higher early troughs did track with acute renal dysfunction at one month[2]. The dissociation matters. It tells us that a single steady state trough, the number on which most clinical surveillance still turns, is a poor surrogate for chronic risk. The signal lies elsewhere, in cumulative exposure and in instability. Cumulative exposure expressed as the area under the concentration curve has been shown to predict progressive renal impairment more faithfully than spot troughs[14], and intrapatient variability, the restlessness of the concentration over time, has been associated with both CKD and graft loss after LT[15]. That association warrants a careful reading. Intrapatient variability is not a single biological exposure but a common final path for several distinct processes, among them non-adherence, diarrhoea and altered enteric absorption, drug and food interactions, intercurrent illness, inconsistent timing of sampling in relation to dosing, and clinician-initiated dose adjustment. It is best regarded at present as a marker of risk and a plausible causal pathway rather than as a proven independent treatment target, and whether reducing variability itself reduces renal injury remains to be demonstrated prospectively. There is a mechanistic thread worth pulling here as well, since tacrolimus raises blood pressure in part by activating the renal sodium chloride cotransporter[16], which means that some of the renal harm attributed to direct toxicity is mediated through a blood pressure pathway that is itself eminently treatable. Surveillance that follows variability and time within the therapeutic range, rather than isolated troughs, would align monitoring with the biology the data describe.

Widening the therapeutic frame

The therapeutic discussion in the current synthesis is competent but conservative, and it stops at the boundary of immunosuppression. Within that boundary the conclusions are sound. Tacrolimus is preferable to cyclosporine, largely because it controls blood pressure better and thereby blunts a powerful mediator of renal decline[17]. Minimisation of calcineurin inhibitor with mycophenolate improves filtration without forfeiting graft survival, a result supported by meta-analysis[18] and by randomised comparison of reduced dose tacrolimus with mycophenolate against standard exposure[19], and reinforced by induction strategies that pair low dose prolonged release tacrolimus with mycophenolate from the outset[20]. Conversion to an inhibitor of the mammalian target of rapamycin offers a further route to calcineurin inhibitor sparing, with renal benefit balanced against a measurable excess of acute rejection and a malignancy signal that demands careful patient selection[21]. None of this is contentious, and most of it is a decade or more old.

What is conspicuously missing is the pharmacology that has transformed the care of CKD everywhere outside transplantation. Sodium glucose cotransporter 2 (SGLT2) inhibitors slow the progression of CKD and reduce mortality across both diabetic and non-diabetic cause, and that benefit has been extended to broad CKD populations defined by reduced filtration or albuminuria[22,23], although these data are not derived from liver transplant recipients. The nonsteroidal mineralocorticoid receptor antagonist finerenone reduces progression in diabetic kidney disease[24], again in a population that does not include transplant recipients. These agents have not been formally tested in liver transplant recipients, and the omission is no longer defensible, because the very population that Sessa et al[1] rightly identify as carrying the greatest renal jeopardy, those transplanted for metabolic dysfunction associated steatotic liver disease and the dual burden of metabolic and vascular injury it imposes, is precisely the phenotype in which these drugs deliver their largest gains in the general population[25]. What follows is a research prescription rather than a recommendation for practice. SGLT2 inhibitors, and where indicated nonsteroidal mineralocorticoid receptor antagonists, should be evaluated prospectively in stable liver transplant recipients beyond the early postoperative period, in adequately powered trials with safety endpoints defined in advance. Those endpoints should include euglycaemic ketoacidosis, urinary and genital infection, volume depletion and its haemodynamic consequences, hyperkalaemia where mineralocorticoid receptor antagonism is used, interactions with calcineurin inhibitors and inhibitors of the mammalian target of rapamycin, and hepatic graft outcomes. Until such data exist, the case for these agents after LT is a case for trials rather than for extrapolation. Current guidance from the major hepatology and transplantation societies has begun to foreground renal preservation and to formalise criteria for simultaneous liver and kidney transplantation, and the recent AASLD-AST practice guideline on the diagnosis and post-transplant management of non-graft-related complications addresses CKD after LT directly[26,27]. Cystatin C based stratification is nevertheless not yet embedded in routine post-transplant renal surveillance, and no recommendation on these agent classes can at present rest on evidence generated within the transplant population itself. The next revisions should be in a position to close both gaps. Table 2 proposes a staged agenda along these lines.

Table 2 A staged agenda for renal preservation after liver transplantation.

Domain	Prevailing practice	Proposed refinement	Principal supporting evidence
Measurement	Creatinine based estimation against a single threshold	Cystatin C based estimation, measured filtration for pivotal decisions, routine albuminuria	[5-8]
Perioperative care	Early standard dose calcineurin inhibitor	Haemodynamic optimisation, nephrotoxin avoidance, basiliximab enabled delay of calcineurin inhibitor in those at risk	[12,13]
Maintenance and monitoring	Static trough targeting	Mycophenolate enabled minimisation, surveillance by intrapatient variability and time in therapeutic range	[14,15,18,19]
Pharmacological protection	Confined to immunosuppression adjustment	Prospective evaluation of sodium glucose cotransporter 2 inhibitors and, where indicated, nonsteroidal mineralocorticoid receptor antagonists	[22-24]
Trial endpoints	Creatinine derived estimation	Measured filtration and validated injury biomarkers	[6,8]

CONCLUSION

Sessa et al[1] are right in their essential claim, that CKD after LT is a modifiable determinant of survival rather than an inevitable tax on immunosuppression, and they state it without equivocation. Our argument is that the next advance will come not from another revision of the catalogue of predictors, most of which are either fixed or already known, but from three upstream corrections. The first is measurement. Creatinine alone is an inadequate instrument in this population, and cystatin C based estimation, measured glomerular filtration where decisions are pivotal, and albuminuria as a routine second axis should be adopted, with the limitations of each acknowledged. The second is prevention of the transition from AKI to CKD, which places the emphasis on perioperative haemodynamic protection, the avoidance of nephrotoxins, and the induction enabled delay of calcineurin inhibitor exposure in those at risk, monitored by intrapatient variability and time within the therapeutic range rather than by isolated troughs. The third is the formal evaluation of nephroprotective pharmacotherapy, foremost the SGLT2 inhibitors, in adequately powered trials in liver transplant recipients, with prespecified safety endpoints and with renal outcomes judged by measured filtration and validated injury markers rather than by creatinine alone. These three priorities, accurate measurement, early prevention and evidence-based nephroprotection, define the agenda on which the renal fortunes of the liver transplant recipient now depend.